Download MBA (Master of Business Administration) IB and Marketing 1st and 2nd Semester Quantitative Techniques for Managers Notes
MBA-H2040 Quantitative Techniques for Managers
UNIT I
1 INTRODUCTION TO OPERATIONS RESEARCH
ON
S
S
E
L
LESSON STRUCTURE
1.1 Introduction
1.2 History of Operations Research
1.3 Stages of Development of Operations
Research
1.4 Relationship Between Manager and
OR Specialist
1.5 OR Tools and Techniques
1.6 Applications of Operations Research
1.7 Limitations of Operations Research
1.8 Summary
Objectives
1.9 Key Terms
After Studying this lesson, you should be able
1.10 Self Assessment Questions
to:
1.11 Further References
Understand the meaning, purpose, and
tools of Operations Research
Describe the history of Operations
Research
Describe the Stages of O.R
Explain the Applications of Operations
Research
Describe the Limitations of Operation
Research
Understand the OR specialist and
Manager relationship
MBA-H2040 Quantitative Techniques for Managers
1.1 Introduction
The British/Europeans refer to "operational research", the Americans to "operations research" - but both
are often shortened to just "OR" - which is the term we will use.
Another term which is used for this field is "management science" ("MS"). The Americans
sometimes combine the terms OR and MS together and say "OR/MS" or "ORMS". Yet other terms
sometimes used are "industrial engineering" ("IE") and "decision science" ("DS"). In recent years there
has been a move towards a standardization upon a single term for the field, namely the term "OR".
Operation Research is a relatively new discipline. The contents and the boundaries of the OR are
not yet fixed. Therefore, to give a formal definition of the term Operations Research is a difficult task.
The OR starts when mathematical and quantitative techniques are used to substantiate the decision being
taken. The main activity of a manager is the decision making. In our daily life we make the decisions
even without noticing them. The decisions are taken simply by common sense, judgment and expertise
without using any mathematical or any other model in simple situations. But the decision we are
concerned here with are complex and heavily responsible. Examples are public transportation network
planning in a city having its own layout of factories, residential blocks or finding the appropriate product
mix when there exists a large number of products with different profit contributions and production
requirement etc.
Operations Research tools are not from any one discipline. Operations Research takes tools from
different discipline such as mathematics, statistics, economics, psychology, engineering etc. and
combines these tools to make a new set of knowledge for decision making. Today, O.R. became a
professional discipline which deals with the application of scientific methods for making decision, and
especially to the allocation of scarce resources. The main purpose of O.R. is to provide a rational basis
for decisions making in the absence of complete information, because the systems composed of human,
machine, and procedures may do not have complete information.
Operations Research can also be treated as science in the sense it describing, understanding and
predicting the systems behaviour, especially man-machine system. Thus O.R. specialists are involved in
three classical aspect of science, they are as follows:
i) Determining the systems behaviour
ii) Analyzing the systems behaviour by developing appropriate models
iii) Predict the future behaviour using these models
The emphasis on analysis of operations as a whole distinguishes the O.R. from other research
and engineering. O.R. is an interdisciplinary discipline which provided solutions to problems of military
operations during World War II, and also successful in other operations. Today business applications are
MBA-H2040 Quantitative Techniques for Managers
primarily concerned with O.R. analysis for the possible alternative actions. The business and industry
befitted from O.R. in the areas of inventory, reorder policies, optimum location and size of warehouses,
advertising policies, etc.
As stated earlier defining O.R. is a difficult task. The definitions stressed by various experts and
Societies on the subject together enable us to know what O.R. is, and what it does. They are as follows:
1. According to the Operational Research Society of Great Britain (OPERATIONAL RESEARCH
QUARTERLY, l3(3):282, l962), Operational Research is the attack of modern science on
complex problems arising in the direction and management of large systems of men, machines,
materials and money in industry, business, government and defense. Its distinctive approach is to
develop a scientific model of the system, incorporating measurements of factors such as change
and risk, with which to predict and compare the outcomes of alternative decisions, strategies or
controls. The purpose is to help management determine its policy and actions scientifically.
2. Randy Robinson stresses that Operations Research is the application of scientific methods to
improve the effectiveness of operations, decisions and management. By means such as analyzing
data, creating mathematical models and proposing innovative approaches, Operations Research
professionals develop scientifically based information that gives insight and guides decision-
making. They also develop related software, systems, services and products.
3. Morse and Kimball have stressed O.R. is a quantitative approach and described it as " a scientific
method of providing executive departments with a quantitative basis for decisions regarding the
operations under their control".
4. Saaty considers O.R. as tool of improving quality of answers. He says, "O.R. is the art of giving
bad answers to problems which otherwise have worse answers".
5. Miller and Starr state, "O.R. is applied decision theory, which uses any scientific, mathematical
or logical means to attempt to cope with the problems that confront the executive, when he tries
to achieve a thorough-going rationality in dealing with his decision problem".
6. Pocock stresses that O.R. is an applied Science. He states "O.R. is scientific methodology
(analytical, mathematical, and quantitative) which by assessing the overall implication of various
alternative courses of action in a management system provides an improved basis for
management decisions".
1.2 History of Operations Research
Operation Research is a relatively new discipline. Whereas 70 years ago it would have been possible to
study mathematics, physics or engineering (for example) at university it would not have been possible to
study Operation Research, indeed the term O.R. did not exist then. It was really only in the late 1930's
that operational research began in a systematic fashion, and it started in the UK. As such it would be
interesting to give a short history of O.R.
1936
MBA-H2040 Quantitative Techniques for Managers
Early in 1936 the British Air Ministry established Bawdsey Research Station, on the east coast, near
Felixstowe, Suffolk, as the centre where all pre-war radar experiments for both the Air Force and the
Army would be carried out. Experimental radar equipment was brought up to a high state of reliabilit y
and ranges of over 100 miles on aircraft were obtained.
It was also in 1936 that Royal Air Force (RAF) Fighter Command, charged specifically with the
air defense of Britain, was first created. It lacked however any effective fighter aircraft - no Hurricanes
or Spitfires had come into service - and no radar data was yet fed into its very elementary warning and
control system.
It had become clear that radar would create a whole new series of problems in fighter direction
and control so in late 1936 some experiments started at Biggin Hill in Kent into the effective use of such
data. This early work, attempting to integrate radar data with ground based observer data for fighter
interception, was the start of OR.
1937
The first of three major pre-war air-defence exercises was carried out in the summer of 1937. The
experimental radar station at Bawdsey Research Station was brought into operation and the information
derived from it was fed into the general air-defense warning and control system. From the early warning
point of view this exercise was encouraging, but the tracking information obtained from radar, after
filtering and transmission through the control and display network, was not very satisfactory.
1938
In July 1938 a second major air-defense exercise was carried out. Four additional radar stations had been
installed along the coast and it was hoped that Britain now had an aircraft location and control system
greatly improved both in coverage and effectiveness. Not so! The exercise revealed, rather, that a new
and serious problem had arisen. This was the need to coordinate and correlate the additional, and often
conflicting, information received from the additional radar stations. With the outbreak of war apparently
imminent, it was obvious that something new - drastic if necessary - had to be attempted. Some new
approach was needed.
Accordingly, on the termination of the exercise, the Superintendent of Bawdsey Research Station, A.P.
Rowe, announced that although the exercise had again demonstrated the technical feasibility of the radar
system for detecting aircraft, its operational achievements still fell far short of requirements. He
therefore proposed that a crash program of research into the operational - as opposed to the technical -
aspects of the system should begin immediately. The term "operational research" [RESEARCH into
MBA-H2040 Quantitative Techniques for Managers
(military) OPERATIONS] was coined as a suitable description of this new branch of applied science.
The first team was selected from amongst the scientists of the radar research group the same day.
1939
In the summer of 1939 Britain held what was to be its last pre-war air defence exercise. It involved some
33,000 men, 1,300 aircraft, 110 antiaircraft guns, 700 searchlights, and 100 barrage balloons. This
exercise showed a great improvement in the operation of the air defence warning and control system.
The contribution made by the OR team was so apparent that the Air Officer Commander-in-Chief RAF
Fighter Command (Air Chief Marshal Sir Hugh Dowding) requested that, on the outbreak of war, they
should be attached to his headquarters at Stanmore in north London.
Initially, they were designated the "Stanmore Research Section". In 1941 they were redesignated
the "Operational Research Section" when the term was formalised and officially accepted, and similar
sections set up at other RAF commands.
1940
On May 15th 1940, with German forces advancing rapidly in France, Stanmore Research Section was
asked to analyses a French request for ten additional fighter squadrons (12 aircraft a squadron - so 120
aircraft in all) when losses were running at some three squadrons every two days (i.e. 36 aircraft every 2
days). They prepared graphs for Winston Churchill (the British Prime Minister of the time), based upon
a study of current daily losses and replacement rates, indicating how rapidly such a move would deplete
fighter strength. No aircraft were sent and most of those currently in France were recalled.
This is held by some to be the most strategic contribution to the course of the war made by OR
(as the aircraft and pilots saved were consequently available for the successful air defense of Britain, the
Battle of Britain).
1941 onward
In 1941, an Operational Research Section (ORS) was established in Coastal Command which was to
carry out some of the most well-known OR work in World War II.
The responsibility of Coastal Command was, to a large extent, the flying of long-range sorties by
single aircraft with the object of sighting and attacking surfaced U-boats (German submarines). The
technology of the time meant that (unlike modern day submarines) surfacing was necessary to recharge
batteries, vent the boat of fumes and recharge air tanks. Moreover U-boats were much faster on the
surface than underwater as well as being less easily detected by sonar.
MBA-H2040 Quantitative Techniques for Managers
Thus the Operation Research started just before World War II in Britain with the establishment
of teams of scientists to study the strategic and tactical problems involved in military operations. The
objective was to find the most effective utilization of limited military resources by the use of quantitative
techniques. Following the end of the war OR spread, although it spread in different ways in the UK and
USA.
In 1951 a committee on Operations Research formed by the National Research Council of USA,
and the first book on "Methods of Operations Research", by Morse and Kimball, was published. In 1952
the Operations Research Society of America came into being.
Success of Operations Research in army attracted the attention of the industrial mangers who
were seeking solutions to their complex business problems. Now a days, almost every organization in all
countries has staff applying operations research, and the use of operations research in government has
spread from military to wide variety of departments at all levels. The growth of operations research has
not limited to the U.S.A. and U.K., it has reached many countries of the world.
India was one the few first countries who started using operations research. In India, Regional
Research Laboratory located at Hyderabad was the first Operations Research unit established during
1949. At the same time another unit was set up in Defense Science Laboratory to solve the Stores,
Purchase and Planning Problems. In 1953, Operations Research unit was established in Indian Statistical
Institute, Calcutta, with the objective of using Operations Research methods in National Planning and
Survey. In 1955, Operations Research Society of India was formed, which is one of the first members of
International Federation of Operations Research societies. Today Operations Research is a popular
subject in management institutes and schools of mathematics.
1.3 Stages of Development of Operations Research
The stages of development of O.R. are also known as phases and process of O.R, which has six
important steps. These six steps are arranged in the following order:
Step I: Observe the problem environment
Step II: Analyze and define the problem
Step III: Develop a model
Step IV: Select appropriate data input
Step V: Provide a solution and test its reasonableness
Step VI: Implement the solution
MBA-H2040 Quantitative Techniques for Managers
Step I: Observe the problem environment
The first step in the process of O.R. development is the problem environment observation. This step
includes different activities; they are conferences, site visit, research, observations etc. These activities
provide sufficient information to the O.R. specialists to formulate the problem.
Step II: Analyze and define the problem
This step is analyzing and defining the problem. In this step in addition to the problem definition the
objectives, uses and limitations of O.R. study of the problem also defined. The outputs of this step are
clear grasp of need for a solution and its nature understanding.
Step III: Develop a model
This step develops a model; a model is a representation of some abstract or real situation. The models
are basically mathematical models, which describes systems, processes in the form of equations,
formula/relationships. The different activities in this step are variables definition, formulating equations
etc. The model is tested in the field under different environmental constraints and modified in order to
work. Some times the model is modified to satisfy the management with the results.
Step IV: Select appropriate data input
A model works appropriately when there is appropriate data input. Hence, selecting appropriate input
data is important step in the O.R. development stage or process. The activities in this step include
internal/external data analysis, fact analysis, and collection of opinions and use of computer data banks.
The objective of this step is to provide sufficient data input to operate and test the model developed in
Step_III.
Step V: Provide a solution and test its reasonableness
This step is to get a solution with the help of model and input data. This solution is not implemented
immediately, instead the solution is used to test the model and to find there is any limitations. Suppose if
the solution is not reasonable or the behaviour of the model is not proper, the model is updated and
modified at this stage. The output of this stage is the solution(s) that supports the current organizational
objectives.
Step VI: Implement the solution
MBA-H2040 Quantitative Techniques for Managers
At this step the solution obtained from the previous step is implemented. The implementation of the
solution involves mo many behavioural issues. Therefore, before implementation the implementation
authority has to resolve the issues. A properly implemented solution results in quality of work and gains
the support from the management.
The process, process activities, and process output are summarized in the following Table 1-1.
Process Activities
Process
Process Output
Site visits, Conferences,
Sufficient information and
Step 1:
Observations, Research
support to proceed
Observe the problem
environment
Define: Use, Objectives,
Clear grasp of need for and
limitations
Step 2:
nature of solution requested
Analyze and define
the problem
Define interrelationships,
Models that works under stated
Formulate equations,
environmental constraints
Step 3:
Develop a Model
Use known O.R. Model ,
Search alternate Model
Analyze: internal-external data,
Sufficient inputs to operate and
facts
Step 4:
test model
Select appropriate data
Collect options,
input
Use computer data banks
Test the model
Solution(s) that support current
Step 5:
organizational goals
find limitations
Provide a solution and
test its reasonableness
update the model
Step 6:
Implement the
solution
MBA-H2040 Quantitative Techniques for Managers
Resolve behavioural issues
Improved working and
Management support for longer
Sell the idea
run operation of model
Give explanations
Management involvement
Table 1-1: Process, Process activities, Process output of O.R. development stages
1.4 Relationship between the Manager and O.R. Specialist
The key responsibility of manager is decision making. The role of the O.R. specialist is to help the
manager make better decisions. Figure 1-1 explains the relationship between the O.R. specialist and the
manager/decision maker.
STEPS IN PROBLEM RECOGNITION,
INVOLVEMENT: O.R. SPECIALIST or
FORMULATION AND SOLUTION
MANAGER
Recognize
from
organizational
Manager
symptoms that a problem exists.
Decide what variables are involved; state
the problem in quantitative relationships
Manager and O.R. Specialist
among the variables.
I
nvestigat
e met
hods f
or sol
ving t
he
problems as stated above; determine
O.R. Specialist
appropriate quantitative tools to be used.
Attempt solutions to the problems; find
various solutions; state assumptions
O.R. Specialist
underlying
these
solutions;
test
alternative solutions.
Determine which solution is most
effective because of practical constraints
Manager and O.R. Specialist
wit hin the organi zation; decide what t
he
solution means for the organization.
MBA-H2040 Quantitative Techniques for Managers
Choose the solution to be used.
Manager
`
S
e l l ' t
h
e
de
c i s
i
o
n
t
o
o
pe
r
a
t i
n
g
m
a
na
ge
r
s ;
Manager and O.R. Specialist
get their understanding and cooperation.
Figure 1-1 Relationship between Manager/Decision Maker and O.R. Specialists
1.5 O.R. Tools and Techniques
Operations Research uses any suitable tools or techniques available. The common frequently used
tools/techniques are mathematical procedures, cost analysis, electronic computation. However,
operations researchers given special importance to the development and the use of techniques like linear
programming, game theory, decision theory, queuing theory, inventory models and simulation. In
addition to the above techniques, some other common tools are non-linear programming, integer
programming, dynamic programming, sequencing theory, Markov process, network scheduling
(PERT/CPM), symbolic Model, information theory, and value theory. There is many other Operations
Research tools/techniques also exists. The brief explanations of some of the above techniques/tools are
as follows:
Linear Programming:
This is a constrained optimization technique, which optimize some criterion within some constraints. In
Linear programming the objective function (profit, loss or return on investment) and constraints are
linear. There are different methods available to solve linear programming.
Game Theory:
This is used for making decisions under conflicting situations where there are one or more
players/opponents. In this the motive of the players are dichotomized. The success of one player tends to
be at the cost of other players and hence they are in conflict.
Decision Theory:
Decision theory is concerned with making decisions under conditions of complete certainty about the
future outcomes and under conditions such that we can make some probability about what will happen
in future.
Queuing Theory:
MBA-H2040 Quantitative Techniques for Managers
This is used in situations where the queue is formed (for example customers waiting for service, aircrafts
waiting for landing, jobs waiting for processing in the computer system, etc). The objective here is
minimizing the cost of waiting without increasing the cost of servicing.
Inventory Models:
Inventory model make a decisions that minimize total inventory cost. This model successfully reduces
the total cost of purchasing, carrying, and out of stock inventory.
Simulation:
Simulation is a procedure that studies a problem by creating a model of the process involved in the
problem and then through a series of organized trials and error solutions attempt to determine the best
solution. Some times this is a difficult/time consuming procedure. Simulation is used when actual
experimentation is not feasible or solution of model is not possible.
Non-linear Programming:
This is used when the objective function and the constraints are not linear in nature. Linear relationships
may be applied to approximate non-linear constraints but limited to some range, because approximation
becomes poorer as the range is extended. Thus, the non-linear programming is used to determine the
approximation in which a solution lies and then the solution is obtained using linear methods.
Dynamic Programming:
Dynamic programming is a method of analyzing multistage decision processes. In this each elementary
decision depends on those preceding decisions and as well as external factors.
Integer Programming:
If one or more variables of the problem take integral values only then dynamic programming method is
used. For example number or motor in an organization, number of passenger in an aircraft, number of
generators in a power generating plant, etc.
Markov Process:
Markov process permits to predict changes over time information about the behavior of a system is
known. This is used in decision making in situations where the various states are defined. The
probability from one state to another state is known and depends on the current state and is independent
of how we have arrived at that particular state.
MBA-H2040 Quantitative Techniques for Managers
Network Scheduling:
This technique is used extensively to plan, schedule, and monitor large projects (for example computer
system installation, R & D design, construction, maintenance, etc.). The aim of this technique is
minimize trouble spots (such as delays, interruption, production bottlenecks, etc.) by identifying the
critical factors. The different activities and their relationships of the entire project are represented
diagrammatically with the help of networks and arrows, which is used for identifying critical activities
and path. There are two main types of technique in network scheduling, they are:
Program Evaluation and Review Technique (PERT) ? is used when activities time is not known
accurately/ only probabilistic estimate of time is available.
Critical Path Method (CPM) ? is used when activities time is know accurately.
Information Theory:
This analytical process is transferred from the electrical communication field to O.R. field. The
objective of this theory is to evaluate the effectiveness of flow of information with a given system. This
is used mainly in communication networks but also has indirect influence in simulating the examination
of business organizational structure with a view of enhancing flow of information.
1.6 Applications of Operations Research
Today, almost all fields of business and government utilizing the benefits of Operations Research. There
are voluminous of applications of Operations Research. Although it is not feasible to cover all
applications of O.R. in brief. The following are the abbreviated set of typical operations research
applications to show how widely these techniques are used today:
Accounting:
Assigning audit teams effectively
Credit policy analysis
Cash flow planning
Developing standard costs
Establishing costs for byproducts
Planning of delinquent account strategy
Construction:
Project scheduling, monitoring and control
Determination of proper work force
Deployment of work force
Allocation of resources to projects
Facilities Planning:
Factory location and size decision
Estimation of number of facilities required
Hospital planning
MBA-H2040 Quantitative Techniques for Managers
International logistic system design
Transportation loading and unloading
Warehouse location decision
Finance:
Building cash management models
Allocating capital among various alternatives
Building financial planning models
Investment analysis
Portfolio analysis
Dividend policy making
Manufacturing:
Inventory control
Marketing balance projection
Production scheduling
Production smoothing
Marketing:
Advertising budget allocation
Product introduction timing
Selection of Product mix
Deciding most effective packaging alternative
Organizational Behavior / Human Resources:
Personnel planning
Recruitment of employees
Skill balancing
Training program scheduling
Designing organizational structure more effectively
Purchasing:
Optimal buying
Optimal reordering
Materials transfer
Research and Development:
R & D Projects control
R & D Budget allocation
Planning of Product introduction
1.7 Limitations of Operations Research
Operations Research has number of applications; similarly it also has certain limitations. These
limitations are mostly related to the model building and money and time factors problems involved in its
application. Some of them are as given below:
i)
Distance between O.R. specialist and Manager
Operations Researchers job needs a mathematician or statistician, who might not be aware of
the business problems. Similarly, a manager is unable to understand the complex nature of
Operations Research. Thus there is a big gap between the two personnel.
ii)
Magnitude of Calculations
MBA-H2040 Quantitative Techniques for Managers
The aim of the O.R. is to find out optimal solution taking into consideration all the factors. In
this modern world these factors are enormous and expressing them in quantitative model and
establishing relationships among these require voluminous calculations, which can be
handled only by machines.
iii)
Money and Time Costs
The basic data are subjected to frequent changes, incorporating these changes into the
operations research models is very expensive. However, a fairly good solution at present may
be more desirable than a perfect operations research solution available in future or after some
time.
iv)
Non-quantifiable Factors
When all the factors related to a problem can be quantifiable only then operations research
provides solution otherwise not. The non-quantifiable factors are not incorporated in O.R.
models. Importantly O.R. models do not take into account emotional factors or qualitative
factors.
v)
Implementation
Once the decision has been taken it should be implemented. The implementation of decisions
is a delicate task. This task must take into account the complexities of human relations and
behavior and in some times only the psychological factors.
1.8 Summary
Operations Research is relatively a new discipline, which originated in World War II, and became very
popular throughout the world. India is one of the few first countries in the world who started using
operations research. Operations Research is used successfully not only in military/army operations but
also in business, government and industry. Now a day's operations research is almost used in all the
fields.
Proposing a definition to the operations research is a difficult one, because its boundary and
content are not fixed. The tools for operations search is provided from the subject's viz. economics,
engineering, mathematics, statistics, psychology, etc., which helps to choose possible alternative courses
of action. The operations research tool/techniques include linear programming, non-linear programming,
dynamic programming, integer programming, Markov process, queuing theory, etc.
Operations Research has a number of applications. Similarly it has a number of limitations,
which is basically related to the time, money, and the problem involves in the model building. Day-by-
day operations research gaining acceptance because it improve decision making effectiveness of the
managers. Almost all the areas of business use the operations research for decision making.
1.9 Key Terms
MBA-H2040 Quantitative Techniques for Managers
OR: Operations Research.
MS: Management Science.
Symbolic Model: An abstract model, generally using mathematical symbols.
Criterion: is measurement, which is used to evaluation of the results.
Integer Programming: is a technique, which ensures only integral values of variables in the problem.
Dynamic Programming: is a technique, which is used to analyze multistage decision process.
Linear Programming: is a technique, which optimizes linear objective function under limited
constraints.
Inventory Model: these are the models used to minimize total inventory costs.
Optimization: Means maximization or minimization.
1.10 Self Assessment Questions
Q1. Define Operations Research.
Q2. Describe the relationship between the manager and O.R. specialist.
Q3. Explain the various steps in the O.R. development process.
Q4. Discuss the applications of O.R.
Q5. Discuss the limitation of O.R.
Q6. Describe different techniques of O.R.
Q7. Discuss few areas of O.R. applications in your organization or organization you are familiar with.
1.11 Further References
Hamdy A Taha, 1999. Introduction to Operations Research, PHI Limited, New Delhi.
Sharma, J.K., 1989. Mathematical Models in Operations Research, Tata McGraw Hill Publishing
Company Ltd., New Delhi.
Beer, Stafford, 1966. Decision and Control, John Wiley & Sons, Inc., New York.
Levin, Rubin, Stinson, Gardner, 1992. Quantitative Approaches to Management, Tata McGraw Hill
Publishing Company Ltd. New Delhi.
Wagner, Harvery M., 1975. Principles of Operations Research, PHI, Egnlewood Cliffs, N.J.
MBA-H2040 Quantitative Techniques for Managers
UNIT I
ON
S
2 LINEAR PROGRAMMING ?GRAPHICAL METHOD
S
E
L
LESSON STRUCTURE
2.1 Introduction to Linear Programming
2.2 Linear Programming Problem
Formulation
2.3 Formulation with Different Types of
Constraints
2.4 Graphical Analysis of Linear
Programming
2.5 Graphical Linear Programming Solution
2.6 Multiple Optimal Solutions
2.7 Unbounded Solution
2.8 Infeasible Solution
2.9 Summary
2.10 Key Terms
2.11 Self Assessment Questions
2.12 Key Solutions
2.13 Further References
MBA-H2040 Quantitative Techniques for Managers
Objectives
After studying this lesson, you should be able
to:
Formulate Linear Programming Problem
Identify the characteristics of linear
programming problem
Make a graphical analysis of the linear
programming problem
Solve the problem graphically
Identify the various types of solutions
MBA-H2040 Quantitative Techniques for Managers
2.1 Introduction to Linear Programming
Linear Programming is a special and versatile technique which can be applied to a variety of
management problems viz. Advertising, Distribution, Investment, Production, Refinery Operations, and
Transportation analysis. The linear programming is useful not only in industry and business but also in
non-profit sectors such as Education, Government, Hospital, and Libraries. The linear programming
method is applicable in problems characterized by the presence of decision variables. The objective
function and the constraints can be expressed as linear functions of the decision variables. The
decision variables represent quantities that are, in some sense, controllable inputs to the system being
modeled. An objective function represents some principal objective criterion or goal that measures the
effectiveness of the system such as maximizing profits or productivity, or minimizing cost or
consumption. There is always some practical limitation on the availability of resources viz. man,
material, machine, or time for the system. These constraints are expressed as linear equations involving
the decision variables. Solving a linear programming problem means determining actual values of the
decision variables that optimize the objective function subject to the limitation imposed by the
constraints.
The main important feature of linear programming model is the presence of linearity in the
problem. The use of linear programming model arises in a wide variety of applications. Some model
may not be strictly linear, but can be made linear by applying appropriate mathematical transformations.
Still some applications are not at all linear, but can be effectively approximated by linear models. The
ease with which linear programming models can usually be solved makes an attractive means of dealing
with otherwise intractable nonlinear models.
2.2 Linear Programming Problem Formulation
The linear programming problem formulation is illustrated through a product mix problem. The product
mix problem occurs in an industry where it is possible to manufacture a variety of products. A product
has a certain margin of profit per unit, and uses a common pool of limited resources. In this case the
linear programming technique identifies the products combination which will maximize the profit
subject to the availability of limited resource constraints.
Example 2.1:
Suppose an industry is manufacturing tow types of products P1 and P2. The profits per Kg of the two
products are Rs.30 and Rs.40 respectively. These two products require processing in three types of
machines. The following table shows the available machine hours per day and the time required on each
18
MBA-H2040 Quantitative Techniques for Managers
machine to produce one Kg of P1 and P2. Formulate the problem in the form of linear programming
model.
Profit/Kg
P1
P2
Total available Machine
Rs.30
Rs.40
hours/day
Machine 1
3
2
600
Machine 2
3
5
800
Machine 3
5
6
1100
Solution:
The procedure for linear programming problem formulation is as follows:
Introduce the decision variable as follows:
Let x1 = amount of P1
x2 = amount of P2
In order to maximize profits, we establish the objective function as
30x1 + 40x2
Since one Kg of P1 requires 3 hours of processing time in machine 1 while the corresponding
requirement of P2 is 2 hours. So, the first constraint can be expressed as
3x1 + 2x2 600
Similarly, corresponding to machine 2 and 3 the constraints are
3x1 + 5x2 800
5x1 + 6x2 1100
In addition to the above there is no negative production, which may be represented algebraically as
x1 0
;
x2 0
Thus, the product mix problem in the linear programming model is as follows:
Maximize
30x1 + 40x2
Subject to:
3x1 + 2x2 600
3x1 + 5x2 800
5x1 + 6x2 1100
x1 0, x2 0
2.3 Formulation with Different Types of Constraints
19
MBA-H2040 Quantitative Techniques for Managers
The constraints in the previous example 2.1 are of "less than or equal to" type. In this section we are
going to discuss the linear programming problem with different constraints, which is illustrated in the
following Example 2.2.
Example 2.2:
A company owns two flour mills viz. A and B, which have different production capacities for high,
medium and low quality flour. The company has entered a contract to supply flour to a firm every month
with at least 8, 12 and 24 quintals of high, medium and low quality respectively. It costs the company
Rs.2000 and Rs.1500 per day to run mill A and B respectively. On a day, Mill A produces 6, 2 and 4
quintals of high, medium and low quality flour, Mill B produces 2, 4 and 12 quintals of high, medium
and low quality flour respectively. How many days per month should each mill be operated in order to
meet the contract order most economically.
Solution:
Let us define x1 and x2 are the mills A and B. Here the objective is to minimize the cost of the machine
runs and to satisfy the contract order. The linear programming problem is given by
Minimize
2000x1 + 1500x2
Subject to:
6x1 + 2x2 8
2x1 + 4x2 12
4x1 + 12x2 24
x1 0, x2 0
2.4 Graphical Analysis of Linear Programming
This section shows how a two-variable linear programming problem is solved graphically, which is
illustrated as follows:
Example 2.3:
Consider the product mix problem discussed in section 2.2
Maximize
30x1 + 40x2
Subject to:
3x1 + 2x2 600
3x1 + 5x2 800
5x1 + 6x2 1100
x1 0, x2 0
20
MBA-H2040 Quantitative Techniques for Managers
From the first constraints 3x1 + 2x2 600, draw the line 3x1 + 2x2 = 600 which passes through the point
(200, 0) and (0, 300). This is shown in the following graph as line 1.
300
3x1 + 2x2 = 600(line 1)
200
B
X2
C
100
3x1 + 5x2 = 800(line 2)
5x1 + 6x2 = 1100(line 3)
A
D
0
50
100
150
X1
200
275
Graph 1: Three closed half planes and Feasible Region
Half Plane
- A linear inequality in two variables is called as a half plane.
Boundary
- The corresponding equality (line) is called as the boundary of the half plane.
Close Half Plane ? Half plane with its boundary is called as a closed half plane.
21
MBA-H2040 Quantitative Techniques for Managers
In this case we must decide in which side of the line 3x1 + 2x2 = 600 the half plane is located. The
easiest way to solve the inequality for x2 is
3x1 600 ? 2x2
And for the fixed x1, the coordinates satisfy this inequality are smaller than the corresponding ordinate
on the line and thus the inequality is satisfied for all the points below the line 1.
Similarly, we have to determine the closed half planes for the inequalities 3x1 + 5x2 800 and
5x1 + 6x2 1100 (line 2 and line 3 in the graph). Since all the three constraints must be satisfied
simultaneously we have consider the intersection of these three closed half planes. The complete
intersection of these three closed half planes is shown in the above graph as ABCD. The region ABCD
is called the feasible region, which is shaded in the graph.
Feasible Solution:
Any non-negative value of x1, x2 that is x1 0 and x2 0 is known as feasible solution of the linear
programming problem if it satisfies all the existing constraints.
Feasible Region:
The collection of all the feasible solution is called as the feasible region.
Example 2.4:
In the previous example we discussed about the less than or equal to type of linear programming
problem, i.e. maximization problem. Now consider a minimization (i.e. greater than or equal to type)
linear programming problem formulated in Example 2.2.
Minimize
2000x1 + 1500x2
Subject to:
6x1 + 2x2 8
2x1 + 4x2 12
4x1 + 12x2 24
x1 0, x2 0
The three lines 6x1 + 2x2 = 8, 2x1 + 4x2 = 12, and 4x1 + 12x2 = 24 passes through the point
(1.3,0) (0,4), (6,0) (0,3) and (6,0) (0,2). The feasible region for this problem is shown in the following
Graph 2. In this problem the constraints are of greater than or equal to type of feasible region, which is
bounded on one side only.
22
MBA-H2040 Quantitative Techniques for Managers
8
6
X2
A
4
6x
1 + 2x2 8
B
2
2x1 + 4x2 12
C
4x1 + 12x2 24
0
X1
2
4
6
8
Graph 2: Feasible Region
2.5 Graphical Liner Programming Solution
A two variable linear programming problem can be easily solved graphically. The method is simple but
the principle of solution is depends on certain analytical concepts, they are:
Convex Region:
A region R is convex if and only if for any two points on the region R the line connecting those points
lies entirely in the region R.
Extreme Point:
The extreme point E of a convex region R is a point such that it is not possible to locate two distinct
points in R, so that the line joining them will include E. The extreme points are also called as corner
points or vertices.
Thus, the following result provides the solution to the linear programming model:
"If the minimum or maximum value of a linear function defined over a convex region exists,
then it must be on one of the extreme points".
23
MBA-H2040 Quantitative Techniques for Managers
In this section we are going to describe linear programming graphical solution for both the
maximization and minimization problems, discussed in Example 2.3 and Example 2.4.
Example 2. 5:
Consider the maximization problem described in Example 2.3.
Maximize
30x1 + 40x2
Subject to:
3x1 + 2x2 600
3x1 + 5x2 800
5x1 + 6x2 1100
M = 30x
x
1 +40x2
1 0, x2 0
The feasible region identified in the Example 2.3 is a convex polygon, which is illustrated in the
following Graph 3. The extreme point of this convex region are A, B, C, D and E.
300
200
B
X2
C
100
D
A
E
0
50
100
150
X1
200
275
24
MBA-H2040 Quantitative Techniques for Managers
Graph 3: Graphical Linear Programming Solution
In this problem the objective function is 30x1 + 40x2. Let be M is a parameter, the graph 30x1 +
40x2 = M is a group of parallel lines with slope ? 30/40. Some of these lines intersects the feasible region
and contains many feasible solutions, whereas the other lines miss and contain no feasible solution. In
order to maximize the objective function, we find the line of this family that intersects the feasible
region and is farthest out from the origin. Note that the farthest is the line from the origin the greater will
be the value of M.
Observe that the line 30x1 + 40x2 = M passes through the point D, which is the intersection of the
lines 3x1 + 5x2 = 800 and 5x1 + 6x2 = 1100 and has the coordinates x1 = 170 and x2 = 40. Since D is the
only feasible solution on this line the solution is unique.
The value of M is 6700, which is the objective function maximum value. The optimum value
variables are x1 = 170 and X2 = 40.
The following Table 1 shows the calculation of maximum value of the objective function.
Extreme Point
Coordinates
Objective Function
X1 X2
30x1 + 40x2
A
X1 = 0
X2 = 0
0
B
X1 = 0
X2 = 160
6400
C
X1 = 110
X2 = 70
6100
D
X1 = 170
X2 = 40
6700
E
X1 = 200
X2 = 0
6000
Table 1: Shows the objective function Maximum value calculation
Example 2.6:
Consider the minimization problem described in Example 2.4.
Minimize
2000x1 + 1500x2
Subject to:
6x1 + 2x2 8
2x1 + 4x2 12
4x1 + 12x2 24
x1 0, x2 0
The feasible region for this problem is illustrated in the following Graph 4. Here each of the half
planes lies above its boundary. In this case the feasible region is infinite. In this case, we are concerned
with the minimization; also it is not possible to determine the maximum value. As in the previous
25
MBA-H2040 Quantitative Techniques for Managers
example let us introduce a parameter M in the objective function i.e. 2000x1 + 1500x2 = M and draw the
lines for different values of M, which is shown in the following Table 2.
8
6
X2
A
4
B
2
C
0
2000x1+ 1500x
2=M
X1
2
4
6
8
Graph 4: Graphical Linear Programming Solution
Extreme Point
Coordinates
Objective Function
X1 X2
2000x1 + 1500x2
A
X1 = 0
X2 = 4
6000
B
X1 = 0.5
X2 = 2.75
5125
C
X1 = 6
X2 = 0
12000
Table 2: Shows the objective function Minimum value computation
The minimum value is 5125 at the extreme point B, which is the value of the M (objective
function). The optimum values variables are X1 = 0.5 and X2 = 2.75.
2.6 Multiple Optimal Solutions
26
MBA-H2040 Quantitative Techniques for Managers
When the objective function passed through only the extreme point located at the intersection of two
half planes, then the linear programming problem possess unique solutions. The previous examples i.e.
Example 2.5 and Example 2.6 are of this types (which possessed unique solutions).
When the objective function coincides with one of the half planes generated by the constraints in
the problem, will possess multiple optimal solutions. In this section we are going to discuss about the
multiple optimal solutions of linear programming problem with the help of the following Example 2.7.
Example 2.7:
A company purchasing scrap material has two types of scarp materials available. The first type has 30%
of material X, 20% of material Y and 50% of material Z by weight. The second type has 40% of
material X, 10% of material Y and 30% of material Z. The costs of the two scraps are Rs.120 and
Rs.160 per kg respectively. The company requires at least 240 kg of material X, 100 kg of material Y
and 290 kg of material Z. Find the optimum quantities of the two scraps to be purchased so that the
company requirements of the three materials are satisfied at a minimum cost.
Solution
First we have to formulate the linear programming model. Let us introduce the decision variables x1 and
x2 denoting the amount of scrap material to be purchased. Here the objective is to minimize the
purchasing cost. So, the objective function here is
Minimize
120x1 + 160x2
Subject to:
0.3x1 + 0.4x2 240
0.2x1 + 0.1x2 100
0.5x1 + 0.3x2 290
x1 0; x2 0
Multiply by 10 both sides of the inequalities, then the problem becomes
Minimize
120x1 + 160x2
Subject to:
3x1 + 4x2 2400
2x1 + x2 1000
5x1 + 3x2 2900
x1 0; x2 0
27
MBA-H2040 Quantitative Techniques for Managers
Let us introduce parameter M in the objective function i.e. 120x1 + 160x2 = M. Then we have to
determine the different values for M, which is shown in the following Table 3.
Extreme Point
Coordinates
Objective Function
X1 X2
120x1 + 160x2
A
X1 = 0
X2 = 1000
160000
B
X1 = 150
X2 = 740
136400
C
X1 = 400
X2 = 300
96000
D
X1=800
X2=0
96000
Table 3: Shows the calculation of Minimum objective function value
Note that there are two minimum value for the objective function (M=96000). The feasible
region and the multiple solutions are indicated in the following Graph 5.
1100
A
1000
900
800
B
700
5x1 + 3x2 2900
600
X2
500
2x
1 + x2 1000
400
300
C
200
3x1 + 4x2 2400
100
28
0
D
MBA-H2040 Quantitative Techniques for Managers
X1
Graph 5: Feasible Region, Multiple Optimal Solutions
The extreme points are A, B, C, and D. One of the objective functions 120x1 + 160x2 = M family
coincides with the line CD at the point C with value M=96000, and the optimum value variables are x1 =
400, and x2 = 300. And at the point D with value M=96000, and the optimum value variables are x1 =
800, and x2 = 0.
Thus, every point on the line CD minimizes objective function value and the problem contains
multiple optimal solutions.
2.7 Unbounded Solution
When the feasible region is unbounded, a maximization problem may don't have optimal solution, since
the values of the decision variables may be increased arbitrarily. This is illustrated with the help of the
following problem.
Maximize
3x1 + x2
Subject to:
x1 + x2 6
-x1 + x2 6
-x1 + 2x2 -6
and
x1, x2 0
Graph 6 shows the unbounded feasible region and demonstrates that the objective function can
be made arbitrarily large by increasing the values of x1 and x2 within the unbounded feasible region. In
this case, there is no point (x1, x2) is optimal because there are always other feasible points for which
objective function is larger. Note that it is not the unbounded feasible region alone that precludes an
optimal solution. The minimization of the function subject to the constraints shown in the Graph 6
would be solved at one the extreme point (A or B).
The unbounded solutions typically arise because some real constraints, which represent a
practical resource limitation, have been missed from the linear programming formulation. In such
situation the problem needs to be reformulated and re-solved.
29
MBA-H2040 Quantitative Techniques for Managers
-x1 + x2 = 6
A
6
x1 + x2 = 6
5
4
-x1 + 2x2 = 6
X2
3
2
Graph 6: Unbounded Feasible Region
1
1
2
3
4
5
6
B
X1
2.8 Infeasible Solution
A linear programming problem is said to be infeasible if no feasible solution of the problem exists. This
section describes infeasible solution of the linear programming problem with the help of the following
Example 2.8.
Example 2.8:
30
MBA-H2040 Quantitative Techniques for Managers
Minimize
200x1 + 300x2
Subject to:
0.4x1 + 0.6x2 240
0.2x1 + 0.2x2 80
0.4x1 + 0.3x2 180
x1, x2 0
On multiplying both sides of the inequalities by 10, we get
4x1 + 6x2 2400
2x1 + 2x2 800
4x1 + 3x2 1800
700
A
600
4x1 + 3x2 = 1800
500
B
400
X2
300
200
F
2x1 + 2x2 = 800
100
4x1 + 6x2 = 2400
0
D
C
E
500
600
100
200
300
400
X1
Graph 7: Infeasible Solution
The region right of the boundary AFE includes all the solutions which satisfy the first (4x1 + 6x2
2400) and the third (4x1 + 3x2 1800) constraints. The region left of the BC contains all solutions
which satisfy the second constraint (2x1 + 2x2 800).
31
MBA-H2040 Quantitative Techniques for Managers
Hence, there is no solution satisfying all the three constraints (first, second, and third). Thus, the
linear problem is infeasible. This is illustrated in the above Graph 7.
2.9 Summary
In Operations Research linear programming is a versatile technique with wide applications in various
management problems. Linear Programming problem has a number of characteristics. That is first we
have to identify the decision variable. The problem must have a well defined objective function, which
are expressed in terms of the decision variables.
The objective function may have to be maximized when it indicates the profit or production or
contribution. If the objective function represents cost, in this case the objective function has to be
minimized.
The management problem is expressed in terms of the decision variables with the objective
function and constraints. A linear programming problem is solved graphically if it contains only two
variables.
2.10 Key Terms
Objective Function: is a linear function of the decision variables representing the objective of the
manager/decision maker.
Constraints: are the linear equations or inequalities arising out of practical limitations.
Decision Variables: are some physical quantities whose values indicate the solution.
Feasible Solution: is a solution which satisfies all the constraints (including the non-negative) presents
in the problem.
Feasible Region: is the collection of feasible solutions.
Multiple Solutions: are solutions each of which maximize or minimize the objective function.
Unbounded Solution: is a solution whose objective function is infinite.
Infeasible Solution: means no feasible solution.
2.11 Self Assessment Questions
Q1. A juice company has its products viz. canned apple and bottled juice with profit margin Rs.4 and
Rs.2 respectively pre unit. The following table shows the labour, equipment, and ingredients to produce
each product per unit.
Canned Apple
Bottled Juice
Total
Labour
2.0
3.0
12.0
Equipment
3.2
1.0
8.0
32
MBA-H2040 Quantitative Techniques for Managers
Ingredients
2.4
2.0
9.0
Formulate the linear programming problem (model) specifying the product mix which will maximize the
profit without exceeding the levels of resources.
Q2. An organization is interested in the analysis of two products which can be produces from the idle
time of labour, machine and investment. It was notified on investigation that the labour requirement of
the first and the second products was 4 and 5 units respectively and the total available man hours was
48. Only first product required machine hour utilization of one hour per unit and at present only 10 spare
machine hours are available. Second product needs one unit of byproduct per unit and the daily
availability of the byproduct is 12 units. According to the marketing department the sales potential of
first product cannot exceed 7 units. In a competitive market, first product can be sold at a profit of Rs.6
and the second product at a profit of Rs.10 per unit.
Formulate the problem as a linear programming model. Also determine graphically the feasible
region. Identify the redundant constraints if any.
Q3. Find graphically the feasible region of the linear programming problem given in Q1.
Q4. A bed mart company is in the business of manufacturing beds and pillows. The company has 40
hours for assembly and 32 hours for finishing work per day. Manufacturing of a bed requires 4 hours for
assembly and 2 hours in finishing. Similarly a pillow requires 2 hours for assembly and 4 hours for
finishing. Profitability analysis indicates that every bed would contribute Rs.80, while a pillow
contribution is Rs.55 respectively. Find out the daily production of the company to maximize the
contribution (profit).
Q5. Maximize
1170x1 + 1110x2
Subject to:
9x1 + 5x2 500
7x1 + 9x2 300
5x1 + 3x2 1500
7x1 + 9x2 1900
2x1 + 4x2 1000
x1, x2 0
Find graphically the feasible region and the optimal solution.
Q6. Solve the following LP problem graphically
Minimize
2x1 +1.7x2
Subject to:
0.15x1 + 0.10x2 1.0
0.75x1 + 1.70x2 7.5
1.30x1 + 1.10x2 10.0
x1, x2 0
Q7. Solve the following LP problem graphically
33
MBA-H2040 Quantitative Techniques for Managers
Maximize
2x1 + 3x2
Subject to:
x1 ? x2 1
x1 + x2 3
x1, x2 0
Q8. Graphically solve the following problem of LP
Maximize
3x1 + 2x2
Subject to:
2x1 ? 3x2 0
3x1 + 4x2 -12
x1, x2 0
Q9. Solve the following problem graphically
Maximize
4x1 + 4x2
Subject to:
-2x1 + x2 1
x1 2
x1 + x2 3
x1, x2 0
2.12 Key Solutions
Q1.
Canned Apple x1
Bottled Juice x2
Maximize
4x1 + 2x2
Subject to:
2x1 + 3x2 12
3.2x1 + x2 8
2.4x1 + 2x2 9
x1, x2 0
Q2.
First Product x1
Second Product x2
Maximize
6x1 + 10x2
Subject to:
4x1 + 5x2 48
x1 10
34
MBA-H2040 Quantitative Techniques for Managers
x2 12
x1 7
x1, x2 0
The constraints x1 10 is redundant.
Q4.
Beds = 8
Pillows = 4
Maximum Profits is: Rs.860
Q5.
Optimum variables values are: x1=271.4, x2=0
The maximum value is: 317573
Q6.
Optimum variables values are: x1=6.32, x2=1.63
The minimum values is: 15.4
Q7.
The solution is unbounded
Q8.
The problem has no feasible solution
Q9.
The problem has multiple solutions with the following optimum variable values:
x1=2, x2 =1 or x1=2/3, x2=7/3
The Maximum objective function value is: 12
2.13 Further References
Mittal, K.V. 1976. Optimization Methods in Operations Research and Systems Analysis, Wiley Eastern
Ltd, New Delhi.
Taha, H.A1999. Operations Research An Introduction, PHI Ltd., New Delhi.
Richard I.Levin, David S. Rubin, Joel P. Stinson, Everette S.Gardner, Jr.1992. Quantitative Approaches
to Management, McGraw Hill, NJ.
35
MBA-H2040 Quantitative Techniques for Managers
UNIT I
ON
S
3 LINEAR PROGRAMMING ? SIMPLEX METHOD
S
E
L
LESSON STRUCTURE
3.1 Introduction
3.2 Basics of Simplex Method
3.3 Simplex Method Computation
3.4 Simplex Method with More Than Two
Variables
3.5 Two Phase and M Method
3.5.1 Two Phase Method
3.5.2 M Method
3.6 Multiple Solutions
3.7 Unbounded Solution
3.8 Infeasible Solution
3.9 Summary
3.10
Key Terms
3.11
Self Assessment Questions
3.12
Key Solutions
3.13
3.13 Further References
36
MBA-H2040 Quantitative Techniques for Managers
Objectives
After Studying this lesson, you should be able
to:
Understand the basics of simplex method
Explain the simplex calculations
Describe various solutions of Simplex
Method
Understand two phase and M method
37
MBA-H2040 Quantitative Techniques for Managers
3.1 Introduction
The Linear Programming with two variables can be solved graphically. The graphical method of solving
linear programming problem is of limited application in the business problems as the number of
variables is substantially large. If the linear programming problem has larger number of variables, the
suitable method for solving is Simplex Method. The simplex method is an iterative process, through
which it reaches ultimately to the minimum or maximum value of the objective function.
The simplex method also helps the decision maker/manager to identify the following:
Redundant Constraints
Multiple Solutions
Unbounded Solution
Infeasible Problem
3.2 Basics of Simplex Method
The basic of simplex method is explained with the following linear programming problem.
Example 3.1:
Maximize
60x1 + 70x2
Subject to:
2x1 + x2 300
3x1 + 4x2 509
4x1 + 7x2 812
x1, x2 0
Solution
First we introduce the variables
s3, s4, s5 0
So that the constraints becomes equations, thus
2x1 + x2 + s3 = 300
3x1 + 4x2 + s4 = 509
4x1 + 7x2 + s5 = 812
Corresponding to the three constraints, the variables s3, s4, s5 are called as slack variables. Now, the
system of equation has three equations and five variables.
There are two types of solutions they are basic and basic feasible, which are discussed as follows:
Basic Solution
We may equate any two variables to zero in the above system of equations, and then the system will
have three variables. Thus, if this system of three equations with three variables is solvable such a
solution is called as basic solution.
38
MBA-H2040 Quantitative Techniques for Managers
For example suppose we take x1=0 and x2=0, the solution of the system with remaining three
variables is s3=300, s4=509 and s5=812, this is a basic solution and the variables s3, s4, and s5 are known
as basic variables where as the variables x1, x2 are known as non-basic variables.
The number of basic solution of a linear programming problem is depends on the presence of the
number of constraints and variables. For example if the number of constraints is m and the number of
variables including the slack variables is n then there are at most nCn-m = nCm basic solutions.
Basic Feasible Solution
A basic solution of a linear programming problem is called as basic feasible solutions if it is feasible it
means all the variables are non-negative. The solution s3=300, s4=509 and s5=812 is a basic feasible
solution.
The number of basic feasible solution of a linear programming problem is depends on the
presence of the number of constraints and variables. For example if the number of constraints is m and
the number of variables including the slack variables is n then there are at most nCn-m = nCm basic
feasible solutions.
Every basic feasible solution is an extreme point of the convex set of feasible solutions and every
extreme point is a basic feasible solution of the set of given constraints. It is impossible to identify the
extreme points geometrically if the problem has several variables but the extreme points can be
identified using basic feasible solutions. Since one the basic feasible solution will maximize or minimize
the objective function, the searching of extreme points can be carry out starting from one basic feasible
solution to another.
The Simplex Method provides a systematic search so that the objective function increases in the
cases of maximization progressively until the basic feasible solution has been identified where the
objective function is maximized.
3.3 Simplex Method Computation
This section describes the computational aspect of simplex method. Consider the following linear
programming problem
Maximize
60x1 + 70x2
Subject to:
2x1 + x2 + s3 = 300
3x1 + 4x2 + s4 = 509
4x1 + 7x2 + s5 = 812
x1, x2, s3, s4 ,s5 0
39
MBA-H2040 Quantitative Techniques for Managers
The profit Z=60x1 + 70x2 i.e. Maximize 60x1 + 70x2
The standard form can be summarized in a compact table form as
In this problem the slack variables s3, s4, and s5 provide a basic feasible solution from which the
simplex computation starts. That is s3==300, s4=509 and s5=812. This result follows because of the
special structure of the columns associated with the slacks.
If z represents profit then z=0 corresponding to this basic feasible solution. We represent by CB
the coefficient of the basic variables in the objective function and by XB the numerical values of the
basic variable.
So that the numerical values of the basic variables are: XB1=300, XB2=509, XB3=812. The profit
z=60x1+70x2 can also expressed as z-60x1-70x2=0. The simplex computation starts with the first
compact standard simplex table as given below:
CB
Basic
Cj
60 70 0 0 0
Variables XB
x1 x2 s3 s4 s5
0
s3
300
2 1 1 0 0
0
s4
509
3 4 0 1 0
0
s5
812
4 7 0 0 1
z
-60 -70 0 0 0
Table 1
In the objective function the coefficients of the variables are CB1=CB2=CB3=0. The topmost row
of the Table 1 denotes the coefficient of the variables x1, x2, s3, s4, s5 of the objective function
respectively. The column under x1 indicates the coefficient of x1 in the three equations respectively.
Similarly the remaining column also formed.
On seeing the equation z=60x1+70x2 we may observe that if either x1 or x2, which is currently
non-basic is included as a basic variable so that the profit will increase. Since the coefficient of x2 is
higher we choose x2 to be included as a basic variable in the next iteration. An equivalent criterion of
choosing a new basic variable can be obtained the last row of Table 1 i.e. corresponding to z.
Since the entry corresponding to x2 is smaller between the two negative values, x2 will be
included as a basic variable in the next iteration. However with three constraints there can be only three
basic variables.
Thus, by bringing x2 a basic variable one of the existing basic variables becomes non-basic. The
question here is How to identify this variable? The following statements give the solution to this
question.
Consider the first equation i.e. 2x1 + x2 + s3 = 300
40
MBA-H2040 Quantitative Techniques for Managers
From this equation
2x1+s3=300-x2
But x1=0. Hence, in order that s30
300-x20
i.e. x2300
Similarly consider the second equation i.e. 3x1 + 4x2 + s4 = 509
From this equation
3x1+s4=509-4x2
But, x1=0. Hence, in order that s40
509-4x20
i.e. x2509/9
Similarly consider the third equation i.e. 4x1 + 7x2 + s5 = 812
From this equation
4x1+s5=812-7x2
But x1=0. Hence, in order that s50
812-7x20
i.e. x2812/7
Therefore the three equation lead to
x2300,
x2509/9,
x2812/7
Thus x2=Min (x2300, x2509/9, x2812/7) it means
x2=Min (x2300/1, x2509/9, x2812/7)=116
Therefore x2=116
If x2=116, you may be note from the third equation
7x2+s5=812
i.e. s5=0
Thus, the variable s5 becomes non-basic in the next iteration.
So that the revised values of the other two basic variables are
S3=300-x2=184
S4=509-4*116=45
Refer to Table 1, we obtain the elements of the next Table i.e. Table 2 using the following rules:
1. We allocate the quantities which are negative in the z-row. Suppose if all the quantities are
positive, the inclusion of any non-basic variable will not increase the value of the objective
function. Hence the present solution maximizes the objective function. If there are more than one
negative values we choose the variable as a basic variable corresponding to which the z value is
least as this is likely to increase the more profit.
41
MBA-H2040 Quantitative Techniques for Managers
2. Let xj be the incoming basic variable and the corresponding elements of the jth row column be
denoted by Y1j, Y2j and Y3j respectively. If the present values of the basic variables are XB1,
XB2 and XB3 respectively, then we can compute.
Min [XB1/Y1j, XB2/Y2j, XB3/Y3j] for Y1j, Y2j, Y3j>0.
Note that if any Yij0, this need not be included in the comparison. If the minimum occurs
corresponding to XBr/Yrj then the rth basic variable will become non-basic in the next iteration.
3. Using the following rules the Table 2 is computed from the Table 1.
i.
The revised basic variables are s3, s4 and x2. Accordingly, we make CB1=0, CB2=0
and CB3=70.
ii.
As x2 is the incoming basic variable we make the coefficient of x2 one by dividing
each element of row-3 by 7. Thus the numerical value of the element corresponding
to x1 is 4/7, corresponding to s5 is 1/7 in Table 2.
iii.
The incoming basic variable should appear only in the third row. So we multiply the
third-row of Table 2 by 1 and subtract it from the first-row of Table 1 element by
element. Thus the element corresponding to x2 in the first-row of Table 2 is 0.
Therefore the element corresponding to x1 is
2-1*4/7=10/7 and the element corresponding to s5 is
0-1*1/7=-1/7
In this way we obtain the elements of the first and the second row in Table 2. In Table 2 the
numerical values can also be calculated in a similar way.
CB
Basic
Cj
60 70 0 0 0
Variables XB
x1 x2 s3 s4 s5
0
s3
184
10/7 0 1 0 -1/7
0
s4
45
5/7 0 0 1 -4/7
70
x2
116
4/7 1 0 0 1/7
zj-cj
-140/7 0 0 0 70/7
Table 2
Let CB1, CB2, Cb3 be the coefficients of the basic variables in the objective function. For
example in Table 2 CB1=0, CB2=0 and CB3=70. Suppose corresponding to a variable J, the quantity zj is
defined as zj=CB1, Y1+CB2, Y2j+CB3Y3j. Then the z-row can also be represented as Zj-Cj.
For example:
z1 - c1 = 10/7*0+5/7*0+70*4/7-60 = -140/7
z5 ? c5 = -1/7*0-4/7*0+1/7*70-0 = 70/7
1. Now we apply rule (1) to Table 2. Here the only negative zj-cj is z1-c1 = -140/7
Hence x1 should become a basic variable at the next iteration.
42
MBA-H2040 Quantitative Techniques for Managers
2. We compute the minimum of the ratio
184 , 45, 116
644 , 63 , 203
Min 10 5 4
= Min 5
= 63
7 7 7
This minimum occurs corresponding to s4, it becomes a non basic variable in next iteration.
3. Like Table 2, the Table 3 is computed sing the rules (i), (ii), (iii) as described above.
CB
Basic
Cj
60 70 0 0 0
Variables XB
x1 x2 s3 s4 s5
0
s3
94
0 0 1 -2 1
60
x1
63
1 0 0 7/5 -4/5
70
x2
80
0 1 0 -4/5 3/5
zj-cj
0 0 0 28 -6
Table 3
1. z5 ? c5 < 0 should be made a basic variable in the next iteration.
2. Now compute the minimum ratios
94, 80
Min 1 3
= 94
5
Note: Since y25 = -4/5 < 0, the corresponding ratio is not taken for comparison.
The variable s3 becomes non basic in the next iteration.
3. From the Table 3, Table 4 is calculated following the usual steps.
CB
Basic
Cj
60 70 0 0 0
Variables XB
x1 x2 s3 s4 s5
0
s5
94
0 0 1 -2 1
60
x1
691/5
1 0 4/5 -1/5 0
70
x2
118/5
0 1 -3/5 2/5 0
zj-cj
0 0 6 16 0
43
MBA-H2040 Quantitative Techniques for Managers
Note that zj ? cj 0 for al j, so that the objective function can't be improved any further.
Thus, the objective function is maximized for x1 = 691/5 and x2=118/5 and
The maximum value of the objective function is 9944.
3.4
Simplex Method with More Than Two Variables
In previous section we discussed the simplex method of linear programming problem with two decision
variables. The simplex method computational procedure can be readily extended to linear programming
problems with more than two variables. This is illustrated in this section with the help of the product
mix problem given in the following Example 3.2.
Example 3.2
An organization has three machine shops viz. A, B and C and it produces three product viz. X, Y and Z
using these three machine shops. Each product involves the operation of the machine shops. The time
available at the machine shops A, B and C are 100, 72 and 80 hours respectively. The profit per unit of
product X, Y and Z is $22, $6
and $2 respectively. The
following table shows the time
required for each operation for
10
7
2
unit amount of each product.
Determine
an
appropriate
product mix so as to maximize
the profit.
2
3
4
Machine
1
2
1
Shops
A B C
Products
Profit/unit
X $22
Y $6
Z $2
Available Hours 100 72 80
Solution
First we have to develop linear programming formulation. The linear programming formulation of the
product mix problem is:
Maximize
22x1 + 6x2 + 2x3
Subject to:
10x1 + 2x2 + x3 100
7x1 + 3x2 + 2x3 72
2x1 + 4x2 + x3 80
x1, x2, x3 0
44
MBA-H2040 Quantitative Techniques for Managers
We introduce slack variables s4, s5 and s6 to make the inequalities equation.
Thus, the problem can be stated as
Maximize
22x1 + 6x2 + 2x3
Subject to:
10x1 + 2x2 + x3 + s4 = 100
7x1 + 3x2 + 2x3 + s5 = 72
2x1 + 4x2 + x3 + s6 = 80
x1, x2, x3, s4, s5, s6 0
From the above equation the simplex Table 1 can be obtained in a straight forward manner. Here
the basic variables are s4, s5 and s6. Therefore CB1 = CB2 = CB3 = 0.
CB
Basic
Cj
22
6
2
0
0
0
Variable XB
x1
x2
x3
s4
s5
s6
0
s4
100
10
2
1
1
0
0
0
s5
72
7
3
2
0
1
0
0
s6
80
2
4
1
0
0
1
zj-cj
-22
-6
-2
0
0
0
Table 1
1. z1-c1 = -22 is the smallest negative value. Hence x1 should be taken as a basic variable in the next
iteration.
2. Calculate the minimum of the ratios
Min 100 , 72 , 80 = 10
10 7 2
The variable s4 corresponding to which minimum occurs is made a non basic variable.
3. From the Table 1, the Table 2 is calculated using the following rules:
i.
The revised basic variables are x1, s5, s6. Accordingly we make CB1=22, CB2=0 and
CB3=0.
ii.
Since x1 is the incoming variable we make x1 coefficient one by dividing each
element of row 1 by 10. Thus the numerical value of the element corresponding to x2
is 2/10, corresponding to x3 is 1/10, corresponding to s4 is 1/10, corresponding to s5 is
0/10 and corresponding to s6 is 0/10 in Table 2.
iii.
The incoming basic variable should only appear in the first row. So we multiply first
row of Table 2 by 7 and subtract if from the second row of Table 1 element by
element.
Thus,
The element corresponding to x1 in the second row of Table 2 is zero
The element corresponding to x2 is 3 ? 7 * 2 = 16
10 10
By using this way we get the elements of the second and the third row in Table 2.
45
MBA-H2040 Quantitative Techniques for Managers
Similarly, the calculation of numerical values of basic variables in Table 2 is done.
CB
Basic
Cj
22
6
2
0
0
0
Variable XB
x1
x2
x3
s4
s5
s6
22
x1
10
1
2/10
1/10
1/10
0
0
0
s5
2
0
16/10 13/10 -7/10
1
0
0
s6
60
0
18/5
4/5
-1/5
0
1
zj-cj
0
-8/5
1/5
12/5
0
0
Table 2
1. z2-c2 = -8/5. So x2 becomes a basic variable in the next iteration.
2. Calculate the minimum of the ratios
10, 7 , 60
Min 2 16 18 = Min 50, 70, 300 = 70
10 10 5 16 18 16
Hence the variable s5 will be a non basic variable in the next iteration.
3. From Table 2, the Table 3 is calculated using the rules (i), (ii) and (iii) mentioned above.
CB
Basic
Cj
22
6
2
0
0
0
Variable XB
x1
x2
x3
s4
s5
s6
22
x1
73/8
1
0
-1/16
3/16
-1/8
0
6
x2
30/8
0
1
13/16 -7/16
5/8
0
0
s6
177/4 0
0
-17/8
11/8
-9/4
1
zj-cj
0
0
24/16 24/16 1
0
Table 3
Note that all zj ? cj 0, so that the solution is x1 = 73/8, x2 = 30/8 and s6 = 177/4 maximizes the
objective function.
The Maximum Profit is: 22*73/8 + 6*30/8 = 1606/8 + 180/8
= 1786/8 = $223.25
3.5 Tow Phase and M-Method
In the last two section we discussed the simplex method was applied to linear programming problems
with less than or equal to () type constraints. Thus, there we could introduce slack variables which
provide an initial basic feasible solution of the problem.
46
MBA-H2040 Quantitative Techniques for Managers
Generally, the linear programming problem can also be characterized by the presence of both
`less than or equal to' type or `greater than or equal to ()' type constraints. In such case it is not always
possible to obtain an initial basic feasible solution using slack variables.
The greater than or equal to type of linear programming problem can be solved by using the
following methods:
1. Two Phase Method
2. M Method
In this section we will discuss these two methods.
3.5.1 Two Phase Method
We discuss the Two Phase Method with the help of the following Example 3.3.
Example 3.3
Minimize
12.5x1 + 14.5x2
Subject to:
x1 + x2 2000
0.4x1 + 0.75x2 1000
0.075x1 + 0.1x2 200
x1, x2 0
Solution
Here the objective function is to be minimized; the values of x1 and x2 which minimized this objective
function are also the values which maximize the revised objective function i.e.
Maximize
-12.5x1 ? 14.5x2
We can multiply the second and the third constraints by 100 and 1000 respectively for the
convenience of calculation.
Thus, the revised linear programming problem is:
Maximize
-12.5x1 ? 14.5x2
Subject to:
x1 + x2 2000
40x1 + 75x2 100000
75x1 + 100x2 200000
x1, x2 0
Now we convert the two inequalities by introducing surplus variables s3 and s4 respectively.
The third constraint is changed into an equation by introducing a slack variable s5.
Thus, the linear programming problem becomes as
47
MBA-H2040 Quantitative Techniques for Managers
Maximize
-12.5x1 ? 14.5x2 = -25/2x1 ? 29/2x2
Subject to:
x1 + x2 -s3
= 2000
40x1 + 75x2 -s4 = 100000
75x1 + 100x2 +s5 = 200000
x1, x2,s3,s4,s5 0
Even though the surplus variables can convert greater than or equal to type constraints into
equations they are unable to provide initial basic variables to start the simplex method calculation. So
we may have to introduce two more additional variables a6 and a7 called as artificial variable to
facilitate the calculation of an initial basic feasible solution.
In this method the calculation is carried out in tow phases hence tow phase method.
Phase I
In this phase we will consider the following linear programming problem
Maximize
-a6 ?a7
Subject to:
x1 + x2 -s3 +a6 = 2000
40x1 + 75x2 -s4 + a7 = 100000
75x1 + 100x2 +s5 = 200000
x1, x2.s3,s4,s5,a6,a7 0
The initial basic feasible solution of the problem is
a6 = 2000, a7=100000 and s5 = 200000.
As the minimum value of the objective function of the Phase I is zero at the end of the Phase I
calculation both a6 and a7 become zero.
CB
Basic
C
0
0
0
0
0
-1
-1
j
variables XB
x
x
s
s
s
a
a
1
2
3
4
5
6
7
-1
a
2000
1
1
-1
0
0
1
0
6
-1
a
100000 40
75
0
-1
0
0
1
7
0
s
200000 75
100
0
0
1
0
0
5
z
j-cj
-41
-76
1
1
0
0
0
Table 1
Here x2 becomes a basic variable and a7 becomes non basic variable in the next iteration. It is no
longer considered for re-entry in the table.
48
MBA-H2040 Quantitative Techniques for Managers
CB
Basic
Cj
0
0
0
0
0
-1
variables XB
x1
x2
s3
s4
s5
a6
-1
a6
2000/3
7/15
0
-1
1/75
0
1
0
x2
4000/3
8/15
1
0
-1/75
0
0
0
s5
200000/3 65/3
0
0
4/3
1
0
zj-cj
-1/15
0
1
-1/75
0
0
Table 2
Then x1 becomes a basic variable and a6 becomes a non basic variable in the next iteration.
CB
Basic
Cj
0
0
0
0
0
variables XB
x1
x2
s3
s4
s5
0
x1
10000/7
1
0
-15/7
1/35
0
0
x2
4000/7
0
1
8/7
-1/35
0
0
s5
250000/7 0
0
325/7
16/21
1
zj-cj
0
0
0
0
0
Table 3
The calculation of Phase I end at this stage. Note that, both the artificial variable have been
removed and also found a basic feasible solution of the problem.
The basic feasible solution is:
x1 = 10000/7, x2 = 4000/2, s5 = 250000/7.
Phase II
The initial basic feasible solution obtained at the end of the Phase I calculation is used as the initial basic
feasible solution of the problem. In this Phase II calculation the original objective function is introduced
and the usual simplex procedure is applied to solve the linear programming problem.
49
MBA-H2040 Quantitative Techniques for Managers
CB
Basic
Cj
-25/2
-29/2
0
0
0
variables XB
x1
x2
s3
s4
s5
-25/2
x1
10000/7
1
0
-15/7
1/35
0
-29/2
x2
4000/7
0
1
8/7
-1/35
0
0
s5
250000/7 0
0
325/7
5/7
1
zj-cj
0
0
143/14 2/35
0
Table 1
In this Table 1 all zj-cj 0 the current solution maximizes the revised objective function.
Thus, the solution of the problem is:
x1 = 10000/7 = 1428 and x2 = 4000/7 = 571.4 and
The Minimum Value of the objective function is: 26135.3
3.5.2 M Method
In this method also we need artificial variables for determining the initial basic feasible solution. The M
method is explained in the next Example 3.4 with the help of the previous Example 3.3.
Example 3.4
Maximize
-12.5x1 ? 14.5x2
Subject to:
x1 + x2 ?s3 = 2000
40x1 + 75x2 -s4 = 100000
75x1 + 100x2 +s5 = 200000
x1, x2, s3, s4, s5 0.
Introduce the artificial variables a6 and a7 in order to provide basic feasible solution in the second and
third constraints. The objective function is revised using a large positive number say M.
Thus, instead of the original problem, consider the following problem i.e.
Maximize
-12.5x1 ? 14.5x2 ? M (a6 + a7)
Subject to:
x1 + x2 ?s3 + a6 = 2000
40x1 + 75x2 -s4 +a7 = 100000
75x1 + 100x2 +s5 = 200000
x1, x2, s3, s4, s5, a6, a7 0.
50
MBA-H2040 Quantitative Techniques for Managers
The coefficient of a6 and a7 are large negative number in the objective function. Since the objective
function is to be maximized in the optimum solution, the artificial variables will be zero. Therefore, the
basic variable of the optimum solution are variable other than the artificial variables and hence is a basic
feasible solution of the original problem.
The successive calculation of simplex tables is as follows:
CB
Basic
Cj
-12.5
-14.5 0
0
0
-M
-M
variables XB
x1
x2
s3
s4
s5
a6
a7
-M
a6
2000
1
1
-1
0
0
1
0
-M
a7
100000 40
75
0
-1
0
0
1
0
s5
200000 75
100
0
0
1
0
0
zj-cj
-41M
-76M M
M
0
0
0
+12.5
+14.5
Table 1
Since M is a large positive number, the coefficient of M in the zj ? cj row would decide the
entering basic variable. As -76M < -41M, x2 becomes a basic variable in the next iteration replacing a7.
The artificial variable a7 can't be re-entering as basic variable.
CB
Basic
Cj
-12.5
-14.5 0
0
0
-M
variables XB
x1
x2
s3
s4
s5
a6
-M
a6
2000/3
7/15
0
-1
1/75
0
1
-14.5
x2
4000/3
8/15
1
0
-1/75
0
0
0
s5
200000/3 65/3
0
0
4/3
1
0
zj-cj
-7/15M
0
M
-M/75
0
0
+143/30
+29/150
Table 2
Now x1 becomes a basic variable replacing a6. Like a7 the variable a6 also artificial variable so it
can't be re-entering in the table.
51
MBA-H2040 Quantitative Techniques for Managers
CB
Basic
Cj
-12.5
-14.5 0
0
0
variables XB
x1
x2
s3
s4
s5
-12.5
x1
10000/7
1
0
-15/7
1/35
0
-14.5
x2
4000/7
0
1
8/7
-1/35
0
0
s5
250000/7 0
0
325/7
16/21
1
zj-cj
0
0
143/14 2/35
0
Table 3
Hence
The optimum solution of the problem is x1 = 10000/7, x2 = 4000/7 and
The Minimum Value of the Objective Function is: 26135.3
3.6 Multiple Solutions
The simplex method also helps in identifying multiple solutions of a linear programming problem. This
is explained with the help of the following Example 3.5.
Example 3.5
Consider the following linear programming problem.
Maximize
2000x1 + 3000x2
Subject to:
6x1 + 9x2 100
2x1 + x2 20
x1, x2 0.
Solution
Introduce the slack variables s3 and s4, so that the inequalities can be converted in to equation as
follows:
6x1 + 9x2 + s3 = 100
2x1 + x2 + s4 = 20
x1, x2, s3, s4 0.
The computation of simple procedure and tables are as follows:
52
MBA-H2040 Quantitative Techniques for Managers
CB
Basic
Cj
2000
3000
0
0
variables XB
x1
x2
s3
s4
0
s3
100
6
9
1
0
0
s4
20
2
1
0
1
zj-cj
-2000 -3000 0
0
Table 1
CB
Basic
Cj
2000
3000
0
0
variables XB
x1
x2
s3
s4
0
x2
100/9
2/3
1
1/9
0
0
s4
80/9
4/3
0
-1/9
1
zj-cj
0
0
3000/9 0
Table 2
Here zj-cj 0 for al the variables so that we can't improve the simplex table any more. Hence it
is optimum.
The optimum solution is x1 = 0, x2 = 100/9 and
The maximum value of the objective function is: 100000/3 = 33333.33.
However, the zj-cj value corresponding to the non basic variable x1 is also zero. This indicates
that there is more than one optimum solution for the problem exists.
In order to calculate the value of the alternate optimum solution we have to introduce x1 as a
basic variable replacing s4. The next Table 3 shows the computation of this.
53
MBA-H2040 Quantitative Techniques for Managers
CB
Basic
C
2000
3000
0
0
j
variables XB
x
x
s
s
1
2
3
4
3000
x
20/3
0
1
1/6
1/2
2
2000
x
20/3
1
0
-1/12
3/4
1
z
j-cj
0
0
1000/3 3000
Table 3
Thus,
x1 = 20/3, x2 = 20/3 also maximize the objective function and
The Maximum value of the objective function is: 100000/3 = 33333.33
Thus, the problem has multiple solutions.
3.7 Unbounded Solution
In this section we will discuss how the simplex method is used to identify the unbounded solution. This
is explained with the help of the following Example 3.6.
Example 3.6
Consider the following linear programming problem.
Maximize
5x1 + 4x2
Subject to:
x1 ? x2 8
x1 7
x1, x2 0.
Solution :
Introduce the slack variables s3 and s4, so that the inequalities becomes as equation as follows:
x1 + s3 = 7
x1 ? x2 + s4 = 8
x1, x2, s3, s4 0.
The calculation of simplex procedures and tables are as follows:
54
MBA-H2040 Quantitative Techniques for Managers
CB
Basic
C
5
4
0
0
j
variables XB
x
x
s
s
1
2
3
4
0
s
7
1
0
1
0
3
0
s
8
1
-1
0
1
4
z
j-cj
-5
-4
0
0
Table 1
CB
Basic
C
5
4
0
0
j
variables XB
x
x
s
s
1
2
3
4
5
x
7
1
0
1
0
1
0
s
1
0
-1
-1
1
4
z
j-cj
0
-4
5
0
Table 2
Note that z2-c2 < 0 which indicates x2 should be introduced as a basic variable in the next
iteration. However, both y120, y220.
Thus, it is not possible to proceed with the simplex method of calculation any further as we
cannot decide which variable will be non basic at the next iteration. This is the criterion for unbounded
solution.
NOTE: If in the course of simplex computation zj-cj < 0 but yij 0 for al i then the problem has no
finite solution.
But in this case we may observe that the variable x2 is unconstrained and can be increased
arbitrarily. This is why the solution is unbounded.
3.8 Infeasible Solution
This section illustrates how to identify the infeasible solution using simplex method. This is explained
with the help of the following Example 3.7.
Example 3.7
55
MBA-H2040 Quantitative Techniques for Managers
Consider the following problem.
Minimize
200x1 + 300x2
Subject to:
2x1 + 3x2 1200
x1 + x2 400
2x1 + 3/2x2 900
x1, x2 0
Solution
Since it is a minimization problem we have to convert it into maximization problem and introduce the
slack, surplus and artificial variables. The problem appears in the following manner after doing all these
procedure.
Maximize
-200x1 - 300x2
Subject to:
2x1 + 3x2 -s3 +a6 = 1200
x1 + x2 +s4 = 400
2x1 + 3/2x2 - s5 + a7 = 900
x1, x2, s3, s4, s5, a6, a7 0
Here the a6 and a7 are artificial variables. We use two phase method to solve this problem.
Phase I
Maximize
-a6 ?a7
Subject to:
2x1 + 3x2 -s3 +a6 = 1200
x1 + x2 +s4 = 400
2x1 + 3/2x2 - s5 + a7 = 900
x1, x2, s3, s4, s5, a6, a7 0
The calculation of simplex procedures and tables are as follows:
CB
Basic
C
0
0
0
0
0
-1
-1
j
variables XB
x
x
s
s
s
a
a
1
2
3
4
5
6
7
-1
a
1200
2
3
-1
0
0
1
0
6
0
s
400
1
1
0
1
0
0
1
4
-1
a
900
2
3/2
0
0
-1
0
0
7
z
j-cj
-4
-9/2
1
0
1
0
0
Table 1
56
MBA-H2040 Quantitative Techniques for Managers
CB
Basic
C
0
0
0
0
0
-1
j
variables XB
x
x
s
s
s
a
1
2
3
4
5
7
0
x
400
2/3
1
-1/3
0
0
0
2
0
s
0
1/3
0
1/3
1
0
0
4
-1
a
300
1
0
1/2
0
-1
1
7
z
j-cj
-1
0
-1/2
0
1
0
Table 2
CB
Basic
C
0
0
0
0
0
-1
j
variables XB
x
x
s
s
s
a
1
2
3
4
5
7
0
x
400
0
1
-1
-2
0
0
2
0
x
0
1
0
1
3
0
0
1
-1
a
300
0
0
-1/2
-3
-1
1
7
z
j-cj
0
0
1/2
3
1
0
Table 3
Note that zj-cj 0 for al the variables but the artificial variable a7 is still a basic variable. This
situation indicates that the problem has no feasible solution.
3.9 Summary
The simplex method is very useful and appropriate method for solving linear programming problem
having more than tow variables. The slack variables are introduced for less that or equal to type, surplus
variables are introduce for greater than or equal to type of linear programming problem. The basic
feasible solution is important in order to solve the problem using the simplex method.
A basic feasible solution of a system with m-equations and n-variables has m non-negative
variables called as basic variables and n-m variables with value zero known as non-basic variables. The
objective function is maximized or minimized at one of the basic feasible solutions.
Surplus variables can't provide the basic feasible solution instead artificial variables are used to
get the basic feasible solutions and it initiate the simplex procedure. Two phase and M-Method are
available to solve linear programming problem in these case.
The simplex method also used to identify the multiple, unbounded and infeasible solutions.
3.10 Key Terms
Basic Variable: Variable of a basic feasible solution has n non-negative value.
Non Basic Variable: Variable of a feasible solution has a value equal to zero.
57
MBA-H2040 Quantitative Techniques for Managers
Artificial Variable: A non-negative variable introduced to provide basic feasible solution and initiate
the simplex procedures.
Slack Variable: A variable corresponding to a type constraint is a non-negative variable introduced to
convert the inequalities into equations.
Surplus Variable: A variable corresponding to a type constraint is a non-negative variable introduced
to convert the constraint into equations.
Basic Solution: System of m-equation and n-variables i.e. m<n is a solution where at least n-m
variables are zero.
Basic Feasible Solution: System of m-equation and n-variables i.e. m<n is a solution where m variables
are non-negative and n-m variables are zero.
Optimum Solution: A solution where the objective function is minimized or maximized.
3.11 Self Assessment Questions
Q1. A soft drinks company has a two products viz. Coco-cola and Pepsi with profit of $2 an $1 per unit.
The following table illustrates the labour, equipment and materials to produce per unit of each product.
Determine suitable product mix which maximizes the profit using simplex method.
Pepsi Coco-cola Total Resources
Labour
12
3
2
1
2.3
Equipment
6.9
1
1.4
Material
4.9
Q2. A factory produces three using three types of ingredients viz. A, B and C in different proportions.
The following table shows the requirements o various ingredients as inputs per kg of the products.
Ingredients
Products
A
B
C
1
4
8
8
2
4
6
4
3
8
4
0
The three profits coefficients are 20, 20 and 30 respectively. The factory has 800 kg of ingredients A,
1800 kg of ingredients B and 500 kg of ingredient C.
Determine the product mix which will maximize the profit and also find out maximum profit.
58
MBA-H2040 Quantitative Techniques for Managers
Q3. Solve the following linear programming problem using two phase and M method.
Maximize
12x1 + 15x2 + 9x3
Subject to:
8x1 + 16x2 + 12x3 250
4x1 + 8x2 + 10x3 80
7x1 + 9x2 + 8x3 =105
x1, x2, x3 0
Q4. Solve the following linear programming problem using simplex method.
Maximize
3x1 + 2x2
Subject to:
x1 ?x2 1
x1 + x2 3
x1, x2 0
Q5. Solve the following linear programming problem using simplex method.
Maximize
x1 + x2
Subject to:
-2x1 + x2 1
x1 2
x1 + x2 3
x1, x2, x3 0
Q6. Maximize
P = 3x1 + 4x2 + x3
Subject to:
x1 + 2x2 + x3 6
2x1 +2x3 4
3x1 + x2 + x3 9
x1, x2, x3 0
3.12 Key Solutions
Q1. Coco-Cola = 20/9, Pepsi = 161/90
Maximum Profit = $6.23
Q2. x1 = 0, x2 = 125, x3 = 75/2
Maximum Profit = 5375
Q3. x1 = 6, x2 = 7, x3 = 0
Maximum Profit = 177
Q4. Unbounded Solution
Q5. x1 = 2, x2 = 1 or x1 = 2/3, x2 = 7/3
59
MBA-H2040 Quantitative Techniques for Managers
Maximum Profit = 3.
Q6. x1 = 2, x2 = 2, x3 = 0
Maximum P = 14
3.13. Further References
Hamdy A Taha, 1999. Introduction to Operations Research, PHI Limited, New Delhi.
Mustafi, C.K. 1988. Operations Research Methods and Practices, Wiley Eastern Limited, New Delhi.
Levin, R and Kirkpatrick, C.A. 1978. Quantitative Approached to Management, Tata McGraw Hill,
Kogakusha Ltd., International Student Edition.
Peterson R and Silver, E. A. 1979. Decision Systems for Inventory Management and Production
Planning.
Handley, G and T.N. Whitin. 1983. Analysis of Inventory Systems, PHI.
Starr, M.K. and D.W. Miller. 1977. Inventory Control Theory and Practice, PHI.
UNIT I
60
MBA-H2040 Quantitative Techniques for Managers
ON
S
4 DUAL LINEAR PROGRAMMING PROBLEMS
S
E
L
LESSON STRUCTURE
4.1 Introduction
4.2 Dual Problem Formulation
4.3 Dual Problem Properties
4.4 Simple Way of Solving Dual Problem
4.5 Summary
4.6 Key Terms
4.7 Self Assessment Questions
4.8 Key Solutions
4.9 Further References
Objectives
After Studying this lesson, you should be able
to:
Understand the Dual Linear programming
Problem
Formulate a Dual Problem
Solve the Dual Linear Programming
Problem
Understand the Properties of a Dual
Problem
61
MBA-H2040 Quantitative Techniques for Managers
4.1 Introduction
For every linear programming problem there is a corresponding linear programming problem called the
dual. If the original problem is a maximization problem then the dual problem is minimization problem
and if the original problem is a minimization problem then the dual problem is maximization problem.
In either case the final table of the dual problem will contain both the solution to the dual problem and
the solution to the original problem.
The solution of the dual problem is readily obtained from the original problem solution if the
simplex method is used.
The formulation of the dual problem also sometimes referred as the concept of duality is helpful
for the understanding of the linear programming. The variable of the dual problem is known as the dual
variables or shadow price of the various resources. The dual problem is easier to solve than the original
problem. The dual problem solution leads to the solution of the original problem and thus efficient
computational techniques can be developed through the concept of duality. Finally, in the competitive
strategy problem solution of both the original and dual problem is necessary to understand the complete
problem.
4.2 Dual Problem Formulation
If the original problem is in the standard form then the dual problem can be formulated using the
following rules:
The number of constraints in the original problem is equal to the number of dual variables. The
number of constraints in the dual problem is equal to the number of variables in the original
problem.
The original problem profit coefficients appear on the right hand side of the dual problem
constraints.
If the original problem is a maximization problem then the dual problem is a minimization
problem. Similarly, if the original problem is a minimization problem then the dual problem is a
maximization problem.
The original problem has less than or equal to () type of constraints while the dual problem has
greater than or equal to () type constraints.
The coefficients of the constraints of the original problem which appear from left to right are
placed from top to bottom in the constraints of the dual problem and vice versa.
The Dual Linear Programming Problem is explained with the help of the following Example 4.1.
Example 4.1
Consider the following product mix problem:
62
MBA-H2040 Quantitative Techniques for Managers
Three machine shops A, B, C produces three types of products X, Y, Z respectively. Each product
involves operation of each of the machine shops. The time required for each operation on various
products is given as follows:
Machine Shops
Products
A
B
C
Profit per unit
X
$12
10
7
2
Y
$3
2
3
4
Z
$1
1
2
1
Available Hours
100
77 80
The available hours at the machine shops A, B, C are 100, 77, and 80 only. The profit per unit of
products X, Y, and Z is $12, $3, and $1 respectively.
Solution:
The formulation of Linear Programming (original problem) is as follows:
Maximize
12x1 + 3x2 + x3
Subject to:
10x1 + 2x2 + x3 100
7x1 + 3x2 + 2x3 77
2x1 + 4x2 + x3 80
x1, x2, x3 0
We introduce the slack variables s4, s5 and s6 then the equalities becomes as:
Maximize
12x1 + 3x2 + x3
Subject to:
10x1 + 2x2 + x3
+ s4 = 100
7x1 + 3x2 + 2x3
+ s5 = 77
2x1 + 4x2 + x3
+s6 = 80
x1, x2, x3, s4, s5, s6 0
Form the above equations, the first simplex table is obtained is as follows:
CB
Basic
Cj
12
3
1
0
0
0
Variable XB
x1
x2
x3
s4
s5
s6
0
s4
100
10
2
1
1
0
0
63
MBA-H2040 Quantitative Techniques for Managers
0
s5
77
7
3
2
0
1
0
0
s6
80
2
4
1
0
0
1
zj-cj
-12
-3
-1
0
0
0
Table 1
Note that the basic variables are s4, s5 and s6. Therefore CB1 = 0, CB2 = 0, CB3 = 0.
1. The smallest negative element in the above table of z1 ? c1 is -12. Hence, x1 becomes a basic
variable in the next iteration.
2. Determine the minimum ratios
Min 100, 72, 80 = 10
10 7 2
Here the minimum value is s4, which is made as a non-basic variable.
3. The next Table 2 is calculated using the following rules:
(i)
The revised basic variables are x1, s5, s6. Accordingly we make CB1=22, CB2=0 and
CB3=0.
(ii)
Since x1 is the incoming variable we make x1 coefficient one by dividing each
element of row 1 by 10. Thus the numerical value of the element corresponding to x2
is 2/10, corresponding to x3 is 1/10, corresponding to s4 is 1/10, corresponding to s5 is
0/10 and corresponding to s6 is 0/10 in Table 2.
(iii)
The incoming basic variable should only appear in the first row. So we multiply first
row of Table 2 by 7 and subtract if from the second row of Table 1 element by
element.
Thus,
The element corresponding to x1 in the second row of Table 2 is zero
The element corresponding to x2 is 3 ? 7 * 2 = 16
10 10
By using this way we get the elements of the second and the third row in Table 2.
Similarly, the calculation of numerical values of basic variables in Table 2 is done.
CB
Basic
Cj
22
6
2
0
0
0
Variable XB
x1
x2
x3
s4
s5
s6
12
x1
10
1
2/10
1/10
1/10
0
0
0
s5
7
0
16/10 13/10 -7/10
1
0
0
s6
60
0
18/5
4/5
-1/5
0
1
zj-cj
0
-3/5
1/5
6/5
0
0
Table 2
64
MBA-H2040 Quantitative Techniques for Managers
4. z2-c2 = -3/5. So x2 becomes a basic variable in the next iteration.
5. Determine the minimum of the ratios
10, 7 , 60
Min 2 16 18 = Min 50, 70, 300 = 70
10 10 5 16 18 16
So that the variable s5 will be a non basic variable in the next iteration.
6. From Table 2, the Table 3 is calculated using the rules (i), (ii) and (iii) mentioned above.
CB
Basic
Cj
12
3
1
0
0
0
Variable XB
x1
x2
x3
s4
s5
s6
12
x1
73/8
1
0
-1/16
3/16
-1/8
0
3
s5
35/8
0
1
13/16 -7/16
5/8
0
0
s6
177/4 0
0
-17/8
11/8
-9/4
1
zj-cj
0
0
11/16 15/16 3/8
0
Table 3
Since all the zi ? cj 0, the optimum solution is as:
x1 = 73/8 and x2 = 35/8 and
The Maximum Profit is: $981/8 = $122.625
Suppose an investor is deciding to purchase the resources A, B, C. What offers is he going to produce?
Let, assume that W1, W2 and W3 are the offers made per hour of machine time A, B and C respectively.
Then these prices W1, W2 and W3 must satisfy the conditions given below:
1. W1, W2, W3 0
2. Assume that the investor is behaving in a rational manner; he would try to bargain as much as
possible so that the total annual payable to the produces would be as little as possible. This leads
to the following condition:
Minimize
100W1 + 77W2 + 80W3
65
MBA-H2040 Quantitative Techniques for Managers
3. The total amount offer by the investor to the three resources viz. A, B and C required to produce
one unit of each product must be at least as high as the profit gained by the producer per unit.
Since, these resources enable the producer to earn the specified profit corresponding to the
product he would not like to sell it for anything less assuming he is behaving rationally. This
leads to the following conditions:
10w1 + 7w2 + 2w3 12
2w1 + 3w2 + 4w3 3
w1 + 2w2 + w3 1
Thus, in this case we have a linear problem to ascertain the values of the variable w1, w2, w3. The
variables w1, w2 and w3 are called as dual variables.
Note:
The original (primal) problem illustrated in this example
a. considers the objective function maximization
b. contains type constraints
c. has non-negative constraints
This original problem is called as primal problem in the standard form.
4.3 Dual Problem Properties
The following are the different properties of dual programming problem:
i.
If the original problem is in the standard form, then the dual problem solution is obtained
from the zj ? cj values of slack variables.
For example: In the Example 4.1, the variables s4, s4 and s6 are the slack variables. Hence
the dual problem solution is w1 = z4 ? c4 = 15/16, w2 = z5 ? c5 = 3/8 and w3 = z6 ? c6 = 0.
ii.
The original problem objective function maximum value is the minimum value of the dual
problem objective function.
For example:
From the above Example 4.1 we know that the original problem maximum values
is 981/8 = 122.625. So that the minimum value of the dual problem objective function is
100*15/16 + 77*3/8 + 80*0 = 981/8
Here the result has an important practical implication. If both producer and investor analyzed
the problem then neither of the two can outmaneuver the other.
iii.
Shadow Price: A resource shadow price is its unit cost, which is equal to the increase in
profit to be realized by one additional unit of the resource.
For example:
Let the minimum objective function value is expressed as:
66
MBA-H2040 Quantitative Techniques for Managers
100*15/16 + 77*3/8 + 80*0
If the first resource is increased by one unit the maximum profit also increases by 15/16,
which is the first dual variable of the optimum solution. Therefore, the dual variables are also
referred as the resource shadow price or imputed price. Note that in the previous example the
shadow price of the third resource is zero because there is already an unutilized amount, so
that profit is not increased by more of it until the current supply is totally exhausted.
iv.
In the originals problem, if the number of constraints and variables is m and n then the
constraint and variables in the dual problem is n and m respectively. Suppose the slack
variables in the original problem is represented by y1, y1, ....., yn and the surplus variables are
represented by z1, z2, ..., zn in the dual problem.
v.
Suppose, the original problem is not in a standard form, then the dual problem structure is
unchanged. However, if a constraint is greater than or equal to type, the corresponding dual
variable is negative or zero. Similarly, if a constraint in the original problem is equal to type,
then the corresponding dual variables is unrestricted in sign.
Example 4.2
Consider the following linear programming problem
Maximize
22x1 + 25x2 +19x3
Subject to:
18x1 + 26x2 + 22x3 350
14x1 + 18x2 + 20x3 180
17x1 + 19x2 + 18x3 = 205
x1, x2, x3 0
Note that this is a primal or original problem.
The corresponding dual problem for this problem is as follows:
Minimize
250w1 + 80w2 +105w3
Subject to:
18w1 + 4w2 + 7w3 22
26w1 + 18w2 + 19w3 25
22w1 + 20w2 + 18w3 19
w1 0, w2 , and w3 is unrestricted in sign (+ or -).
Now, we can solve this using simplex method as usual.
4.4 Simple Way of Solving Dual Problem
Solving of dual problem is simple; this is illustrated with the help of the following Example 4.3.
67
MBA-H2040 Quantitative Techniques for Managers
Example 4.3:
Minimize
P = x1 + 2x2
Subject to:
x1 + x2 8
2x1 + y 12
x1 1
Solution:
Step 1: Set up the P-matrix and its transpose
1 1 8
2 1 12
P = 1 0 1
1 2 0
1 2 1 1
PT = 1 1 0 2
8
12 1 0
w1 w2 w3 s1 s2 g z
1 2 1 1 0 0 1
Step 2: Form the
constraints and the
1 1 0 0 1 0 2
objective function
for the dual
-8 -12 -1 0 0 1 0
w1
+ 2w2 + w3 1
w1 + w2 2
z = 8w1 = 12w2 + 2
Step 3: Construct the initial simplex tableau for the dual
Since there are no negative entries in the last column above the third row, we have a standard
simplex problem. The most negative number in the bottom row to the left of the last column is
-12. This establishes the pivot column. The smallest nonnegative ratio is 1/2. The pivot element
is 2 in the w2-column.
68
MBA-H2040 Quantitative Techniques for Managers
Step 4: Pivoting
w1 w2 w3 s1 s2 g z
Pivoting about the
2 we get:
1 2 1 1 0 0 1
1/2 0 -1/2 -1/2 1 0 3/2
-2 0 5 6 0 1 6
w1 w2 w3 s1 s2 g z
1/2 1 1/2 1/2 0 0 1/2
1 1 0 0 1 0 2
-8 -12 -1 0 0 1 0
w1 w2 w3 s1 s2 g z
1/2 1 1/2 1/2 0 0 1/2
1/2 0 -1/2 -1/2 1 0 3/2
-2 0 5 6 0 1 6
The most negative entry in the bottom row to the left of the last column is -2. The smallest non-
negative ratio is the 1/2 in the first row. This is the next pivot element.
w1 w2 w3 s1 s2 g z
1/2 1 1/2 1/2 0 0 1/2
1/2 0 -1/2 -1/2 1 0 3/2
Pivoting about the
-2 0 5 6 0 1 6
1/2:
69
MBA-H2040 Quantitative Techniques for Managers
Since there are no negative entries in the bottom row and to the left of the last column, the
process is complete. The solutions are at the feet of the slack variable columns.
Therefore,
w1 w2 w3 s1 s2 g z
The
optimum solution
1 2 1 1 0 0 1
provided by x1 = 8
and x2 = 0
0 -1 -1 -1 1 0 1
The
Minimum Value
is: 8
0 4 7 8 0 1 8
4.5 Summary
For every linear programming problem there is a dual problem. The variables of the dual problem are
called as dual variables. The variables have economic value, which can be used for planning its
resources. The dual problem solution is achieved by the simplex method calculation of the original
(primal) problem. The dual problem solution has certain properties, which may be very useful for
calculation purposes.
4.6 Key Terms
Original Problem: This is the original linear programming problem, also called as primal problem.
Dual Problem: A dual problem is a linear programming problem is another linear programming
problem formulated from the parameters of the primal problem.
Dual Variables: Dual programming problem variables.
Optimum Solution: The solution where the objective function is minimized or maximized.
Shadow Price: Price of a resource is the change in the optimum value of the objective function per unit
increase of the resource.
4.7 Self Assessment Questions
Q1. An organization manufactures three products viz. A, B and c. The required raw material per piece of
product A, B and C is 2kg, 1kg, and 2kg. Assume that the total weekly availability is 50 kg. In order to
produce the products the raw materials are processed on a machine by the labour force and on a weekly
70
MBA-H2040 Quantitative Techniques for Managers
the availability of machine hours is 30. Assume that the available total labour hour is 26. The following
table illustrates time required per unit of the three products.
The profit per unit from the products A, B and C are #25, #30 and #40.
Formulate the dual linear programming problem and determine the optimum values of the dual
variables.
Q2. Consider the following dual problem
Minimize
3w1 + 4w2
Subject to:
3w1 + 4w2 24
2w1 + w2 10
5w1 + 3w2 29
w1, w2 0
4.8 Key Solutions
Q1. Minimize
50w1 + 30w2 + 26w3
Subject to:
2w1 + 0.5w2 + w3 25
w1 + 3w2 + 2w3 30
2w1 + w2 + w3 40
w1 = 50/3 = 16.6, w2 = 0 and w3 = 20/3 = 6.6
Q2.
Maximize
24x1 + 10x2 + 29x3
Subject
to:
Product
Labour
Machine
3x1 + 2x2 + 5x3 3
Hour
Hour
4x1 + x2 + 3x3 4
A
0.5
1
x1, x2, x3 0 and w1 = 4, w2 = 3
B
3
2
Objective Function Maximum
Value is: 24
C
1
1
71
MBA-H2040 Quantitative Techniques for Managers
4.9 Further References
Mustafi, C.K. 1988. Operations Research Methods and Practices, Wiley Eastern Limited, New Delhi.
Hamdy A Taha, 1999. Introduction to Operations Research, PHI Limited, New Delhi.
Peterson R and Silver, E. A. 1979. Decision Systems for Inventory Management and Production
Planning.
Levin, R and Kirkpatrick, C.A. 1978. Quantitative Approached to Management, Tata McGraw Hill,
Kogakusha Ltd., International Student Edition.
72
MBA-H2040 Quantitative Techniques for Managers
UNIT II
ON
S
S
1 TRANSPORTATION PROBLEM
E
L
LESSON STRUCTURE
1.5 Introduction
1.6 Transportation Algorithm
1.7 Basic Feasible Solution of a
Transportation Problem
1.8 Modified Distribution Method
1.9 Unbalanced Transportation Problem
1.10
Degenerate
Transportation
Problem
1.11
Transshipment Problem
1.12
Transportation
Problem
Maximization
1.13
Summary
1.14
Key Terms
1.15
Self Assessment Questions
1.16
Key Solutions
1.17
Further References
Objectives
After Studying this lesson, you should be able
to:
Formulation of a Transportation Problem
Determine basic feasible solution using
various methods
Understand the MODI, Stepping Stone
Methods for cost minimization
Make unbalanced Transportation Problem
into balanced one using appropriate
method
Solve Degenerate Problem
Formulate and Solve Transshipment
Problem
Describe suitable method for maximizing
the objective function instead of
minimizing
73
MBA-H2040 Quantitative Techniques for Managers
1.1 Introduction
A special class of linear programming problem is Transportation Problem, where the objective is to
minimize the cost of distributing a product from a number of sources (e.g. factories) to a number of
destinations (e.g. warehouses) while satisfying both the supply limits and the demand requirement.
Because of the special structure of the Transportation Problem the Simplex Method of solving is
unsuitable for the Transportation Problem. The model assumes that the distributing cost on a given rout
is directly proportional to the number of units distributed on that route. Generally, the transportation
model can be extended to areas other than the direct transportation of a commodity, including among
others, inventory control, employment scheduling, and personnel assignment.
The transportation problem special feature is illustrated here with the help of following Example 1.1.
Example 1.1:
Suppose a manufacturing company owns three factories (sources) and distribute his products to five
different retail agencies (destinations). The following table shows the capacities of the three factories,
the quantity of products required by the various retail agencies and the cost of shipping one unit of the
product from each of the three factories to each of the five retail agencies.
Retail Agency
Factories
1
2
3
4
5
Capacity
1
1
9
13
36
51
50
2
24
12
16
20
1
100
3
14
33
1
23
26
150
Requirement
100
60
50
50
40
300
Usually the above table is referred as Transportation Table, which provides the basic information
regarding the transportation problem. The quantities inside the table are known as transportation cost per
unit of product. The capacity of the factories 1, 2, 3 is 50, 100 and 150 respectively. The requirement of
the retail agency 1, 2, 3, 4, 5 is 100,60,50,50, and 40 respectively.
In this case, the transportation cost of one unit
from factory 1 to retail agency 1 is 1,
from factory 1 to retail agency 2 is 9,
from factory 1 to retail agency 3 is 13, and so on.
A transportation problem can be formulated as linear programming problem using variables with
two subscripts.
Let
74
MBA-H2040 Quantitative Techniques for Managers
x11=Amount to be transported from factory 1 to retail agency 1
x12= Amount to be transported from factory 1 to retail agency 2
........
........
........
........
x35= Amount to be transported from factory 3 to retail agency 5.
Let the transportation cost per unit be represented by C11, C12, .....C35 that is C11=1, C12=9, and so on.
Let the capacities of the three factories be represented by a1=50, a2=100, a3=150.
Let the requirement of the retail agencies are b1=100, b2=60, b3=50, b4=50, and b5=40.
Thus, the problem can be formulated as
Minimize
C11x11+C12x12+...............+C35x35
Subject to:
x11 + x12 + x13 + x14 + x15 = a1
x21 + x22 + x23 + x24 + x25 = a2
x31 + x32 + x33 + x34 + x35 = a3
x11 + x21 + x31 = b1
x12 + x22 + x32 = b2
x13 + x23 + x33 = b3
x14 + x24 + x34 = b4
x15 + x25 + x35 = b5
x11, x12, ......, x35 0.
Thus, the problem has 8 constraints and 15 variables. So, it is not possible to solve such a
problem using simplex method. This is the reason for the need of special computational procedure to
solve transportation problem. There are varieties of procedures, which are described in the next section.
1.2 Transportation Algorithm
The steps of the transportation algorithm are exact parallels of the simplex algorithm, they are:
Step 1: Determine a starting basic feasible solution, using any one of the following three methods
1. North West Corner Method
2. Least Cost Method
3. Vogel Approximation Method
Step 2: Determine the optimal solution using the following method
1. MODI (Modified Distribution Method) or UV Method.
1.3 Basic Feasible Solution of a Transportation Problem
75
MBA-H2040 Quantitative Techniques for Managers
The computation of an initial feasible solution is illustrated in this section with the help of the
example1.1 discussed in the previous section. The problem in the example 1.1 has 8 constraints and 15
variables we can eliminate one of the constraints since a1 + a2 + a3 = b1 + b2 + b3 + b4 +b5. Thus now the
problem contains 7 constraints and 15 variables. Note that any initial (basic) feasible solution has at
most 7 non-zero Xij. Generally, any basic feasible solution with m sources (such as factories) and n
destination (such as retail agency) has at most m + n -1 non-zero Xij.
The special structure of the transportation problem allows securing a non artificial basic feasible
solution using one the following three methods.
4. North West Corner Method
5. Least Cost Method
6. Vogel Approximation Method
The difference among these three methods is the quality of the initial basic feasible solution they
produce, in the sense that a better that a better initial solution yields a smaller objective value. Generally
the Vogel Approximation Method produces the best initial basic feasible solution, and the North West
Corner Method produces the worst, but the North West Corner Method involves least computations.
North West Corner Method:
The method starts at the North West (upper left) corner cell of the tableau (variable x11).
Step -1: Allocate as much as possible to the selected cell, and adjust the associated amounts of capacity
(supply) and requirement (demand) by subtracting the allocated amount.
Step -2: Cross out the row (column) with zero supply or demand to indicate that no further assignments
can be made in that row (column). If both the row and column becomes zero simultaneously, cross out
one of them only, and leave a zero supply or demand in the uncrossed out row (column).
Step -3: If exactly one row (column) is left uncrossed out, then stop. Otherwise, move to the cell to the
right if a column has just been crossed or the one below if a row has been crossed out. Go to step -1.
Example 1.2:
Consider the problem discussed in Example 1.1 to illustrate the North West Corner Method of
determining basic feasible solution.
Retail Agency
Factories
1
2
3
4
5
Capacity
1
1
9
13
36
51
50
2
24
12
16
20
1
100
3
14
33
1
23
26
150
Requirement
100
60
50
50
40
300
76
MBA-H2040 Quantitative Techniques for Managers
The allocation is shown in the following tableau:
Capacity
1 9 13 36 51
50
50
24 12 16 20 1
100 50
50 50
14 33 1 23 26
10
50
50
40
150 140 90 40
Requirement 100
60
50
50
40
50 10
The arrows show the order in which the allocated (bolded) amounts are generated. The starting
basic solution is given as
x11 = 50,
x21 = 50, x22 = 50
x32 = 10, x33 = 50, x34 = 50, x35 = 40
The corresponding transportation cost is
50 * 1 + 50 * 24 + 50 * 12 + 10 * 33 + 50 * 1 + 50 * 23 + 40 * 26 = 4420
It is clear that as soon as a value of Xij is determined, a row (column) is eliminated from further
consideration. The last value of Xij eliminates both a row and column. Hence a feasible solution
computed by North West Corner Method can have at most m + n ? 1 positive Xij if the transportation
problem has m sources and n destinations.
Least Cost Method
The least cost method is also known as matrix minimum method in the sense we look for the row and
the column corresponding to which Cij is minimum. This method finds a better initial basic feasible
solution by concentrating on the cheapest routes. Instead of starting the allocation with the northwest
cell as in the North West Corner Method, we start by allocating as much as possible to the cell with the
smallest unit cost. If there are two or more minimum costs then we should select the row and the column
corresponding to the lower numbered row. If they appear in the same row we should select the lower
numbered column. We then cross out the satisfied row or column, and adjust the amounts of capacity
and requirement accordingly. If both a row and a column is satisfied simultaneously, only one is crossed
out. Next, we look for the uncrossed-out cell with the smallest unit cost and repeat the process until we
are left at the end with exactly one uncrossed-out row or column.
77
MBA-H2040 Quantitative Techniques for Managers
Example 1.3:
The least cost method of determining initial basic feasible solution is illustrated with the help of problem
presented in the section 1.1.
Capacity
1 9 13 36 51
50
50
24 12 16 20 1
100 60
60
40
14 33 1 23 26
50
50
50
150 100 50
Requirement 100
60
50
50
40
50
The Least Cost method is applied in the following manner:
We observe that C11=1 is the minimum unit cost in the table. Hence X11=50 and the first row is
crossed out since the row has no more capacity. Then the minimum unit cost in the uncrossed-out row
and column is C25=1, hence X25=40 and the fifth column is crossed out. Next C33=1is the minimum unit
cost, hence X33=50 and the third column is crossed out. Next C22=12 is the minimum unit cost, hence
X22=60 and the second column is crossed out. Next we look for the uncrossed-out row and column now
C31=14 is the minimum unit cost, hence X31=50 and crossed out the first column since it was satisfied.
Finally C34=23 is the minimum unit cost, hence X34=50 and the fourth column is crossed out.
So that the basic feasible solution developed by the Least Cost Method has transportation cost is
1 * 50 + 12 * 60 + 1 * 40 + 14 * 50 + 1 * 50 + 23 * 50 = 2710
Note that the minimum transportation cost obtained by the least cost method is much lower than
the corresponding cost of the solution developed by using the north-west corner method.
Vogel Approximation Method (VAM):
VAM is an improved version of the least cost method that generally produces better solutions. The steps
involved in this method are:
Step 1: For each row (column) with strictly positive capacity (requirement), determine a penalty by
subtracting the smallest unit cost element in the row (column) from the next smallest unit cost element
in the same row (column).
78
MBA-H2040 Quantitative Techniques for Managers
Step 2: Identify the row or column with the largest penalty among all the rows and columns. If the
penalties corresponding to two or more rows or columns are equal we select the topmost row and the
extreme left column.
Step 3: We select Xij as a basic variable if Cij is the minimum cost in the row or column with largest
penalty. We choose the numerical value of Xij as high as possible subject to the row and the column
constraints. Depending upon whether ai or bj is the smaller of the two ith row or jth column is crossed out.
Step 4: The Step 2 is now performed on the uncrossed-out rows and columns until all the basic variables
have been satisfied.
Example 1.4:
Consider the following transportation problem
Destination
Origin
1
2
3
4
ai
1
20
22
17
4
120
2
24
37
9
7
70
3
32
37
20
15
50
bj
60
40
30
110
240
Note: ai=capacity (supply)
bj=requirement (demand)
Now, compute the penalty for various rows and columns which is shown in the following table:
Destination
Origin
1
2
3
4
ai Column
Penalty
1
20
22
17
4
120 13
2
24
37
9
7
70 2
3
32
37
20
15
50 5
bj
60
40
30
110
240
Row Penalty
4
15
8
3
Look for the highest penalty in the row or column, the highest penalty occurs in the second column and
the minimum unit cost i.e. cij in this column is c12=22. Hence assign 40 to this cell i.e. x12=40 and cross
out the second column (since second column was satisfied. This is shown in the following table:
Destination
Origin
1
2
3
4
ai Column
Penalty
1
20
22 40
17
4
80 13
2
24
37
9
7
70 2
3
32
37
20
15
50 5
bj
60
40
30
110
240
79
MBA-H2040 Quantitative Techniques for Managers
Row Penalty
4
15
8
3
The next highest penalty in the uncrossed-out rows and columns is 13 which occur in the first row and
the minimum unit cost in this row is c14=4, hence x14=80 and cross out the first row. The modified table
is as follows:
Destination
Origin
1
2
3
4
ai Column
Penalty
1
20
22
17
4 0 13
40
80
2
24
37
9
7
70 2
3
32
37
20
15
50 5
bj
60
40
30
110
240
Row Penalty
4
15
8
3
The next highest penalty in the uncrossed-out rows and columns is 8 which occurs in the third column
and the minimum cost in this column is c23=9, hence x23=30 and cross out the third column with
adjusted capacity, requirement and penalty values. The modified table is as follows:
Destination
Origin
1
2
3
4
ai Column
Penalty
1
20
22
17
4 0 13
40
80
2
24
37
9
7
40 17
30
3
32
37
20
15
50 17
bj
60
40
30
110
240
Row Penalty
8
15
8
8
The next highest penalty in the uncrossed-out rows and columns is 17 which occurs in the second row
and the smallest cost in this row is c24=15, hence x24=30 and cross out the fourth column with the
adjusted capacity, requirement and penalty values. The modified table is as follows:
Destination
Origin
1
2
3
4
ai Column
Penalty
1
20
22
17
4 0 13
40
80
2
24
37
9
7
10 17
30
30
3
32
37
20
15
50 17
bj
60
40
30
110
240
80
MBA-H2040 Quantitative Techniques for Managers
Row Penalty
8
15
8
8
The next highest penalty in the uncrossed-out rows and columns is 17 which occurs in the second row
and the smallest cost in this row is c21=24, hence xi21=10 and cross out the second row with the adjusted
capacity, requirement and penalty values. The modified table is as follows:
Destination
Origin
1
2
3
4
ai Column
Penalty
1
20
22
17
4 0 13
40
80
2
24
37
9
7
0 17
10
30
30
3
32
37
20
15
50 17
bj
60
40
30
110
240
Row Penalty
8
15
8
8
The next highest penalty in the uncrossed-out rows and columns is 17 which occurs in the third row and
the smallest cost in this row is c31=32, hence xi31=50 and cross out the third row or first column. The
modified table is as follows:
Destination
Origin
1
2
3
4
ai Column
Penalty
1
20
22
17
4 0 13
40
80
2
24
37
9
7
0 17
10
30
30
3
32
37
20
15
0 17
50
bj
60
40
30
110
240
Row Penalty
8
15
8
8
The transportation cost corresponding to this choice of basic variables is
22 * 40 + 4 * 80 + 9 * 30 + 7 * 30 + 24 * 10 + 32 * 50 = 3520
1.4 Modified Distribution Method
The Modified Distribution Method, also known as MODI method or u-v method, which provides a
minimum cost solution (optimal solution) to the transportation problem. The following are the steps
involved in this method.
81
MBA-H2040 Quantitative Techniques for Managers
Step 1: Find out the basic feasible solution of the transportation problem using any one of the three
methods discussed in the previous section.
Step 2: Introduce dual variables corresponding to the row constraints and the column constraints. If
there are m origins and n destinations then there will be m+n dual variables. The dual variables
corresponding to the row constraints are represented by ui, i=1,2,.....m where as the dual variables
corresponding to the column constraints are represented by vj, j=1,2,.....n. The values of the dual
variables are calculated from the equation given below
ui + vj = cij if xij > 0
Step 3: Any basic feasible solution has m + n -1 xij > 0. Thus, there will be m + n -1 equation to
determine m + n dual variables. One of the dual variables can be chosen arbitrarily. It is also to be noted
that as the primal constraints are equations, the dual variables are unrestricted in sign.
Step 4: If xij=0, the dual variables calculated in Step 3 are compared with the cij values of this allocation
as cij ? ui ? vj. If al cij ? ui ? vj 0, then by the theorem of complementary slackness it can be shown that
the corresponding solution of the transportation problem is optimum. If one or more cij ? ui ? vj < 0, we
select the cell with the least value of cij ? ui ? vj and allocate as much as possible subject to the row and
column constraints. The allocations of the number of adjacent cell are adjusted so that a basic variable
becomes non-basic.
Step 5: A fresh set of dual variables are calculated and repeat the entire procedure from Step 1 to Step 5.
Example 1.5:
For example consider the transportation problem given below:
Supply
1 9 13 36 51
50
24 12 16 20 1
100
14 33 1 23 26
150
Demand 100
70
50
40
40
300
Step 1: First we have to determine the basic feasible solution. The basic feasible solution using least
cost method is
x11=50, x22=60, x25=40, x31=50, x32=10, x33=50 and x34=40
Step 2: The dual variables u1, u2, u3 and v1, v2, v3, v4, v5 can be calculated from the corresponding cij
values, that is
u1+v1=1
u2+v2=12
u2+v5=1
u3+v1=14
u3+v2=33
u3+v3=1
u3+v4=23
Step 3: Choose one of the dual variables arbitrarily is zero that is u3=0 as it occurs most often in the
above equations. The values of the variables calculated are
82
MBA-H2040 Quantitative Techniques for Managers
u1= -13, u2= -21, u3=0
v1=14, v2=33, v3=1, v4=23, v5=22
Step 4: Now we calculate cij ? ui ? vj values for all the cells where xij=0 (.e. unallocated cell by the basic
feasible solution)
That is
Cell(1,2)= c12-u1-v2 = 9+13-33 = -11
Cell(1,3)= c13-u1-v3 = 13+13-1 = 25
Cell(1,4)= c14-u1-v4 = 36+13-23 = 26
Cell(1,5)= c15-u1-v5 = 51+13-22 = 42
Cell(2,1)= c21-u2-v1 = 24+21-14 = 31
Cell(2,3)= c23-u2-v3 = 16+21-1 = 36
Cell(2,4)= c24-u2-v4 = 20+21-23 = 18
Cell(3,5)= c35-u3-v5 = 26-0-22 = 4
Note that in the above calculation all the cij ? ui ? vj 0 except for cel (1, 2) where c12 ? u1 ? v2 = 9+13-
33 = -11.
Thus in the next iteration x12 will be a basic variable changing one of the present basic variables
non-basic. We also observe that for allocating one unit in cell (1, 2) we have to reduce one unit in cells
(3, 2) and (1, 1) and increase one unit in cell (3, 1). The net transportation cost for each unit of such
reallocation is
-33 -1 + 9 +14 = -11
The maximum that can be allocated to cell (1, 2) is 10 otherwise the allocation in the cell (3, 2)
will be negative. Thus, the revised basic feasible solution is
x11=40, x12=10, x22=60, x25=40, x31=60, x33=50, x34=40
1.5 Unbalanced Transportation Problem
In the previous section we discussed about the balanced transportation problem i.e. the total supply
(capacity) at the origins is equal to the total demand (requirement) at the destination. In this section we
are going to discuss about the unbalanced transportation problems i.e. when the total supply is not equal
to the total demand, which are called as unbalanced transportation problem.
In the unbalanced transportation problem if the total supply is more than the total demand then
we introduce an additional column which will indicate the surplus supply with transportation cost zero.
Similarly, if the total demand is more than the total supply an additional row is introduced in the
transportation table which indicates unsatisfied demand with zero transportation cost.
Example 1.6:
Consider the following unbalanced transportation problem
Warehouses
83
MBA-H2040 Quantitative Techniques for Managers
Plant w1 w2 w3 Supply
20
17
25
X
400
10
10
20
Y
500
Demand 400
400
500
In this problem the demand is 1300 whereas the total supply is 900. Thus, we now introduce an
additional row with zero transportation cost denoting the unsatisfied demand. So that the modified
transportation problem table is as follows:
Warehouses
Plant w1 w2 w3 Supply
20
17
25
X
400
10
10
20
Y
500
Unsatisfied
0
0
0
Demand
400
Demand
400
400
500
1300
Now we can solve as balanced problem discussed as in the previous sections.
1.6. Degenerate Transportation Problem
In a transportation problem, if a basic feasible solution with m origins and n destinations has less than m
+ n -1 positive Xij i.e. occupied cells, then the problem is said to be a degenerate transportation
problem. The degeneracy problem does not cause any serious difficulty, but it can cause computational
problem wile determining the optimal minimum solution.
There fore it is important to identify a degenerate problem as early as beginning and take the
necessary action to avoid any computational difficulty. The degeneracy can be identified through the
following results:
"In a transportation problem, a degenerate basic feasible solution exists if and only if some
partial sum of supply (row) is equal to a partial sum of demand (column). For example the following
transportation problem is degenerate. Because in this problem
a1 = 400 = b1
a2 + a3 = 900 = b2 + b3
Warehouses
Plant w1 w2 w3 Supply (ai)
84
MBA-H2040 Quantitative Techniques for Managers
20
17
25
X
400
10
10
20
Y
500
Unsatisfied
0
0
0
demand
400
Demand (bj)
400
400
500
1300
There is a technique called perturbation, which helps to solve the degenerate problems.
Perturbation Technique:
The degeneracy of the transportation problem can be avoided if we ensure that no partial sum of
ai (supply) and bj (demand) is equal. We set up a new problem where
ai = ai +d
i = 1, 2, ......, m
bj = bj
j = 1, 2, ......, n -1
bn = bn + md
d > 0
This modified problem is constructed in such a way that no partial sum of ai is equal to the bj.
Once the problem is solved, we substitute d = 0 leading to optimum solution of the original problem.
Example: 1.7
Consider the above problem
Warehouses
Plant w1 w2 w3 Supply (ai)
20
17
25
X
400 + d
10
10
20
Y
500 + d
Unsatisfied
0
0
0
demand
400 + d
Demand (bj)
400
400
500 + 3d 1300 + 3d
Now this modified problem can be solved by using any of the three methods viz. North-west Corner,
Least Cost, or VAM.
1.7 Transshipment Problem
There could be a situation where it might be more economical to transport consignments in several sages
that is initially within certain origins and destinations and finally to the ultimate receipt points, instead of
transporting the consignments from an origin to a destination as in the transportation problem.
85
MBA-H2040 Quantitative Techniques for Managers
The movement of consignment involves tow different modes of transport viz. road and railways
or between stations connected by metre gauge and broad gauge lines. Similarly it is not uncommon to
maintain dumps for central storage of certain bulk material. These require transshipment.
Thus for the purpose of transshipment the distinction between an origin and destination is
dropped so that from a transportation problem with m origins and n destinations we obtain a
transshipment problem with m + n origins and m + n destinations.
The formulation and solution of a transshipment problem is illustrated with the following
Example 1.8.
Example 1.8:
Consider the following transportation problem where the origins are plants and destinations are depots.
Table 1
Depot
Plant
X
Y
Z
Supply
A
150
$1
$3
$15
B
300
$3
$5
$25
Demand
150
150
150 450
When each plant is also considered as a destination and each depot is also considered as an
origin, there are altogether five origins and five destinations. So that some additional cost data are
necessary, they are as follows:
Table 2
Unit transportation cost From Plant To Plant
To
Plant A
Plant B
From
Plant A
0
55
Plant B
2
0
Table 3
Unit transportation cost From Depot To Depot
To
Depot X
Depot Y
Depot Z
86
MBA-H2040 Quantitative Techniques for Managers
From
Depot X
0
25
2
Depot Y
2
0
3
Depot Z
55
3
0
Table 4
Unit transportation cost From Depot to Plant
To
Plant A
Plant B
From
Depot X
3
15
Depot Y
25
3
Depot Z
45
55
Now, from the Table
1, Table 2, Table 3, Table 4
we obtain the transportation formulation of the transshipment problem, which is shown in the Table 5.
Table 5
Transshipment Table
Supply
A
B
X
Y
Z
A
150+450=600
0
55
1
3
15
B
300+450=750
2
0
3
5
25
X
450
3
15
0
25
2
Y
450
25
3
2
0
3
Z
450
45
55
55
3
0
Demand
450
450
150+450= 150+450= 150+450=
600
600
600
A buffer stock of 450 which is the total supply and total demand in the original transportation
problem is added to each row and column of the transshipment problem. The resulting transportation
problem has m + n = 5 origins and m + n = 5 destinations.
By solving the transportation problem presented in the Table 5, we obtain
x11=150
x13=300
x14=150
x21=3001
x22=450
x33=300
x35=150
x44=450
x55=450
The transshipment problem explanation is as follows:
87
MBA-H2040 Quantitative Techniques for Managers
1. Transport x21=300 from plant B to plant A. This increase the availability at plant A to 450 units
including the 150 originally available from A.
2. From plant A transport x13=300 to depot X and x14=150 to depot Y.
3. From depot X transport x35=150 to depot Z.
Thus, the total cost of transshipment is:
2*300 + 3 * 150 + 1*300 + 2*150 = $1650
Note: The consignments are transported from pants A, B to depots X, Y, Z only according to the
transportation Table 1, the minimum transportation cost schedule is x13=150 x21=150 x22=150 with a
minimum cost of 3450.
Thus, transshipment reduces the cost of consignment movement.
1.8 Transportation Problem Maximization
There are certain types of transportation problem where the objective function is to be maximized
instead of minimized. These kinds of problems can be solved by converting the maximization problem
into minimization problem. The conversion of maximization into minimization is done by subtracting
the unit costs from the highest unit cost of the table.
The maximization of transportation problem is illustrated with the following Example 1.9.
Example 1.9:
A company has three factories located in three cities viz. X, Y, Z. These factories supplies consignments
to four dealers viz. A, B, C and D. The dealers are spread all over the country. The production capacity
of these factories is 1000, 700 and 900 units per month respectively. The net return per unit product is
given in the following table.
Dealers
Factory
A
B
C
D
capacity
X
6
6
6
4
1000
Y
4
2
4
5
700
Z
900
5
6
7
8
Requirement 900
800 500 400 2600
Determine a suitable allocation to maximize the total return.
88
MBA-H2040 Quantitative Techniques for Managers
This is a maximization problem. Hence first we have to convert this in to minimization problem.
The conversion of maximization into minimization is done by subtracting the unit cost of the table from
the highest unit cost.
Look the table, here 8 is the highest unit cost. So, subtract all the unit cost from the 8, and then
we get the revised minimization transportation table, which is given below.
Dealers
Factory
A
B
C
D
capacity
X
2
2
2
4
1000 = a1
Y
4
6
4
3
700 =a2
Z
900 =a3
3
2
1
0
Requirement 900=b1
800=b2 500=b3 400=b4 2600
Now we can solve the problem as a minimization problem.
The problem here is degenerate, since the partial sum of a1=b2+b3 or a3=b3. So consider the
corresponding perturbed problem, which is shown below.
Dealers
Factory
A
B
C
D
capacity
X
2
2
2
4
1000+d
Y
4
6
4
3
700+d
Z
900+d
3
2
1
0
Requirement 900
800 500 400+3d 2600+3d
First we have to find out the basic feasible solution. The basic feasible solution by lest cost
method is x11=100+d, x22=700-d, x23=2d, x33=500-2d and x34=400+3d.
Once if the basic feasible solution is found, next we have to determine the optimum solution
using MODI (Modified Distribution Method) method. By using this method we obtain
u1+v1=2
u1+v2=2
u2+v2=6
u2+v3=4
u3+v3=1
u3+v4=0
Taking u1=0 arbitrarily we obtain
u1=0, u2=4, u3=1 and
v1=2, v2=3, v3=0
On verifying the condition of optimality, we know that
89
MBA-H2040 Quantitative Techniques for Managers
C12-u1-v2 < 0
and
C32-u3-v2 <0
So, we allocate x12=700-d and make readjustment in some of the other basic variables.
The revised values are:
x11=200+d, x12=800, x21=700-d, x23=2d, x33=500-3d, and x34=400+3d
u1+v1=2
u1+v2=2
u2+v1=4
u2+v3=4
u3+v3=1
u3+v4=0
Taking u1=0 arbitrarily we obtain
u1=0, u2=2, u3=-1
v1=2, v2=2, v3=2, v4=1
Now, the optimality condition is satisfied.
Finally, taking d=0 the optimum solution of the transportation problem is
X11=200, x12=800, x21=700, x33=500 and x34=400
Thus, the maximum return is:
6*200 + 6*800 + 4*700 + 7*500 + 8*400 = 15500
1.9 Summary
Transportation Problem is a special kind of linear programming problem. Because of the transportation
problem special structure the simplex method is not suitable. But which may be utilized to make
efficient computational techniques for its solution.
Generally transportation problem has a number of origins and destination. A certain amount of
consignment is available in each origin. Similarly, each destination has a certain demand/requirements.
The transportation problem represents amount of consignment to be transported from different origins to
destinations so that the transportation cost is minimized with out violating the supply and demand
constraints.
There are two phases in the transportation problem. First is the determination of basic feasible
solution and second is the determination of optimum solution.
There are three methods available to determine the basic feasible solution, they are
1. North West Corner Method
2. Least Cost Method or Matrix Minimum Method
3. Vogel's Approximation Method (VAM)
In order to determine optimum solution we can use either one of the following method
1. Modified Distribution (MODI) Method
Or
2. Stepping Stone Method
Transportation problem can be generalized into a Transshipment Problem where transportation
of consignment is possible from origin to origin or destination as well as destination to origin or
90
MBA-H2040 Quantitative Techniques for Managers
destination. The transshipment problem may be result in an economy way of shipping in some
situations.
1.10 Key Terms
Origin:
is the location from which the shipments are dispatched.
Destination: is the location to which the shipments are transported.
Unit Transportation Cost: is the transportation cost per unit from an origin to destination.
Perturbation Technique: is a method of modifying a degenerate transportation problem in order to
solve the degeneracy.
1.11 Self Assessment Questions
Q1. Four companies viz. W, X, Y and Z supply the requirements of three warehouses viz. A, B and C
respectively. The companies' availability, warehouses requirements and the unit cost of transportation
are given in the following table. Find an initial basic feasible solution using
a. North West Corner Method
b. Least Cost Method
c. Vogel Approximation Method (VAM)
Warehouses
Company
A
B
C
Supply
W
10
8
9
15
X
5
2
3
20
Y
30
6
7
4
Z
35
7
6
9
Requirement
25
26
49 100
Q2. Find the optimum Solution of the following Problem using MODI method.
Destination
Source
1
2
3
Capacity
A 42
8
9
10
B
30
9
11
11
C
28
10
12
9
Demand 35 40 25 100
91
MBA-H2040 Quantitative Techniques for Managers
Q3. The ABT transport company ships truckloads of food grains from three sources viz. X, Y, Z to four
mills viz. A, B, C, D respectively. The supply and the demand together with the unit transportation cost
per truckload on the different routes are described in the following transportation table. Assume that the
unit transportation costs are in hundreds of dollars. Determine the optimum minimum shipment cost of
transportation using MODI method.
Mill
Source
A
B
C
D
Supply
X
10
2
20
11
15
Y
12
7
9
20
25
Z
10
4
14
16
18
Demand 5 15 15 15
Q4. An organization has three plants at X, Y, Z which supply to warehouses located at A, B, C, D, and E
respectively. The capacity of the plants is 800, 500 and 900 per month and the requirement of the
warehouses is 400, 400, 500, 400 and 800 units respectively. The following table shows the unit
transportation cost.
A
B
C
D E
X
$5
$8
$6
$6
$3
Y
$4
$7
$7
$6
$6
Z
$8
$4
$6
$6
$3
Determine an optimum distribution for the organization in order to minimize the total cost of
transportation.
Q5. Solve the following transshipment problem
Consider a transportation problem has tow sources and three depots. The availability, requirements and
unit cost are as follows:
Depot
Source
D1
D2
D3 Availability
S1
30
9
8
1
S2
1
7
8
30
Requirement 20 20 20 60
In addition to the above, suppose that the unit cost of transportation from source to source and from
depot to depot are as:
Source
92
MBA-H2040 Quantitative Techniques for Managers
S1
S2
S1
Source
0
1
S2 2
0
Depot
D1
D2
D3
D1
0
2
1
Depot
D2
2
0
9
D3 1
9
0
Find out minimum transshipment cost of the problem and also compare this cost with the corresponding
minimum transportation cost.
Q6. Saravana Store, T.Nagar, Chennai interested to purchase the following type and quantities of dresses
Dress
V
W
X
Y
Z
Type
Quantity
150
100
75
250
200
Four different dress makers are submitted the tenders, who undertake to supply not more than the
quantities indicated below:
Dress
A
B
C
D
Maker
Dress
300
250
150
200
Quantity
Saravana Store estimates that its profit per dress will vary according to the dress maker as indicates in
the following table:
V W
X
Y Z
A
2.75
3.5
4.25
2.25
1.5
B
3
3.25
4.5
1.75
1
C
2.5
3.5
4.75
2
1.25
D
3.25
2.75
4
2.5
1.75
Determine how should the orders to be places for the dresses so as to maximize the profit.
93
MBA-H2040 Quantitative Techniques for Managers
1.12 Key Solutions
Q1. a. x11 = 15, x21 = 10, x22 = 10, x32 = 16, x33 = 14, x43 = 35
Minimum Cost is: 753
b. x13 = 15, x22 = 20, x33 = 30, x41 = 25, x42 = 6, x43 = 4
Minimum Cost is: 542
c. x13=15, x22=20, x33=30, x41=25, x42=6, x43=4
Minimum Cost is: 542
Q2. x11=2, x12=40, x21=30, x31=3, x33=25
Minimum Transportation Optimal cost is: 901.
Q3. x12=5, x14=10, x22=10, x23=15, x31=5, x34=5
Minimum Optimal Cost is: $435
Q4. x15=800, x21=400, x24=100, x32=400, x33=200, x34=300, x43=300 (supply shortage)
Minimum Cost of Transportation is: $9200
Q5. Transportation Problem
S1-D2=10, S1-D3=20, S2-D1=20, S2-D2=10 and
Minimum Transportation Cost is: 100
Transshipment Problem
x11=60, x12=10, x15=20, x22=50, x23=40, x33=40, x34=20, x44=60, x55=60 and
Minimum Transshipment Cost is: 100
Q6. 150 dresses of V and 50 dresses of Z by Dress Maker A
250 dresses of Y by Dress Maker B
150 dress of Z by Dress Maker C
100 dress of W and 75 dresses of X by Dress Maker D
Maximum Profit is: 1687.50
1.13 Further References
Hamdy A Taha. 1999. Introduction to Operations Research, PHI Limited, New Delhi.
Mustafi, C.K. 1988. Operations Research Methods and Practice, Wiley Eastern Ltd., New Delhi.
Mittal, K.V. 1976. Optimization Methods in Operations Research and Systems Analysis, Wiley Eastern
Ltd., New Delhi.
94
MBA-H2040 Quantitative Techniques for Managers
UNIT II
ON
S
2 ASSIGNMENT PROBLEM
S
E
L
LESSON STRUCTURE
2.1 Introduction
2.2 Assignment Problem Structure and
Solution
2.3 Unbalanced Assignment Problem
2.4 Infeasible Assignment Problem
2.5 Maximization in an Assignment
Problem
2.6 Crew Assignment Problem
2.7 Summary
2.8 Key Solutions
2.9 Self Assessment Questions
2.10 Key Answers
2.11 Further References
95
MBA-H2040 Quantitative Techniques for Managers
Objectives
After Studying this lesson, you should be able
to:
Assignment Problem Formulation
How to solve the Assignment Problem
How to solve the unbalanced problem
using appropriate method
Make appropriate modification when
some problems are infeasible
Modify the problem when the objective is
to maximize the objective function
Formulate and solve the crew assignment
problems
96
MBA-H2040 Quantitative Techniques for Managers
2.1 Introduction
The Assignment Problem can define as follows:
Given n facilities, n jobs and the effectiveness of each facility to each job, here the problem is to assign
each facility to one and only one job so that the measure of effectiveness if optimized. Here the
optimization means Maximized or Minimized. There are many management problems has a assignment
problem structure. For example, the head of the department may have 6 people available for assignment
and 6 jobs to fill. Here the head may like to know which job should be assigned to which person so that
all tasks can be accomplished in the shortest time possible. Another example a container company may
have an empty container in each of the location 1, 2,3,4,5 and requires an empty container in each of the
locations 6, 7, 8,9,10. It would like to ascertain the assignments of containers to various locations so as
to minimize the total distance. The third example here is, a marketing set up by making an estimate of
sales performance for different salesmen as well as for different cities one could assign a particular
salesman to a particular city with a view to maximize the overall sales.
Note that with n facilities and n jobs there are n! possible assignments. The simplest way of
finding an optimum assignment is to write all the n! possible arrangements, evaluate their total cost and
select the assignment with minimum cost. Bust this method leads to a calculation problem of formidable
size even when the value of n is moderate. For n=10 the possible number of arrangements is 3268800.
2.2 Assignment Problem Structure and Solution
The structure of the Assignment problem is similar to a transportation problem, is as follows:
Jobs
1
2
...
n
1
c
1
11 c12 ... c1n
2
c
1
21 c21 ... c2n
. . . . .
Workers
. . . .
. . . .
n
c
1
n1 cn2 ... cnn
1 1 ... 1
97
MBA-H2040 Quantitative Techniques for Managers
The element cij represents the measure of effectiveness when ith person is assigned jth job. Assume that
the overall measure of effectiveness is to be minimized. The element xij represents the number of ith
individuals assigned to the jth job. Since ith person can be assigned only one job and jth job can be
assigned to only one person we have the following
xi1 + xi2 + ................ + xin = 1, where i = 1, 2, . . . . . . . , n
x1j + x2j + ................ + xnj = 1, where j = 1, 2, . . . . . . . , n
and the objective function is formulated as
Minimize c11x11 + c12x12 + ........... + cnnxnn
xij 0
The assignment problem is actually a special case of the transportation problem where m = n and
ai = bj = 1. However, it may be easily noted that any basic feasible solution of an assignment problem
contains (2n ? 1) variables of which (n ? 1) variables are zero. Because of this high degree of degeneracy
the usual computation techniques of a transportation problem become very inefficient. So, hat a separate
computation technique is necessary for the assignment problem.
The solution of the assignment problem is based on the following results:
"If a constant is added to every element of a row/column of the cost matrix of an assignment
problem the resulting assignment problem has the same optimum solution as the original assignment
problem and vice versa". ? This result may be used in two different methods to solve the assignment
problem. If in an assignment problem some cost elements cij are negative, we may have to convert them
into an equivalent assignment problem where all the cost elements are non-negative by adding a suitable
large constant to the cost elements of the relevant row or column, and then we look for a feasible
solution which has zero assignment cost after adding suitable constants to the cost elements of the
various rows and columns. Since it has been assumed that all the cost elements are non-negative, this
assignment must be optimum. On the basis of this principle a computational technique known as
Hungarian Method is developed. The Hungarian Method is discussed as follows.
Hungarian Method:
The Hungarian Method is discussed in the form of a series of computational steps as follows, when the
objective function is that of minimization type.
Step 1:
98
MBA-H2040 Quantitative Techniques for Managers
From the given problem, find out the cost table. Note that if the number of origins is not equal to the
number of destinations then a dummy origin or destination must be added.
Step 2:
In each row of the table find out the smallest cost element, subtract this smallest cost element from each
element in that row. So, that there will be at least one zero in each row of the new table. This new table
is known as First Reduced Cost Table.
Step 3:
In each column of the table find out the smallest cost element, subtract this smallest cost element from
each element in that column. As a result of this, each row and column has at least one zero element. This
new table is known as Second Reduced Cost Table.
Step 4:
Now determine an assignment as follows:
1. For each row or column with a single zero element cell that has not be assigned or
eliminated, box that zero element as an assigned cell.
2. For every zero that becomes assigned, cross out all other zeros in the same row and for
column.
3. If for a row and for a column there are two or more zero and one can't be chosen by
inspection, choose the assigned zero cell arbitrarily.
4. The above procedures may be repeated until every zero element cell is either assigned
(boxed) or crossed out.
Step 5:
An optimum assignment is found, if the number of assigned cells is equal to the number of rows (and
columns). In case we had chosen a zero cell arbitrarily, there may be an alternate optimum. If no
optimum solution is found i.e. some rows or columns without an assignment then go to Step 6.
Step 6:
Draw a set of lines equal to the number of assignments which has been made in Step 4, covering all the
zeros in the following manner
1. Mark check () to those rows where no assignment has been made.
2. Examine the checked () rows. If any zero element cel occurs in those rows, check () the
respective columns that contains those zeros.
3. Examine the checked () columns. If any assigned zero element occurs in those columns,
check () the respective rows that contain those assigned zeros.
4. The process may be repeated until now more rows or column can be checked.
5. Draw lines through all unchecked rows and through all checked columns.
Step 7:
Examine those elements that are not covered by a line. Choose the smallest of these elements and
subtract this smallest from all the elements that do not have a line through them.
99
MBA-H2040 Quantitative Techniques for Managers
Add this smallest element to every element that lies at the intersection of two lines. Then the
resulting matrix is a new revised cost table.
Example 2.1:
Problem
A work shop contains four persons available for work on the four jobs. Only one person can work on
any one job. The following table shows the cost of assigning each person to each job. The objective is to
assign person to jobs such that the total assignment cost is a minimum.
Jobs
1
2
3
4
20 25 22 28
A
15 18 23 17
Persons
B
19 17 21 24
C
25 23 24 24
D
Solution
As per the Hungarian Method
Step 1: The cost Table
Jobs
1
2
3
4
A
20 25 22 28
Persons
B
15 18 23 17
C
19 17 21 24
D
25 23 24 24
100
MBA-H2040 Quantitative Techniques for Managers
Step 2: Find the First Reduced Cost Table
Jobs
1
2
3
4
A 0 5 2 8
0 3 8 2
Persons
B
2 0 4 7
C
2 0 1 1
D
Step 3: Find the Second Reduced Cost Table
Jobs
1
2
3
4
A 0 5 1 7
0 3 7 1
Persons
B
2 0 3 6
C
2 0 0 0
D
Step 4: Determine an Assignment
By examine row A of the table in Step 3, we find that it has only one zero (cell A1) box this zero and
cross out all other zeros in the boxed column. In this way we can eliminate cell B1.
Now examine row C, we find that it has one zero (cell C2) box this zero and cross out (eliminate) the
zeros in the boxed column. This is how cell D2 gets eliminated.
There is one zero in the column 3. Therefore, cell D3 gets boxed and this enables us to eliminate cell
D4.
Therefore, we can box (assign) or cross out (eliminate) all zeros.
The resultant table is shown below:
101
MBA-H2040 Quantitative Techniques for Managers
Jobs
1
2
3
4
A
0 5 1 7
Persons
B
0 3 7 1
C
2 3 6
D
2 0 0
Step 5:
0
The solution obtained in Step 4 is not optimal. Because we were able to make three assignments when
four were required.
Step 6:
0
Cover all the zeros of the table shown in the Step 4 with three lines (since already we made three
assignments).
Check row B since it has no assignment. Note that row B has a zero in column 1, therefore check
column1. Then we check row A since it has a zero in column 1. Note that no other rows and columns are
checked. Now we may draw three lines through unchecked rows (row C and D) and the checked column
(column 1). This is shown in the table given below:
Jobs
1
2
3
4
0
A 5 1 7
Persons
B
0 3 7 1
C
2 3 6
D
2 0 0
102
0
MBA-H2040 Quantitative Techniques for Managers
Step 7:
Develop the new revised table.
Examine those elements that are not covered by a line in the table given in Step 6. Take the smallest
element in this case the smallest element is 1. Subtract this smallest element from the uncovered cells
and add 1 to elements (C1 and D1) that lie at the intersection of two lines. Finally, we get the new
revised cost table, which is shown below:
Jobs
1
2
3
4
A 0 4 0 6
Persons
B
0 2 6 0
C
3 0 3 6
D
3 0 0 0
Step 8:
Now, go to Step 4 and repeat the procedure until we arrive at an optimal solution (assignment).
Step 9:
Determine an assignment
Examine each of the four rows in the table given in Step 7, we may find that it is only row C which has
only one zero box this cell C2 and cross out D2.
Note that all the remaining rows and columns have two zeros. Choose a zero arbitrarily, say A1 and box
this cell so that the cells A3 and B1 get eliminated.
Now row B (cell B4) and column 3 (cell D4) has one zero box these cells so that cell D4 is eliminated.
Thus, all the zeros are either boxed or eliminated. This is shown in the following table
103
MBA-H2040 Quantitative Techniques for Managers
Jobs
1
2
3
4
0
A 4 0 6
Persons
B
0 2 6
C
3 3 6
0
D
3 0 0
Since the number of assignments equal to the number of rows (columns), the assignment shown
in the above tale is optimal.
0
The total cost of assignment is: 78 that is A1 + B4 + C2 + D3
20 + 17 + 17 + 24 = 78
2.3 Unbalanced Assignment Problem
0
In the previous section we assumed that the number of persons to be assigned and the number of jobs
were same. Such kind of assignment problem is called as balanced assignment problem. Suppose if
the number of person is different from the number of jobs then the assignment problem is called as
unbalanced.
If the number of jobs is less than the number of persons, some of them can't be assigned any job. So that
we have to introduce on or more dummy jobs of zero duration to make the unbalanced assignment
problem into balanced assignment problem. This balanced assignment problem can be solved by using
the Hungarian Method as discussed in the previous section. The persons to whom the dummy jobs are
assigned are left out of assignment.
Similarly, if the number of persons is less than number of jobs then we have introduce one or more
dummy persons with zero duration to modify the unbalanced into balanced and then the problem is
solved using the Hungarian Method. Here the jobs assigned to the dummy persons are left out.
104
MBA-H2040 Quantitative Techniques for Managers
Example 2.2:
Solve the following unbalanced assignment problem of minimizing the total time for performing all the
jobs.
Jobs
1 2 3 4 5
A 5 2 4 2 5
B
2 4 7 6 6
Workers
C
6 7 5 8 7
D
5 2 3 3 4
E
8 3 7 8 6
F
3 6 3 5 7
Solution
In this problem the number of jobs is less than the number of workers so we have to introduce a dummy
job with zero duration.
105
MBA-H2040 Quantitative Techniques for Managers
The revised assignment problem is as follows:
Jobs
1 2 3 4 5 6
A 5 2 4 2 5 0
B
2 4 7 6 6 0
Workers
C
6 7 5 8 7 0
D
5 2 3 3 4 0
E
8 3 7 8 6 0
F
3 6 3 5 7 0
Now the problem becomes balanced one since the number of workers is equal to the number jobs. So
that the problem can be solved using Hungarian Method.
Step 1: The cost Table
Jobs
1 2 3 4 5 6
A 5 2 4 2 5 0
Workers
B
2 4 7 6 6 0
C
6 7 5 8 7 0
D
5 2 3 3 4 0
E
8 3 7 8 6 0
F
3 6 3 5 7 0
106
MBA-H2040 Quantitative Techniques for Managers
Step 2: Find the First Reduced Cost Table
Jobs
1 2 3 4 5 6
A 5 2 4 2 5 0
B
2 4 7 6 6 0
Workers
C
6 7 5 8 7 0
D
5 2 3 3 4 0
E
8 3 7 8 6 0
F
3 6 3 5 7 0
Step 3: Find the Second Reduced Cost Table
Jobs
1 2 3 4 5 6
A 3 0 1 0 1 0
B
0 2 4 4 2 0
Workers
C
4 5 2 6 3 0
D
3 0 0 1 0 0
E
6 1 4 6 2 0
F
1 4 0 3 3 0
107
MBA-H2040 Quantitative Techniques for Managers
Step 4: Determine an Assignment
By using the Hungarian Method the assignment is made as follows:
Jobs
1 2 3 4 5 6
A 3 0 1
0 1 0
B
2 4 4 2 0
Workers
C
4
0
5 2 6 3
D
3 0 0 1 0
E
6 1 4 6 2 0
0
F
1 4 3 3 0
0
Step 5:
The solution obtained in Step 4 is not optimal. Because we were able to make five assignments
when six were required.
Step 6:
Cover all the zeros of the table shown in the Step 4 with five lines (since already we made five
assignments).
Check row E since it has no assignment. Not
0 e that row B has a zero in column 6, therefore check
column6. Then we check row C since it has a zero in column 6. Note that no other rows and columns are
checked. Now we may draw five lines through unchecked rows (row A, B, D and F) and the checked
column (column 6). This is shown in the table given below:
108
MBA-H2040 Quantitative Techniques for Managers
Jobs
1 2 3 4 5 6
A 3 0 1
0 1 0
B
2 4 4 2 0
Workers
C
4
0
5 2 6 3
D
3 0 0 1 0
E
6 1 4 6 2 0
0
F
1 4 3 3 0
0
Step 7:
Develop the new revised table.
Examine those elements that are not covered by a line in the table given in Step 6. Take the
smallest element in this case the smallest element is 1. Subtract this smallest element from the
uncovered cells and add 1 to elements (A6, B6, D6 and F6) that lie at the intersection of two lines.
Finally, we get the new revised cost table, which is shown below:
Jobs
1 2 3 4 5 6
0
109
MBA-H2040 Quantitative Techniques for Managers
A 3 0 1 0 1 1
B
0 2 4 4 2 1
Workers
C
3 4 1 5 2 0
D
3 0 0 1 0 1
E
5 0 3 5 1 0
F
1 4 0 3 3 1
Step 8:
Now, go to Step 4 and repeat the procedure until we arrive at an optimal solution (assignment).
Step 9:
Determine an assignment
Jobs
1 2 3 4 5 6
0
A 3 0 1 1 1
B
2 4 4 2 1
Workers
C
3
0 4 1 5 2
D
3 0 0 1 1
E
5 3 5 1 0 0
F
1 4 3 3 1
0
110
MBA-H2040 Quantitative Techniques for Managers
Since the number of assignments equal to the number of rows (columns), the assignment shown in the
above tale is optimal.
Thus, the worker A is assigned to Job4, worker B is assigned to job 1, worker C is assigned to
job 6, worker D is assigned to job 5, worker E is assigned to job 2, and worker F is assigned to job 3.
Since the Job 6 is dummy so that worker C can't be assigned.
The total minimum time is: 14 that is A4 + B1 + D5 + E2 + F3
2 + 2 + 4 + 3 + 3 = 14
Example 2.3:
A marketing company wants to assign three employees viz. A, B, and C to four offices located at
W, X, Y and Z respectively. The assignment cost for this purpose is given in following table.
Offices
W
X
Y Z
A
160 220 240 200
Employees
B
100 320 260 160
C
100 200 460 250
Solution
Since the problem has fewer employees than offices so that we have introduce a dummy
employee with zero cost of assignment.
The revised problem is as follows:
Offices
W
X
Y Z
111
MBA-H2040 Quantitative Techniques for Managers
A
160 220 240 200
Employees
B
100 320 260 160
C
100 200 460 250
D
0 0 0 0
Now the problem becomes balanced. This can be solved by using Hungarian Method as in the
case of Example 2.2. Thus as per the Hungarian Method the assignment made as follows:
Employee A is assigned to Office X, Employee B is assigned to Office Z, Employee C is
assigned to Office W and Employee D is assigned to Office Y. Note that D is empty so that no one is
assigned to Office Y.
The minimum cost of assignment is: 220 + 160 + 100 = 480
2.4 Infeasible Assignment Problem
Sometimes it is possible a particular person is incapable of performing certain job or a specific
job can't be performed on a particular machine. In this case the solution of the problem takes into
account of these restrictions so that the infeasible assignment can be avoided.
The infeasible assignment can be avoided by assigning a very high cost to the cells where
assignments are restricted or prohibited. This is explained in the following Example 2.4.
Example 2.4:
A computer centre has five jobs to be done and has five computer machines to perform them.
The cost of processing of each job on any machine is shown in the table below.
Jobs
1 2 3 4 5
112
MBA-H2040 Quantitative Techniques for Managers
1
70 30 X 60 30
Computer 2
X 70 50 30 30
Machines
3
60 X 50 70 60
4
60 70 20 40 X
5
30 30 40 X 70
Because of specific job requirement and machine configurations certain jobs can't be done on
certain machines. These have been shown by X in the cost table. The assignment of jobs to the machines
must be done on a one to one basis. The objective here is to assign the jobs to the available machines so
as to minimize the total cost without violating the restrictions as mentioned above.
Solution
Step 1: The cost Table
Because certain jobs cannot be done on certain machines we assign a high cost say for example
500 to these cells i.e. cells with X and modify the cost table. The revised assignment problem is as
follows:
Jobs
1 2 3 4 5
1
70 30 500 60 30
Computer 2
500 70 50 30 30
Machines
3
60 500 50 70 60
4
60 70 20 40 500
5
30 30 40 500 70
113
MBA-H2040 Quantitative Techniques for Managers
Now we can solve this problem using Hungarian Method as discussed in the previous sections.
Step 2: Find the First Reduced Cost Table
Jobs
1 2 3 4 5
1
40 0 470 30 0
Computer 2
470 40 20 0 0
Machines
3
10 450 0 20 10
4
40 50 0 20 480
5
0 0 10 470 40
Step 3: Find the Second Reduced Cost Table
Jobs
1 2 3 4 5
114
MBA-H2040 Quantitative Techniques for Managers
1
40 0 470 30 0
Computer 2
470 40 20 0 0
Machines
3
10 450 0 20 10
4
40 50 0 20 480
5
0 0 10 470 40
Step 4: Determine an Assignment
Jobs
1 2 3 4 5
1
40 0 470 30 0
Computer 2
470 40 20 0
Machines
0
3
10 450 20 10
4
40 50 0 20 480
5
0 10
0
470 40
Step 5:
The solution obtained in Step 4 is not optimal. Because we were able to make four assignments
when five were required.
Step 6:
Cover all the zeros of the table shown in the Step 4 with four lines (since already we made four
assignments).
0
115
MBA-H2040 Quantitative Techniques for Managers
Check row 4 since it has no assignment. Note that row 4 has a zero in column 3, therefore check
column3. Then we check row 3 since it has a zero in column 3. Note that no other rows and columns are
checked. Now we may draw four lines through unchecked rows (row 1, 2, 3 and 5) and the checked
column (column 3). This is shown in the table given below:
Jobs
1 2 3 4 5
1
40 0 470 30 0
Computer 2
470 40 20 0
Machines
0
3
10 450 20 10
4
40 50 0 20 480
5
0 10
0
470 40
Step 7:
Develop the new revised table.
Examine those elements that are not covered by a line in the table given in Step 6. Take the
smallest element in this case the smallest element is 10. Subtract this smallest element from the
uncovered cells and add 1 to elements (A6, B6, D6 and F6) that lie at the intersection of two lines.
Finally, we get the new revised cost table, which is shown below:
Jobs
1 2 3 4 5
0
1
40 0 471 30 0
Computer 2
470 40 21 0 0
Machines
3
0 440 0 10 0
4
30 40 0 10 470
5
0 0 11 470 40
Step 8:
Now, go to Step 4 and repeat the procedure until we arrive at an optimal solution (assignment).
116
MBA-H2040 Quantitative Techniques for Managers
Step 9:
Determine an assignment
Jobs
1 2 3 4 5
1
40 0 471 30
0
Computer 2
470 40 21 0
Machines
3
440 0 10 0
4
30 40 10 470
0
5
0 11 470 40
Since the number of assignments equal to the number of rows (columns), the assignment shown
in the above tale is optimal.
0
Thus, the Machine1 is assigned to Job5, Machine 2 is assigned to job4, Machine3 is assigned to
job1, Machine4 is assigned to job3 and Machine5 is assigned to job2.
The minimum assignment cost is: 170
0
2.5 Maximization in an Assignment Problem
There are situations where certain facilities have to be assigned to a number of jobs so as to
maximize the overall performance of the assignment. In such cases the problem can be converted into a
minimization problem and can be solved by using Hungarian Method. Here the conversion of
maximization problem into a minimization can be done by subtracting all the elements of the cost table
from the highest value of that table.
Example 2.5:
Consider the problem of five different machines can do any of the required five jobs with
different profits resulting from each assignment as illustrated below:
Machines
1 2 3 4 5
117
MBA-H2040 Quantitative Techniques for Managers
1
40 47 50 38 50
2
50 34 37 31 46
Jobs
3
50 42 43 40 45
4
35 48 50 46 46
5
38 72 51 51 49
Find out the maximum profit through optimal assignment.
Solution
This is a maximization problem, so that first we have to find out the highest value in the table
and subtract all the values from the highest value. In this case the highest value is 72.
The new revised table is given below:
Machines
1 2 3 4 5
1
32 35 22 34 22
2
22 38 35 41 26
Jobs
3
22 30 29 32 27
4
37 24 22 26 26
5
34 0 21 21 23
This can be solved by using the Hungarian Method.
By solving this, we obtain the solution is as follows:
118
MBA-H2040 Quantitative Techniques for Managers
Jobs
Machines
1
3
2
5
3
1
4
4
5
2
The maximum profit through this assignment is: 264
2.6 Crew Assignment Problem
The crew assignment problem is explained with the help of the following problem
Problem:
A trip from Chennai to Coimbatore takes six hours by bus. A typical time table of the bus service
in both the direction is given in the Table 1. The cost of providing this service by the company based on
the time spent by the bus crew i.e. driver and conductor away from their places in addition to service
times. The company has five crews. The condition here is that every crew should be provided with more
than 4 hours of rest before the return trip again and should not wait for more than 24 hours for the return
trip. Also the company has guest house facilities for the crew of Chennai as well as at Coimbatore.
Find which line of service is connected with which other line so as to reduce the waiting time to
the minimum.
Table 1
Departure from Route Number
Arrival at
Arrival at
Route Number Departure from
Chennai
Coimbatore
Chennai
Coimbatore
06.00
1
12.00
11.30
a
05.30
07.30
2
13.30
15.00
b
09.00
11.30
3
17.30
21.00
c
15.00
19.00
4
01.00
00.30
d
18.30
00.30
5
06.30
06.00
e
00.00
Solution
For each line the service time is constant so that it does not include directly in the computation. Suppose
if the entire crew resides at Chennai then the waiting times in hours at Coimbatore for different route
connections are given below in Table 2.
If route 1 is combined with route a, the crew after arriving at Coimbatore at 12 Noon start at 5.30
next morning. Thus the waiting time is 17.5 hours. Some of the assignments are infeasible. Route c
119
MBA-H2040 Quantitative Techniques for Managers
leaves Coimbatore at 15.00 hours. Thus the crew of route 1 reaching Coimbatore at 12 Noon are unable
to take the minimum stipulated rest of four hours if they are asked to leave by route c. Hence 1-c is an
infeasible assignment. Thus it cost is M (a large positive number).
Table 2
Route a b c d e
1
17.5 21 M 6.5 12
2
16 19.5 M 5 10.5
3
12 15.5 21.5 M 6.5
4
4.5 8 4 17.5 23
5
23 M 8.5 12 17.5
Similarly, if the crews are assumed to reside at Coimbatore then the waiting times of the crew in
hours at Chennai for different route combinations are given below in Table 3.
Table 3
Route a b c d e
1
18.5 15 9 5.5 M
2
20 16.5 10.5 7 M
3
M 20.5 14.5 11 5.5
4
7.5 M 22 18.5 13
5
13 9.5 M M 18.5
120
MBA-H2040 Quantitative Techniques for Managers
Suppose, if the crew can be instructed to reside either at Chennai or at Coimbatore, minimum
waiting time from the above operation can be computed for different route combination by choosing the
minimum of the two waiting times (shown in the Table 2 and Table 3). This is given in the following
Table 4.
Table 4
Route a b c d e
1
17.5* 15 9 5.5 12*
2
16* 16.5 10.5 5* 10.5*
3
12* 15.5* 14.5 11 5.5
4
4.5* 8* 14* 17.5* 13
5
13 9.5 8.5* 12* 17.5*
Note: The asterisk marked waiting times denotes that the crew are based at Chennai; otherwise they are
based at Coimbatore.
Now we can solve the assignment problem (presented in Table 4) using Hungarian Method.
Step 1: Cost Table (Table 5)
Table 5
Route a b c d e
1
17.5* 15 9 5.5 12*
2
16* 16.5 10.5 5* 10.5*
3
12* 15.5* 14.5 11 5.5
4
4.5* 8* 14* 17.5* 13
5
13 9.5 8.5* 12* 17.5*
121
MBA-H2040 Quantitative Techniques for Managers
Step 2: Find the First Reduced cost table (Table 6)
Table 6
Route a b c d e
1
12 9.5 3.5 0 6.5
2
11 11.5 5.5 0 5.5
3
6.5 10 9 5.5 0
4
0 3.5 9.5 13 8.5
5
4.5 1 0 3.5 9
Step 3: Find the Second Reduced cost table (Table 7)
Table 7
Route a b c d e
1
12 8.5 3.5 0 6.5
2
11 10.5 5.5 0 5.5
3
6.5 9 9 5.5 0
4
0 2.5 9.5 13 8.5
5
4.5 0 0 3.5 9
122
MBA-H2040 Quantitative Techniques for Managers
Step 4: Determine an Assignment (Table 8)
Table 8
Route a b c d e
0
1
12 8.5 3.5 6.5
2
11 10.5 5.5 0 5.5
3
6.5 9 9 5.5
4
2.5 9.5 13 8.5
5
4.5 0 3.5 9 0
Step 5: The solution obtained in Step 4 is not optimal since the number of assignments are less than the
number of rows (columns).
Step 6: Check () row 2 since it
0 has no assignment and note that row 2 has a zero in column d, therefore
check () column d also. Then check row 1 since it has zero in column d. Draw the lines through the
unchecked rows and checked column using 4 lines (only 4 assignments are made). This is shown in
Table 9.
Table 9
Route a b
0 c d e
0
1
12 8.5 3.5 6.5
2
11 10.5 5.5 0 5.5
3
6.5 9 9 5.5
4
2.5 9.5 13 8.5
5
4.5 0 3.5 9 0
123
0
MBA-H2040 Quantitative Techniques for Managers
Step 7: Develop a new revised table (Table 10)
Take the smallest element from the elements not covered by the lines in this case 3.5 is the
smallest element. Subtract all the uncovered elements from 3.5 and add 3.5 to the elements lie at the
intersection of two lines (cells 3d, 4d and 5d). The new revised table is presented in Table 10.
Table 10
Route a b c d e
1
8.5 5 0 0 3
2
7.5 7 2 0 2
3
6.5 9 9 9 0
4
0 2.5 9.5 16.5 8.5
5
4.5 0 0 7 9
Step 8: Go to Step 4 and repeat the procedure until an optimal solution if arrived.
Step 9: Determine an Assignment (Table 11)
Table 11
Route a b c d e
124
MBA-H2040 Quantitative Techniques for Managers
1
8.5 5
0 0 3
2
7.5 7 2 2
3
6.5 9 9 9
0
4
2.5 9.5 16.5 8.5
5
4.5 0 7 9 0
The assignment
illustrated in the above
Table 11 is optimal since the number of assignments is equal to the number of rows (columns).
Thus, the routes to be prepared to achieve the minimum waiting time are as follows:
1
0 ?
c, 2 ? d, 3 ? e, 4 ? a and 5 ? b
By referring Table 5, we can obtain the waiting times of these assignments as well as the residence
(guest house) of the crews. This is presented in the following Table 12.
Table 12
Routes
Residence of the Crew
Waiting Time
0
1 ? c
Coimbatore
9
2 ? d
Chennai
5
3 ? e
Coimbatore
5.5
4 ? a
Chennai
4.5
5 - b
Coimbatore
9.5
2.7 Summary
The assignment problem is used for the allocation of a number of persons to a number of jobs so that the
total time of completion is minimized. The assignment problem is said to be balanced if it has equal
number of person and jobs to be assigned. If the number of persons (jobs) is different from the number
of jobs (persons) then the problem is said to be unbalanced. An unbalanced assignment problem can be
solved by converting into a balanced assignment problem. The conversion is done by introducing
dummy person or a dummy job with zero cost.
Because of the special structure of the assignment problem, it is solved by using a special
method known as Hungarian Method.
2.8 Key Terms
Cost Table: The completion time or cost corresponding to every assignment is written down in a table
form if referred as a cost table.
125
MBA-H2040 Quantitative Techniques for Managers
Hungarian Method: is a technique of solving assignment problems.
Assignment Problem: is a special kind of linear programming problem where the objective is to
minimize the assignment cost or time.
Balanced Assignment Problem: is an assignment problem where the number of persons equal to the
number of jobs.
Unbalanced Assignment Problem: is an assignment problem where the number of jobs is not equal to
the number of persons.
Infeasible Assignment Problem: is an assignment problem where a particular person is unable to
perform a particular job or certain job cannot be done by certain machines.
2.9 Self Assessment Questions
Q1. A tourist company owns a one car in each of the five locations viz. L1, L2, L3, L4, L5 and a
passengers in each of the five cities C1, C2, C3, C4, C5 respectively. The following table shows the
distant between the locations and cities in kilometer. How should be cars be assigned to the passengers
so as to minimize the total distance covered.
Cities
C1
C 2
C3 C4 C5
L1 120 110 115 30 36
Locations
L2
125 100 95 30 16
L3
155 90 135 60 50
L4
160 140 150 60 60
L5
190 155 165 90 85
Q2. Solve the following assignment problem
1 2 3 4 5
126
MBA-H2040 Quantitative Techniques for Managers
1
Rs.3 Rs.8 Rs.2 Rs.10 Rs.3
2
Rs.8 Rs.7 Rs.2 Rs.9 Rs.7
3
Rs.6 Rs.4 Rs.2 Rs.7 Rs.5
4
Rs.8 Rs.4 Rs.2 Rs.3 Rs.5
5
Rs.9 Rs.10 Rs.6 Rs.9 Rs.10
Q3. Work out the various steps of the solution of the Example 2.3.
Q4. A steel company has five jobs to be done and has five softening machines to do them. The cost of
softening each job on any machine is given in the following cost matrix. The assignment of jobs to
machines must be done on a one to one basis. Here is the objective is to assign the jobs to the machines
so as to minimize the total assignment cost without violating the restrictions.
Jobs
1 2 3 4 5
1
80 30 X 70 30
Softening 2
70 X 60 40 30
Machines
3
X 80 60 80 70
4
70 80 30 50 X
5
30 30 50 X 80
Q5. Work out the various steps of the solution of the problem presented in Example 2.5.
Q6. A marketing manager wants to assign salesman to four cities. He has four salesmen of varying
experience. The possible profit for each salesman in each city is given in the following table. Find out an
assignment which maximizes the profit.
127
MBA-H2040 Quantitative Techniques for Managers
Cities
1 2 3 4
1
25 27 28 38
Salesmen 2
28 34 29 40
3
35 24 32 33
4
24 32 25 28
Q7. Shiva's three wife, Rani, Brinda, and Fathima want to earn some money to take care of personal
expenses during a school trip to the local beach. Mr. Shiva has chosen three chores for his wife:
washing, cooking, sweeping the cars. Mr. Shiva asked them to submit bids for what they feel was a fair
pay for each of the three chores. The three wife of Shiva accept his decision. The following table
summarizes the bid received.
Chores
Washing Cooking Sweeping
1 2 3
Rani 25 18 17
Wife's Brinda
17 25 15
Fathima
18 22 32
Q8. Solve the following problem
Office
O1 O2 O3 O4
128
MBA-H2040 Quantitative Techniques for Managers
E1
2600 3200 3400 3000
Employees E2
2000 4200 3600 2600
E3
2000 3000 5600 4000
Q9. The railway operates seven days a week has a time table shown in the following table. Crews
(Driver and Guard) must have minimum rest of six hours between trans. Prepare the combination of
trains that minimizes waiting time away from the city. Note that for any given combination the crew will
be based at the city that results in the smaller waiting time and also find out for each combination the
city where the crew should be based at.
Train No.
Departure at
Arrival at
Train No.
Departure at
Arrival at
Bangalore
Chennai
Chennai
Bangalore
101
7 AM
9 AM
201
9 AM
11 AM
102
9 AM
11 AM
202
10 AM
12 Noon
103
1.30 PM
3.30 PM
203
3.30 PM
5.30 PM
104
7.30 PM
9.30 PM
204
8 PM
10 PM
2.10 Key Solutions
Q1. L1 ? C1, L2 ? C3, L3 ? C2, L4 _ C4, L5 ? C5 and
Minimum Distance is: 450
Q2. 1 ? 5, 2 ? 3, 3 ? 2, 4 ? 4, 5 ? 1 and
Minimum Cost is: Rs.21
Q4. 1 ? 2, 2 ? 4, 3 ? 3, 4 ? 4, 5 ? 1 and
Minimum Assignment Cost is:
Q6. 1 - 1, 2 ? 4, 3 ? 3, 4 ? 2 and
Maximum Profit is: 139
Q7. Rani ? Cooking, Brinda ? Sweeping, Fathima ? Washing and
Minimum Bids Rate is: 51
Q8. E1 ? O2, E2 ? O4, E3 ? O1
Since E4 is empty, Office O3 cannot be assigned to any one.
Minimum Cost is: 7800
Q9.
Trains
Cities
201 ? 103
Bangalore
202 ? 104
Chennai
129
MBA-H2040 Quantitative Techniques for Managers
203 ? 101
Bangalore
204 ? 102
Bangalore
2.11 Further References
Hamdy A Taha, 1999. Introduction to Operations Research, PHI Limited, New Delhi.
Cooper, L and D. Steinberg, 1974. Methods and Applications of Linear Programmings, Saunders,
Philadelphia, USA.
Mustafi, C.K. 1988. Operations Research Methods and Practices, Wiley Eastern Limited, New Delhi.
130
MBA-H2040 Quantitative Techniques for Managers
UNIT II
3 INTRODUCTION TO INVENTORY MANAGEMENT
ON
S
S
E
L
LESSON STRUCTURE
3.1 Introduction
3.2 Objectives of Inventory
3.3 Inventory is an Essential Requirement
3.4 Basic Functions of Inventory
3.5 Types of Inventory
3.6 Factors Affecting Inventory
3.7 Summary
3.8 Key Terms
Objectives
3.9 Self Assessment Questions
After Studying this lesson, you should be able
3.10 Further References
to:
Understand what is inventory
Describe various inventory concepts
Describe the objectives of inventory
Explain the functions of inventory
Describe requirements of inventory
Explain different types of inventory
Describe different factors affecting
inventory
131
MBA-H2040 Quantitative Techniques for Managers
3.1 Introduction
Simply inventory is a stock of physical assets. The physical assets have some economic value, which
can be either in the form of material, men or money. Inventory is also called as an idle resource as long
as it is not utilized. Inventory may be regarded as those goods which are procured, stored and used for
day to day functioning of the organization.
Inventory can be in the form of physical resource such as raw materials, semi-finished goods
used in the process of production, finished goods which are ready for delivery to the consumers, human
resources, or financial resources such as working capital etc.
Inventories means measures of power and wealth of a nation or of an individual during centuries
ago. That is a business man or a nation's wealth and power were assessed in terms of grammes of gold,
heads of cattle, quintals of rice etc.
In recent past, inventories mean measure of business failure. Therefore, businessmen have
started to put more emphasis on the liquidity of assets as inventories, until fast turnover has become a
goal to be pursues for its own sake.
Today inventories are viewed as a large potential risk rather than as a measure of wealth due to
the fast developments and changes in product life. The concept of inventories at present has necessitated
the use of scientific techniques in the inventory management called as inventory control.
Thus, inventory control is the technique of maintaining stock items at desired levels. In other
words, inventory control is the means by which material of the correct quality and quantity is made
available as and when it is needed with due regard to economy in the holding cost, ordering costs, setup
costs, production costs, purchase costs and working capital.
Inventory Management answers two questions viz. How much to order? and when to order?
Management scientist insisting that the inventory is an very essential requirement. Why? This is
illustrated in the next section with the help of materials conversion process diagram.
3.2 Objectives of Inventory
Inventory has the following main objectives:
To supply the raw material, sub-assemblies, semi-finished goods, finished goods, etc. to
its users as per their requirements at right time and at right price.
To maintain the minimum level of waste, surplus, inactive, scrap and obsolete items.
To minimize the inventory costs such as holding cost, replacement cost, breakdown cost
and shortage cost.
To maximize the efficiency in production and distribution.
To maintain the overall inventory investment at the lowest level.
To treat inventory as investment which is risky? For some items, investment may lead to
higher profits and for others less profit.
132
MBA-H2040 Quantitative Techniques for Managers
3.3 Inventory is an Essential Requirement
Inventory is a part and parcel of every facet of business life. Without inventory no business activity can
be performed, whether it being a manufacturing organization or service organization such as libraries,
banks, hospitals etc. Irrespective of the specific organization, inventories are reflected by way of a
conversion process of inputs to outputs. The conversion process is illustrated in the figure 3.1 as given
below.
From the figure 3.1 we can see that there may be stock pints at three stages viz. Input,
Conversion Process and Output. The socks at input are called raw materials whereas the stocks at the
output are called products. The stocks at the conversion process may be called finished or semi-finished
goods or sometimes may be raw material depending on the requirement of the product at conversion
process, where the input and output are based on the market situations of uncertainty, it becomes
physically impossible and economically impractical for each stock item to arrive exactly where it is
required and when it is required.
Random Fluctuations
Conversion
Inputs
Inventory
process Inventory
Sock-points
Comparing Actual Material
Stock levels to planned levels
Material Sock-Points
Land
Outputs
Labour
Capital
Management
Feed Back
Product Inventory
Stock-points
Corrective action: More or less Stocks
Fig: 3.1 Materials Conversion Process
Even it is physically possible to deliver the stock when it is required, it costs more expensive. This is the
basic reason for carrying the inventory. Thus, inventories play an essential and pervasive role in any
organization because they make it possible:
133
MBA-H2040 Quantitative Techniques for Managers
To meet unexpected demand
To achieve return on investment
To order largest quantities of goods, components or materials from the suppliers at
advantageous prices
To provide reasonable customer service through supplying most of the requirements from
stock without delay
To avoid economically impractical and physically impossible delivering/getting right
amount of stock at right time of required
To maintain more work force levels
To facilitate economic production runs
To advantage of shipping economies
To smooth seasonal or critical demand
To facilitate the intermittent production of several products on the same facility
To make effective utilization of space and capital
To meet variations in customer demand
To take the advantage of price discount
To hedge against price increases
To discount quantity
3.4 Basic Functions of Inventory
The important basic function of inventory is
Increase the profitability- through manufacturing and marketing support. But zero
inventory manufacturing- distribution system is not practically possible, so it is
important to remember that each rupee invested in inventory should achieve a specific
goal. The other inventory basic functions are
Geographical Specialization
Decoupling
Balancing supply and demand and
Safety stock
Inventory Investment Alternative
Investment is most important and major part of asset, which should be required to produce a minimum
investment return. The MEC (Marginal Efficiency of Capital) concept holds that an organization should
invest in those alternatives that produce a higher investment return than capital to borrow. The following
figure 3.2 shows that investment alternative A on the MEC curve is acceptable.
134
MBA-H2040 Quantitative Techniques for Managers
100
)
t ( %
A
c
c
e
pt
e
d I
n
v
entory Investment
e
n
s
t
m
ve
A
Cost of Capital
I
n
n
O
r
n
t
u
Re
Reject Inventory Investment
0
20
40
60
80
100
Total Investment Alternatives (in %)
Fig 3.2 Typical MEC curve
The curve shows that about 20% of the inventory investment alternatives will produce a return
on investment above the capital cost.
Geographical Specialization
Another basic inventory function is to allow the geographical specialization individual operating units.
There is a considerable distance between the economical manufacturing location and demand areas due
to factors of production such as raw material, labour, water, power. So that the goods from various
manufacturing locations are collected at a simple warehouse or plant to assemble in final product or to
offer consumers a single mixed product shipment. This also provides economic specialization between
manufacturing and distribution units/locations of an organization.
Decoupling
The provision of maximum efficiency of operations within a single facility is also one of the important
basic functions of the inventory. This is achieved by decoupling, which is done by breaking operations
apart so that one operation(s) supply is independent of another(s) supply.
The decoupling function serves in two ways of purposes, they are
1. Inventories are needed to reduce the dependencies among successive stages of operations so that
shortage of materials, breakdowns or other production fluctuations at one stage do not cause later stage
to shut down. This is illustrated in the following figure 3.3 in an engineering unit.
Raw
D i
e c a
s ti n
g
D r il l i
n
g
D e -
b
ur
ni
n g
P a c
ki n
g Finished
Material
Products
Inventory
Inventory
135
MBA-H2040 Quantitative Techniques for Managers
Fig: 3.3 Decoupling of operation using inventory
The figure shows that the de-burning, packing could continue to operate from inventories should
die-casting and drilling be shut down or they can be decoupled from the production processes that
precede them.
2. One organizational unit schedules its operations independently of another organizational unit.
For example: Consider an automobile organization, here assembly process can be schedule separately
from engine built up operation, and each can be decoupled from final automobile assembly operations
through in process inventories.
Supply and Demand Balancing
The function of Balancing concerns elapsed time between manufacturing and using the product.
Balancing inventories exist to reconcile supply with demand. The most noticeable example of balancing
is seasonal production and year round usage like sugar, rice, woolen textiles, etc. Thus the investment of
balancing inventories links the economies of manufacturing with variations of usage.
Safety Stock
The safety stock also called as buffer stock. The function of safety stock concerns short range variations
in either replacement or demand. Determination of the safety stock size requires a great deal of
inventory planning. Safety stock provides protection against two types of uncertainty, they are
1. Sales in excess of forecast during the replenishment period
2. Delays in replenishment
Thus, the inventories committed to safety stocks denote the greatest potential for improved
performances. There are different techniques are available to develop safety stocks.
3.5 Types of Inventory
Inventory may be classified into manufacturing, service and control aspects, which is illustrated in the
figure 3.4 as given below:
136
MBA-H2040 Quantitative Techniques for Managers
Raw Materials/Production Inventory
Work-in-Process Inventory
Manufacturing M.R.O Inventory
Aspect
Finished Goods Inventory
Miscellaneous Inventory
Lot Size Stocks
y
t
or
n
ve
Anticipation Stocks
I
n
Inventory Service Aspect
e
s
of
yp
T
Fluctuation Stocks
.
4
3
i
g:
F
Risk Stocks
A-Items Inventory
B-Items Inventory
Control Aspect
C-Items Inventory
137
MBA-H2040 Quantitative Techniques for Managers
Each inventory type is discussed in detail are as follows:
Raw Material/Manufacturing Inventory
There are five types of Manufacturing Inventory, they are
Production Inventory
Items going to final product such a raw materials, sub-assemblies purchased from outside are called
production inventory.
Work-in-Process Inventory
The items in the form of semi-finished or products at different stage of production process are known as
work-in-process inventory.
M.R.O. Inventory
Maintenance, Repair and Operating supplies such as spare parts and consumable stores, which do not go
into final product but are consumed during the production process.
Finished Goods Inventory
Finished Goods Inventory includes the products ready for dispatch to the consumers or
distributors/retailers.
Miscellaneous Inventory
Items excluding those mentioned above, such as waste, scrap, obsolete, and un-saleable items arising
from the main production process, stationery used in the office and other items required by office,
factory and other departments, etc. are called miscellaneous inventory.
Service Inventory
The service inventory can be classified into four types, they are:
Lot Size Stocks
Lot size means purchasing in lots. The reasons for this is to
Obtain quantity discounts
Minimize receiving and handling costs
Reduce purchase and transport costs
For example: It would be uneconomical for a textile factory to buy cotton everyday rather than in
bulk during the cotton season.
Anticipation Stocks
Anticipation stocks are kept to meet predictable changes in demand or in availability of raw materials.
138
MBA-H2040 Quantitative Techniques for Managers
For example: The purchase of potatoes in the potato season for sale of roots preservation products
throughout the year.
Fluctuation Stocks
Fluctuation stocks are carried to ensure ready supplies to consumers in the face of irregular fluctuations
in their demands
Risk Stocks
Risk stocks are the items required to ensure that there is no risk of complete production breakdown. Risk
stocks are critical and important for production.
Control Inventory
A good way of examining an inventory control is: to make ABC classification, which is also known as
ABC analysis. ABC analysis means the "control" will be "Always Better" if we start with the ABC
classification of inventory.
The ABC concepts classifies inventories into three groups in terms of percentage of total value
and percentage of number of inventory items, this is illustrated in the figure 3.5 and 3.6 as given below.
Values
e
u
A
A
10% 70%
Val
of
t
age
A items
20% 20%
B
B
e
r
c
e
n
P
B i t e
m
s 70% 10%
C
i t
e
m s
C
C
Percentage of Numbers Numbers
Fig: 3.5 ABC Classification (Frequency Form)
Fig: 3.6 ABC Classification (Tabular Form)
The three groups of inventory items are called A-items group, B-items group, C-items group,
which are explained as follows:
A-items Group: This constitutes 10% of the total number of inventory items and 70% of total money
value for all the items.
139
MBA-H2040 Quantitative Techniques for Managers
B-items Group: This constitutes 20% of the total number of inventory items and 20 % of total money
value for all the items.
C-items Group: This constitutes 70% of the total number of inventory items and 10 % of total money
value for all the items. This is just opposite of A-items group.
The ABC classification provides us clear indication for setting properties of control to the items,
and A-class item receive the importance first in every respect such as tight control, more security, and
high operating doctrine of the inventory control.
The coupling of ABC classification with VED classification enhances the inventory control
efficiency. VED classification means Vital, Essential and Desirable Classification. From the above
description, it may be noted that ABC classification is based on the logic of proportionate value while
VED classification based on experience, judgment, etc. The ABC /VED classification is presented in the
following figure 3.7.
V
E
D
Total
A
3
5
2
10
B
5
7
8
20
C
10
40
20
70
Total
18
52
30
100
Fig: 3.7 ABC/VED Classifications
This is an example of a particular case
The values are expressed in percentage
Note that the total number of categories becomes nine.
3.6 Factors Affecting Inventory
The main problem of inventory control is to answer tow questions viz.
1. How much to order? and
2. When to order?
These questions are answered by developing a inventory model, which is based on the
consideration of the main aspects of inventory viz. demand and cost. There are many factors related with
these tow main factors (Demand and Cost). In this section we will discuss these different factors.
The different factors are:
Economic Parameters
Demand
Ordering Cycle
Delivery Lag
Time Horizon
140
MBA-H2040 Quantitative Techniques for Managers
Stages of Inventory
Number of Supply Echelons
Number and Availability of Items
Government's and Organization's Policy
Economic Parameters
There are different types of economic parameters, they are:
Purchase Price
Procurement Costs
Selling Price
Holding Costs
Shortage costs
Information Processing System Operating Costs
Purchase Cost
The cost of the item is the direct manufacturing cost if it is produced in in-house or the cost paid to the
supplier for the item received. This cost usually equal to the purchase price. When the marketing price
goes on fluctuating, inventory planning is based on the average price mostly it is called as a fixed price.
When price discounts can be secured or when large production runs may result in a decrease in the
production cost, the price factor is of special interest.
Procurement Costs
The costs of placing a purchase order is known as ordering costs and the costs of initial preparation of a
production system (if in-house manufacturing) is called as set up cost. These costs are called as
procurement cost, but these costs vary directly with each purchase order placed or with set up made and
are normally assumed independent of the quantity ordered or produced.
Procurement costs include costs of transportation of items ordered, expediting and follow up,
goods receiving and inspection, administration (includes telephone bills, computer cost, postage, salaries
of the persons working for tendering, purchasing, paper work, etc.), payment processing etc. This cost is
expressed as the cost per order/setup.
Holding Costs
The holding costs also called as carrying costs. The cost associated with holding/carrying of stocks is
called holding cost or carrying cost or possession cost. Holding costs includes handling/carrying cost,
maintenance cost, insurance, safety measures, warehouse rent, depreciation, theft, obsolescence, salaries,
interest on the locked money, etc. Thus, by considering all these elements the storage cost is expressed
either as per unit of item held per unit of time or as a percentage of average money value of investment
held. Therefore the size of all these holding costs usually increases or decreases in proportion to the
amount of inventory that is carried.
Shortage Costs
141
MBA-H2040 Quantitative Techniques for Managers
These costs are penalty costs as a result of running out of stock at the time of item is required. There are
different forms of shortage cost, which is illustrated in the following figure 3.8. One form of the
shortage costs is called as back order on the selling side or backlogging cost on the production side when
the unsatisfied demand can be satisfied at later stage that is consumers has to wait till they gests the
supply.
The second form of shortage costs is called as lost sales costs on the selling side or no
backlogging costs on the production side, when the unsatisfied demand is lost or the consumers goes
some where else instead of waiting for the supply.
Shortage Costs
Selling Side
Production Side
Back order Costs
Lost Sales Costs
Backlogging
No Backlogging
Costs
Costs
Fig: 3.8 Nature of Shortage Costs
These includes the costs of production stoppage, overtime payments, idle machine, loss of
goodwill, loss of sales opportunity, special order at higher price, loss of profits etc.
Information System Operation Costs
Today there are more inventory records should be maintain in the organization, so that some person
must update the records either by hand or by using computer. If the inventory levels are not recorded
daily, this operating cost is incurred in obtaining accurate physical inventory record counts. The
operating costs are fixed.
Demand
A commodity demand pattern may be deterministic or probabilistic.
Deterministic Demand
In this case, the demand is assumed that the quantities of commodity needed over subsequent periods of
time are known with certainty. This is expressed over equal time periods in terms of known constant
demands or in terms of variable demands. The two cases are called as static and dynamic demands.
Probabilistic Demand
This occurs when requirements over a certain time period are not known with certainty but their pattern
can be denoted by a known probability distribution. In this case, the probability distribution is said to be
142
MBA-H2040 Quantitative Techniques for Managers
stationary or non-stationary over time periods. The terms stationary and non-stationary are equivalent to
the terms static and dynamic in the deterministic demand.
For a given time period the demand may be instantaneously satisfied at the beginning of the time period
or uniformly during that time period. The effect of uniform and instantaneous demand directly reflects
on the total cost of carrying inventory.
Ordering Cycle
The ordering cost is related with the inventory situation time measurement. An ordering cycle can be
identified by the time period between two successive placements of orders. The later may be initiated in
one of two ways as:
Periodic Review
Continuous Review
Periodic Review
In this case, the orders are placed at equally intervals of time.
Continuous Review
In this case, an inventory record is updated continuously until a certain lower limit is reached at this
point a new order is placed. Some times this is referred as the two-bin system.
Delivery Lag
The requirement of the inventory is felt and an order is placed, it may be delivered instantaneously or
some times it may be needed before delivery if affected. The time period between the placement of the
requisition for an item and its receipt for actual use is called as delivery lag. The delivery lags also
known as lead time.
There are four different types of lead time, they are
Administrative Lead Time
Transportation Lead Time
Suppliers Lead Time
Inspection Lead Time
The Inspection Lead Time and Administrative Lead Time can be fixed in nature, where as the
Transportation Lead Time and Suppliers Lead Time can never be fixed. It means generally the lead time
may be deterministic or probabilistic.
Time Horizon
Time horizon is, the planning period over which inventory is to be controlled. The planning period may
be finite or infinite in nature. Generally, inventory planning is done on annual basis in most of the
organizations.
143
MBA-H2040 Quantitative Techniques for Managers
Stages of Inventory
In the sequential production process, if the items/parts are stocked at more than one point they are called
multi-stage inventories. This is illustrated in the following figure 3.9.
Raw
D i
e c a
s ti n
g
D r il l i
n
g
D e -
b
ur
ni
n g
P a c
ki n
g Finished
Material
Products
Inventory
Inventory
Fig: 3.9 Multistage Inventories
Number of Supply Echelons
Already we saw that there are several stocking points in the inventory system. These stocking points are
organized in such that one points act as a supply source for some other points. For example, the
production factories supplies the products to warehouse and the warehouse supplies to the retailer and
then to the consumers. In this the each level of movement of the product is called on echelon. This is
illustrated in the figure 3.10 given below.
Consumer
Consumer
Retailer
Consumer
Warehouse
Retailer
Production
Factory
Warehouse
Fig: 3.10 Multi Echelon Supply Systems
Number and Availability of Items
Due to different marker situations some times supply position is poorly affected, which in turn affects
the poison of inventory in the organizations.
Generally inventory includes more than one item. Therefore the number of items in inventory
affects the situation when these items complete for limited total capital or limited space.
Government's and Organization's Policy
144
MBA-H2040 Quantitative Techniques for Managers
There are different governments and as well as organization policies such as import and export,
availability of capital, land, labour, pollutions systems, etc. The government has laid down some policy
norms for items to be imported as well as for other items like highly inflammable, explosive and other
important materials. Similarly, an organization also has certain policies based on the availability of
capital, labour, etc. All these policies affect organization inventories level.
We have discussed different factors that affect the inventory in an organization (above). These
factors are responsible for the development of proper inventory system is called as characteristics of
inventory.
3.7 Summary
This lesson illustrated the introduction of inventory and inventory management/control. This lesson also
illustrated the objectives of inventory, inventory functions, inventory types and the various factors
affecting the inventory.
3.8 Key Terms
Carrying Cost: Cost of maintaining one unit of an item in the stock per unit of time (normally one
year). The carrying cost also called as Holding Cost.
Decoupling: Use of inventories to break apart operations so that one operations supply is independent of
another.
Backlog: Accumulation of unsatisfied demands
Delivery Lag: Time between the placing of an order for the item and receipt of the items for use.
ABC Classification: Classifications of inventories in terms of annual usage value in different categories
of high value (A), medium value (B) and low value (C).
VED Classification: Vital Essential Desirable Classification. This is based on experience/judgment.
VED classification when coupled with ABC classification enhances the inventory control efficiency.
3.9 Self Assessment Questions
Q1. What is inventory and their Objectives?
Q2. Discuss different types of inventory.
Q3. What are the major functions of inventory in an organization?
Q4. Explain different factors affect the inventory?
Q5. Some business peoples think that inventory as necessary evil while others think inventory as an
asset. What is your view?
3.10 Further References
Hamdy A Taha, 1999. Introduction to Operations Research, PHI Limited, New Delhi.
145
MBA-H2040 Quantitative Techniques for Managers
Mustafi, C.K. 1988. Operations Research Methods and Practices, Wiley Eastern Limited, New Delhi.
Levin, R and Kirkpatrick, C.A. 1978. Quantitative Approached to Management, Tata McGraw Hill,
Kogakusha Ltd., International Student Editon.
Levin and Rubin. Quantitative Techniques for Managers, PHI, New Delhi.
UNIT II
4 INVENTORY MODELS
ON
S
S
E
L
LESSON STRUCTURE
4.1 Introduction
4.2 Deterministic Inventory Model
4.3 Deterministic Single Item Inventory Model
4.3.1 Economic Order Quantity Model I
4.3.2 Economic Order Quantity Model II
4.3.3 Economic Production Quantity Model
4.3.4 Price Discounts Model
4.3.5 Dynamic Demand Models
4.4 Deterministic Multi Items Inventory Model
4.4.1 Unknown Cost Structure Model
4.4.2 Known Cost Structure Model
Objectives
4.5 Probabilistic Inventory Models
After Studying this lesson, you should be able
4.5.1 Single Period Probabilistic Model
to:
4.5.2 Single Period Discrete Probabilistic
Understand Simple Deterministic
Demand Model
Inventory Models
4.6 Summary
Understand Simple Probabilistic
4.7 Key Terms
Inventory Models
4.8 Self Assessment Questions
Develop Simple Deterministic Inventory
4.9 Further References
Models
Illustrate the use of Simple Inventory
Models in Practical Situations
Briefly explain Single Item Inventory
Models
Briefly explain Multi Item Inventory
Models
Describe static and dynamic Inventory
Models
Develop Single Period Probabilistic
Models
146
MBA-H2040 Quantitative Techniques
for Managers
4.1 Introduction
The methodology for inventory situation modeling is based on four concepts, they are:
1. Examine the inventory situation, list characteristics and assumption related to the
inventory situation.
2. Develop the total annual relevant cost equation in narrative form as:
Total annual
Procurement
Carrying cost
Stock out costs
r e l
e
v
a
n t
c
o
s
t = I
t
e
m
c
o
s t +
c
o s t +
(
c
y
c
l
e
s t
o
c
k
s +
(cost/sales back
+ safety
order)
stocks)
3. Convert the total annual relevant cost equation from narrative form into the shorthand
logic of mathematics.
4. Optimize the cost equation by finding the optimum for how much to order (also called
order quantity), when to re-order (also called re-order point) and the total annual
relevant cost.
In general, the situation of inventory can be classified into tow types viz. deterministic and
stochastic.
Deterministic- in this variables are known with certainty
Stochastic ? in this variables are probabilistic
This lesson briefly outlines Deterministic Inventory Models and Probabilistic (Discrete
Demand Distribution Model) Inventory Models.
In this section we will discuss deterministic inventory models and later we will discuss
probabilistic inventory models (section 4.5).
4.2 Deterministic Inventory Models
There are different deterministic inventory models, they are:
a. Deterministic single item Inventory Models
i.
EOQ ? Economic Order Quantity Model ? I
ii.
EOQ ? Economic Order Quantity Model ? II (instantaneous supply
when shortages are allowed)
iii.
EPQ ? Economic Production Model (Gradual supply case and shortage
not allowed)
iv.
Price Discounts Model (instantaneous supply with no shortages)
v.
Dynamic Demand Models
b. Deterministic multi-item Inventory Models
i.
Unknown cost structure Model
ii.
Known cost structure Model
147
MBA-H2040 Quantitative Techniques
for Managers
4.3 Deterministic Single Item inventory Models
Inventory models with all the known parameters with certainty are known as deterministic
inventory model. In this section we will discuss the deterministic inventory models for sing
item.
4.3.1 Economic Order Quantity (EOQ) Model I
The EOQ concept applies to the items which are replenished periodically into inventory in
lots covering several periods' needs, subject to the following conditions:
Consumption of item or sales or usage is uniform and continuous
The item is replenished in lots or batches, either by manufacturing or by purchasing
Description
The EOQ model is described under the following situations:
a. Demand is deterministic and it is denoted by D units per year.
b. Price per Unit or cost of purchase is C.
c. Planning period is one year.
d. Ordering Cost or Procurement Cost or Replenishment Cost is Co. Suppose if the items
are manufactured it is known as set up cost.
e. Holding Cost (or carrying cost) is Ch per unit of item per one year time period. The
Ch is expressed either in terms of cost per unit per period or in terms of percentage
charge of the purchase price.
f. Shortage Cost (mostly it is back order cost) is Cs per unit per year.
g. Order Size is Q.
h. Cycle period of replenishment is t.
i. Delivery lad/Lead Time is L (expressed in units of time)
In this section we will discuss about instantaneous supply when shortages are not allowed.
That is whatever is demanded, is supplied immediately after the lead time. If we assume these
constraints, a graph of inventory against time will be look like a regular saw-tooth pattern as
given below (Fig: 4.1).
148
MBA-H2040 Quantitative Techniques
for Managers
Inventory
Inventory
Reorder Point
Q
Q/2
L
L
Time
Reorder Level
t t t
Fig: 4.1. Saw-Tooth Inventory Model
- In this model we assumed that the shortages are not allowed, it means that shortage
cost is prohibitive or Cs is too much large or infinite.
- Everything is so known and regular, there is no need of safety stock.
- Inventory will run out altogether just as the next lot is received.
The different levels of inventories for this model are fixed as follows:
- Minimum level = Safety Stock (Buffer Stock)
- Maximum level = Minimum level + EOQ
- Reorder level = Minimum level + Lead Time Consumption
-- In this case safety stock is not needed, so that safety stock is zero i.e. Minimum
level = 0.
-- Maximum Inventory is the ORDER SIZE (lot size).
-- Maximum Inventory is the ORDER SIZE (lot size).
Therefore, the average inventory per cycle = ?(Maximum level + Minimum level).
Here cycle is the intermittent pattern, in which inventory vary from maximum
to minimum and then back to maximum.
-- Maximum inventory is Q
Therefore, the average inventory per cycle = ?(Q+O) =Q/2, and the average
inventory is time
independent.
149
MBA-H2040 Quantitative Techniques
for Managers
In this case, the Total Annual Relevant Cost is as follows:
Total annual
Annual Purchase
=
+ Annual Ordering
relevant cost
Cost (PC) +
Cost (OC)
(TC)
Annual Carrying
Cost (CC)
Quantity Ordering Number of Carrying Cost
Average
= (Price/Unit) Purchased + Cost/Order Orders placed / + per unit
number of
Per Year
year
units carried
-------------- eq.1
Note that,
Number of Orders/Year = Annual Demand/Order Size
= D/Q
Thus, the eq.1 is written as:
TC = CD + Co D/Q + Ch Q/2 ------------------- eq.2
The EOQ or Order Size is that quantity, which minimizes the Total Cost. Total Cost is the
sum of Fixed Cost and Variable Cost. The Fixed Cost (CD) is independent of Order Size
while the variable Cost is dependent on the Order Size (Q). Since, the fixed cost does not play
any in minimization or maximization process, only variable cost will be minimized here.
For the total cost to be minimum, the first order derivative of TC is zero, that is,
dTC/dQ = -CoD/Q2 +Ch/2 = 0 ----- eq.3 or
CoD/Q = ChQ/2 --------- eq.4
or
Annual Ordering Cost = Annual Carrying Cost ----------- eq.5
The eq.5 may also be obtained from the following fig: 4.2
150
MBA-H2040 Quantitative Techniques
for Managers
TOTAL COST
Total Cost Curve
Annual Carrying
Cost Curve
TC
Annual Ordering Cost
Curve
Annual Purchase Cost
Curve
Q*
ORDER SIZE = Q
Fig: 4.2. Inventory Control Cost Trade-offs
Now, if we examine the eq.2 that is the total cost equation, we obtain the relationship
between the fixed cost and variable cost. This relationship is shown in the above fig.4.2. Note
that the total cost curve has the lowest value just above the intersection of the ordering cost
curve and carrying cost curve, and also at the intersection annual cost is equal to the annual
carrying cost.
From the eq.4, now we will get
EOQ = Q* =
2CoD ----------- eq.6
Ch
The cycle period
t = Optimal Order Quantity or t* = Q* =
2Co ----------
eq.7
Annual Demand D ChD
N = Total number of orders per year , which is the reciprocal of cycle period (1/t*)
That is
N = D =
ChD ------------------ eq.8
151
MBA-H2040 Quantitative Techniques
for Managers
Q 2Co
The annual cost = TC = CD +
2CoChD ----------- eq.9
Lead Time Consumption = (Lead time in years) * (Demand Rate per year)
Minimum Level = O
Maximum Level = Q*
Reorder Level = LD
Let us see few example of this case.
Example 4.1
A manufacturer uses Rs.20, 000 worth of an item during the year. Manufacturer estimated the
ordering cost as Rs.50 per order and holding costs as 12.5% of average inventory value. Find
the optimal order size, number of orders per year, time period per order and total cost.
Solution
Given that:
D = Rs.20, 000
Co = Rs.50
Ch = 12.5% of average inventory value / unit
Total Cost = TC = 25D + (0.125) Q , where Q is order size in Rs.
Q D
By applying the equations (eq.6) to (eq.9), we will get Q*, t*, N
?
Q* =
2CoD
Ch
=
2 * 50 * 20000
= Rs.4000
0.125
t* =
2Co
ChD
=
2 * 50 = 1 years = 73 days
(0.125) * (20000) 5
N = 1 = 5
t*
Note: TC means in this case variable cost only
152
MBA-H2040 Quantitative Techniques
for Managers
TC* =
2 * 50 * 0.125 * 20000
= Rs.500
Therefore
Order Size = Q = Rs.4000
Number of order / year = N = 5
Time period / order = t* = 73 days
Total Cost = TC* = Rs.500
Example 4.2
A manufacturer uses an item at a uniform rate of 25,000 units per year. Assume that no
shortage is allowed and delivery is at an infinite rate. The ordering, receiving and hauling cost
is Rs.23 per order, while inspection cost is Rs.22 per order. Interest costs is Rs.0.056 and
deterioration and obsolescence cost is Rs.0.004 respectively per year for each item actually
held in inventory plus Rs.0.02 per year per unit based on the maximum number of units in
inventory.
Determine the EOQ. If lead time is 40 days, find reorder level.
Solution
Given that
Demand = D = 25000 units/year
Ordering Cost = Co = 23 + 22 = Rs.45 per order
Storage cost Ch = 0.056 + 0.004 = Rs0.060 (based on actual inventory (=average
inventory)
Storage cost Ch = Rs.0.02 per unit/year
(based on maximum inventory)
Total Variable Cost = TC = 25 * 25000 + 0.060 * Q + 0.02 * Q
Q
2
= 625000 + 0.1*Q
Q 2
Thus,
Q* =
2CoD
= Q* =
2*25*25000 =3535.5 units (3535 units
approximately)
Ch
0.1
Reorder level = L*D = 40*25000 = 2739.7 units
365
That is = 2740 units
Therefore
EOQ = Q* = 3535 units
Reorder level = 2740 units
153
MBA-H2040 Quantitative Techniques
for Managers
Look in to the fig.4.2; in this the total cost curve is almost flat near the minimum cost
point. This indicates that small variations in optimal order size will not change the total cost
appreciably. For this purpose we will examine the model of sensitivity in the next section.
Model Sensitivity
In order to examine sensitivity of the model, we compare the sensitivity of the total costs (TC)
for any operating system with the total variable costs for an optimal inventory system (TC*)
by using the ratio TC/TC*. To do this, we have to calculate TC/TC* as a function of Q/Q*.
154
MBA-H2040 Quantitative Techniques
for Managers
Therefore
CoD + ChQ CoD + ChQ*
TC = Q 2 Q* 2 ---------- eq.10
Now by substituting Q* =
2CoD/Ch into equation eq.10 and solving algebraically we get
the relationship as
TC = 1 Q* + Q ---------------------- eq.11
TC* 2 Q Q* , which is shown in the following figure fig.4.3
TC/TC*
2.0
1.5
1.0
0.5
0.5 1.0 1.5 2.0 2.5 3.0 Q/Q*
Fig: 4.3 Inventory Sensitivity (in case of simple lot size)
According to this if Q is off from optimal either direction by a factor of 2; costs are
increased by only 25%. This has an important practical implication.
The model sensitivity is explained with the help of the following example, so that we
will understand more.
Example 4.3
Consider the Example 4.1. The sensitivity of total cost if order size is Rs.4000, then we will
get that
TC/TC* = 1 4000 + 4000 = 1
2 4000 4000
155
MBA-H2040 Quantitative Techniques
for Managers
This indicates that even though order quantity deviates from optimal by Rs.4000 or 100%, the
costs are only 25% higher than the optimal. This excess cost of the non-optimal order quantity
can be found as:
Excess Cost (Marginal Cost) = 0.25(TC*) = 0.25(500) = Rs.125.
4.3.2 The EOQ Model II
Here we are going to discuss, Instantaneous Supply When shortages are allowed. In this case,
stock outs are permitted which means that shortage cost is finite or it is not more. The entire
Model I assumptions (a to i) are also good applicable here. The Inventory situation with
shortages is represented diagrammatically in the following figure fig.4.4.
Look the figure there are two triangles viz. ABC and CEF.
The triangle ABC denotes inventory, whereas
The triangle CEF denotes the shortage
I = inventory level
S = shortage level
Q = order size = I + S
Cycle period = t = t1 + t2
Where t1 is the proportion of cycle period for inventory holding
t2 is the time of stock out
Inventory Reorder Point
A
I
Q t2 t2
B t1 C
F t1 S
S
Shortage
D E
t t
Fig: 4.4 Inventory Situation with Shortages
Total Variable Cost = Annual Ordering Cost + Annual Holding Cost + Annual
Shortage Cost
= CoD + I2Ch + (Q-I)2Cs ---------------- eq.12
156
MBA-H2040 Quantitative Techniques
for Managers
Q 2Q 2Q
From this we will get
EOQ = Q* =
( 2CoD ) ( Co+Ch) -------- eq.13
Ch Cs
Inventory Level = I* =
( 2CoD ) ( Cs ) -------- eq.14
Ch Co+Ch
Shortage Level = Q* - I* ------------------ eq.15
Cycle Period = t* = Q* ----------------- eq.16
D
Number of Orders/Year = 1
t*
Therefore
Total Variable Cost = TC* =
2CoChCsD
(Ch + Cs) -------------- eq.18
Thus, if we compare the total variable cost of Model I and Model II we will see that
2CoChD >
2CoChCsD ---- This implies that the annual
(Ch + Cs)
Cost when shortage is allowed is less than the annual inventory management cost when
shortages are not allowed. That is shortage should be allowed whenever the shortage cost is
not very large for reducing the total cost.
Example 4.4
Consider the following problem.
Problem
The demand for an inventory item each costing Re5, is 20000 units per year. The ordering
cost is Rs.10. The inventory carrying cost is 30% based on the average inventory per year.
Stock out cost is Rs.5 per unit of shortage incurred. Find out various parameters.
Solution
Given that
Demand = D = 20000
Ordering Cost = Co = Rs.10
Carrying Cost = Ch = 30% of Re 5 = 30 * 5 = 1.5
100
157
MBA-H2040 Quantitative Techniques
for Managers
Stock out Cost = Cs = Rs.5
Now we have to determine the various parameter of EOQ Model II such as EOQ, Inventory
Level, Shortage Level, Cycle Period, number of orders/year and Total Cost.
EOQ = Q* =
(2CoD ) ( Co+Ch)
Ch Cs
=
2*10*20000 0.30 + 10 = 1657 units
0.30 5
Inventory Level = I* =
( 2CoD ) (Cs )
Ch Co+Ch
Inventory Level = I* =
2*10*20000 5 = 804 units
0.30 10+0.30
Shortage Level = Q* - I* = 1657 ? 804 = 853 units
Cycle Period = t* = Q* = 1657 = 30.24 days = 30 days
D 20000
Number of Orders/Year = 1 = 1 = D = 20000 = 12 Orders/year
t* Q*/D Q* 1657
Total Cost =
2CoChCsD
(Ch + Cs)
=
2*10*0.30*5*20000 = RS.336.4
0.30+5
4.3.3 Economic Production Quantity (EPQ) Model
Here we will discuss about Graduate Supply case when Shortages are not allowed. EOQ
model is more common in retail situation, while economic production quantity EPQ is
basically associated with manufacturing environment. EPQ shows that over a period of time
inventory gradually built and the consumption go side by side where production rate is higher
than that of consumption rate.
158
MBA-H2040 Quantitative Techniques
for Managers
Assumption (a) to (i) of EOQ Model I also hold good for this model. In this model the
Order Size (Q) is taken as Production Size, the annual production rate is taken as P such that P
> D, otherwise, if P D, the item will be used as fast as it is produced. This situation is
illustrated in the following figure fig: 4.5.
159
MBA-H2040 Quantitative Techniques
for Managers
Maximum Inventory Level
B
Rate of
Decrease of
Inventory=D
Rate of increase of
Inventory = P-D
t1 t2 t1 t2
A D C Time
t t
Fig: 4.5 Gradual Replacement Inventory Situation
In the above figure, t = Cycle Time
t1 = Production Time
t2 = Depletion Time and
t = t1 + t2 of maximum inventory level BD.
Production Time = t1 = Q
P
Cycle Time = t = Q
D
Maximum Inventory Level = BD = (P ? D) * t1
= (P ? D) * Q
P
Minimum Inventory Level = 0
Average Inventory Carried = (P ? D) Q + O
P
(P ? D) Q
=
2
2P
160
MBA-H2040 Quantitative Techniques
for Managers
Total Variable Cost/Year = Annual Setup Cost + Annual Carrying Cost
CoD (P ? D) Q
=
+ Ch
----- eq.19
Q P 2
Thus,
EPQ = Q* =
2CoD
------------------------ eq.20
Ch(P-D)
P
Total Variable Cost = TC* =
2CoCh(P-D) D --------- eq.21
P
The economic production quantity model (gradual supply case and shortage not
allowed) is explained with the help of following Example 4.5.
Example 4.5
Problem
An inventory item unit is used at the rate of 200/day, and can be manufactured at a rate of
700/day. It costs Rs.3000 to set up the manufacturing process and Rs.0.2 per unit per day held
in inventory based on the actual inventory any time. Assume that shortage is not allowed.
Find out the minimum cost and the optimum number of units per manufacturing run.
Solution
Given that
Demand, D = 200 units
Production, P = 700 units
Set up Cost, Co = Rs.3000
Holding Cost, Ch = Rs.0.2
? Minimum Cost = TC* =
2CoCh(P-D) D
P
=
2*3000*0.2(700-200) (200)
= Rs.414
700
EPQ = Q* =
2CoD =
2CoD * P
161
MBA-H2040 Quantitative Techniques
for Managers
Ch(P-D) Ch(P-D)
P
=
2*3000*200*700
= 2898.2 = 2898 units
0.2(700-200)
Minimum Cost = Rs.414
Therefore
Number of units per Manufacturing Run = 2898 units
4.3.4 Price Discounts Model
In this section we will discuss, instantaneous supply with no shortages. We know very well
that whenever we make bulk purchasing of items there may be some discount in price is
usually offered by the suppliers. As far as discount concern, there are two types:
1. Incremental Discount ? discount allowed only for items which are in excess of the
specified amount. In this case, all the prices offered in different slabs are applicable in
finding the total cost.
2. All units Discount ? discount allowed fro all the items purchased. In this, only one
price at any one slab is applicable for finding the total cost.
Here we are going to discuss only all units' discount type.
Advantages of Bulk Purchase
Buying in bulk may results in the following advantages:
less unit price
less ordering cost
cheaper transportation
fewer stock outs
sellers preferential treatment
Disadvantages of Bulk Purchase
Bulk purchase also has the following disadvantages in addition to the above
advantages:
high carrying cost
lower stock turnover
huge capital required
less flexibility
older stocks
heavy deterioration
heavy depreciation
162
MBA-H2040 Quantitative Techniques
for Managers
In case of purchased items, if the discounts are allowed, the price C may vary according to the
following
pattern:
C = P0 if purchase quantity Q = Q0 < q1
= P1 if purchase quantity q1 Q1 < q2
..............................................
...............................................
= Pn if purchase quantity qn < Qn
where Pj-1 is > Pj, for j = 1, 2, ........, n.
-------- eq.22
Pj = Price per unit for the jth lot size
Suppose, shortages are not permitted, the total cost per year is obtained by the following set of
relations:
TC(Q
Q
j) = DPj + CoD + 1 ChQj
, where qj
j<qj+1
----------- eq.23
and Ch=ipj, i being percentage change for j=0 to n. Since price (or unit cost) varies with
purchase size (Q), the fixed cost term CD in equation eq.23 can't be omitted for minimizing
the total cost (TC).
Equation eq.23 for quantity discounts are represented in the figure fig: 4.6.
Total
Cost
(TC)
163
MBA-H2040 Quantitative Techniques
for Managers
O q1 q2
q3
q4 Quantity (Q)
Fig: 4.6 Quantity Discount
The heavy curve on the various price discounts shows feasible portion of the total cost
which is a step function. Therefore, for determining the overall optimum, the following steps
of procedure is adopted:
Step 1: Find EOQ for the lowest price
That is compute Q*n =
2CoD
iPn
If Q* q
n
n, the optimum order quantity is Q*n.
If Q*n < qn, then go to Step 2.
Step 2: Compute Q*n -1 =
2CoD for the next lowest price.
iPn-1
q
If Q*n -1
n-1, then compare the total cost TCn-1 for purchasing Q*n -1 with the total
cost TCn for purchasing quantity qn and select least cost purchase quantity.
If Q*n -1 < qn-1 then go to Step 3.
Step 3: Compute Q*n -2 =
2CoD for the next lowest price.
iPn-2
If Q*n -2 qn-2, then compare the total cost TCn-2, TCn-1 and TCn for purchase
quantities Q*n -2 , qn-1 and qn respectively and select the optimum purchase quantity.
If Q*n -1 < qn-2 then go to Step 4.
Step 4: Continue the procedure until Q*
q
n -j
n-j. Then compare the total costs TCn-j with TCn-
j+1, .... , TCn-1, TCn for purchase quantities Q*n -j , qn-j+1, .... , qn-1, qn respectively, and select
the optimum purchase quantity.
The price discount model is explained with the help of the following examples.
Example 4.6
Consider the following problem, which explains the price discount model.
Problem
Suppose, annual demand for an item is 1500 units, ordering cost is Rs.250, inventory carrying
charge is 12% of the purchase price per year and the purchase prices are:
164
MBA-H2040 Quantitative Techniques
for Managers
P1 = Rs.5
for purchasing Q1 < 250
P2 = Rs.4.25 for purchasing 250 Q2 < 500
P3 = Rs.3.75 for purchasing 500 Q3
Find out the optimum purchase quantity.
Solution
Given that
Demand, D = 1500 units
Ordering Cost, Co = Rs.250
Carrying Cost, Ch = 12% = i
This problem belongs to the Step 1. So that as per Step 1
Q*3 =
2CoD =
2*250*1500 = 1290 units
iPn
(0.12)(3.75)
Note that 1290 > 500, Optimum Purchase Quantity = 1290 units.
Example 4.7
Problem
Consider the problem in Example 4.6 with ordering cost of Rs.10 only. Find out the optimum
purchase quantity.
Solution
Given that
Demand, D = 1500 units
Ordering Cost, Co = Rs.10
Carrying Cost, Ch = 12% = i
In this case,
Q*3 =
2CoD =
2*10*1500
= 258 units
iPn
(0.12)(3.75)
Since 258 < 500 = q3, so that we may have to compute
Q*2 =
2CoD =
2*10*1500 = 242 units
iPn
(0.12)(4.25)
Since 242 < 250 = q2, next we may have to compute
165
MBA-H2040 Quantitative Techniques
for Managers
Q*1 =
2CoD =
2*10*1500 = 224 units
iPn
(0.12)(5)
Since 224 > 0, now we have to compare the total cost for purchasing Q*1 = 224, q2 =
250 and q3 = 500 units respectively.
From equation eq.23
TC (Qj) = DPj + CoD + 1 ChQj
, where qjQj<qj+1
We will get
TC1 (for purchasing 224) = 5*1500 + 10*1500 + 1 (0.12) (5)224 = 7634.2
224 2
TC2 (for Q2=250) = 4.25*1500 + 10*1500 + 1 (0.12) (4.25)250 = 6498.75
250 2
TC3 (for Q3=500) = 3.75*1500 + 10*1500 + 1 (0.12) (3.75)500 = 5767.5
500 2
The EPQ for his problem is Q*3 = 500 units
4.3.5 Dynamic Demand Models
In this model, assume that demand is known with certainty, and although may vary from one
period to the next period. There are five types of dynamic demand inventory models, they are:
i)
Production Inventory Model (Incremental Cost Method)
ii)
Dynamic Inventory Model (Prescribed Rule Method)
iii)
Dynamic Inventory Model (Fixed EOQ Method)
The above five dynamic demand models of inventory are discussed in details in the following
subsequent sections.
i) Production Inventory Model
This is also called as Incremental Cost Method. This situation is explained with the help of
the following example 4.8.
Example 4.8
Consider the following problem.
Problem
A production factory has a fixed weekly cyclic demand as follows:
166
MBA-H2040 Quantitative Techniques
for Managers
Days
Mon
Tue
Wed
Thu
Fri
Sat
Sun
Demand
9
17
2
0
19
9
14
in units
Here the policy is to maintain constant daily production seven days a week. Shortage cost is
Rs.4 per unit per day and storage cost depends upon the size of Q, the quantity carried, as
follows:
Cost per unit for one
=1
=4
=10
day(Rs.)
Q
1Q3
4Q20
20Q
And the chargers are based on the situation at the end of the day.
Determine the optimal starting stock level.
Solution
Let the production rate be the average of the total sales which is
= 9 + 17 + 2 + 0 + 19 + 9 + 14 = 10 units/days
Day
On hand
Demand
Inventory
Shortage
Carrying
stock start
cost (Rs.)
cost (Rs.)
of day
Mon
8
9
-1
4
0
7
Tue
-1+10=9
17
-8
32
0
Wed
-8+10+=2
2
0
0
0
Next
Thu
0+10=10
0
10
0
44
we will
Fri
10+10=20
19
1
0
1
find the
Sat
1+10=11
9
2
0
2
total
Sun
2+10=12
14
-2
8
0
weekly
Mon
-2+10=8
Total Cost 44 + 43 = 87
cost for
different starting stocks, which is illustrated in the following tables.
Table 4.1 shows for starting stock 8, Table 4.2 shows for starting stock 9 and Table
4.3 shows for starting stock 10.
167
MBA-H2040 Quantitative Techniques
Day
On hand
Demand
Inventory
Shortage
Carrying
for Managers
stock start
cost (Rs.)
cost (Rs.)
of day
Mon
9
9
0
0
0
Tue
0+10=10
17
-7
28
0
Wed
-7+10+=3
2
1
0
1
Table
Thu
1+10=11
0
11
0
44
4.1
Fri
11+10=21
19
2
0
2
Cost
Sat
2+10=12
9
3
0
3
Analysi
Sun
3+10=13
14
-1
4
0
s with
Mon
-1+10=9
Total Cost 32 + 50 = 82
Initial
Stock 8
Day
On hand
Demand
Inventory
Shortage
Carrying
stock start
cost (Rs.)
cost (Rs.)
of day
Mon
10
9
1
0
1
Tue
1+10=11
17
-6
24
0
Wed
-6+10+=4
2
2
0
2
Thu
2+10=12
0
12
0
48
Fri
12+10=22
19
3
0
3
Sat
3+10=13
9
4
0
16
Sun
4+10=14
14
0
0
0
Mon
0+10=10
Total Cost 24 + 70 = 94
Table 4.2 Cost Analysis with Initial Stock 9
Table 4.3 Cost Analysis with Initial Stock 10
Therefore
The optimal solution for starting stock of 9 units is on MONDAY.
Minimum Total Cost for this is Rs.82.
168
MBA-H2040 Quantitative Techniques
for Managers
Procedure for solving such incremental problem is cost analysis, which is self
explanatory.
ii) Dynamic Inventory Model (Prescribed Rule Method)
Some times, the organization dealing with inventory may prescribe some rule of procurement
of items of inventory. For example
- Procuring every month
- Procuring every three months
The prescribed rule method is explained with the help of the following Example 4.9.
Example 4.9
Consider the following problem.
169
MBA-H2040 Quantitative Techniques
for Managers
Problem
An organization estimates the demand of an item as follows:
Month
Jan
Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Total
Demand 140 98
62
134 20
72
22
164
139 170 248
51
1320
The organization has decided that their ordering cost is Rs.54 and carrying charge per unit per
month is 2% at the end of each month. The cost of item per unit is Rs.20. Assume that the
supply is instantaneous, there in no stock outs and no lead time. Only full month requirement
is ordered.
Solution
The following table is used for determining the total cost for the policy of quarterly ordering.
Month
Jan
Feb Mar Apr May Jun
Jul
Aug Sep Oct
Nov Dec Total
Starting
0
160 62
0
92
72
0
303 139 0
299 51
--
Inventory
300 --
--
226 --
--
325 --
--
469 --
--
1320
Replenishment
140 98
62
134 20
72
22
164 139 170 248 51
1320
Requirements
160 62
0
92
72
0
303 139 0
299 51
0
1178
Ending
Inventory
Table 4.4 Total Cost Analysis for Quarterly Policy
Therefore
Total Carrying Cost = 1178*(0.02)*(20) = Rs.471.2
Total Replenishment Cost = 4*54 = Rs.216
Total Annual Cost = 216 + 471.2 = Rs.687.2
iii) Dynamic Inventory Model (Fixed EOQ Method)
The Fixed EOQ Method is explained with the help of the following Example 4.10.
Example 4.10
Consider the problem of Example 4.9.
Problem
An organization estimates the demand of an item as follows:
Month
Jan
Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Total
Demand 140 98
62
134 20
72
22
164
139 170 248
51
1320
170
MBA-H2040 Quantitative Techniques
for Managers
The organization has decided that their ordering cost is Rs.54 and carrying charge per unit per
month is 2% at the end of each month. The cost of item per unit is Rs.20. Assume that the
supply is instantaneous, there in no stock outs and no lead time. Only full month requirement
is ordered.
Solution
Average Monthly Demand = Demand = 1320 = 110 units/month
Month 12
EOQ =
2*54*110 = 172 (approximately)
(0.02)*(20)
In this case the organization has to ordered full month requirement, therefore 172 lies between
140 and 248 units, but 172 is more closer to 140 than 248, so that the organization order one
month requirement at the beginning of the January.
Similarly, at the beginning of the February, the organization order two month
requirement.
The detailed result is illustrated in the following Table 4.5
Month
Jan
Feb Mar Apr May Jun
Jul
Aug Sep Oct
Nov Dec Total
Starting
0
0
62
0
92
72
0
164 0
0
0
0
--
Inventory
140 160 --
226 --
--
186 --
139 170 248 51
1320
Replenishment
140 98
62
134 20
72
22
164 139 170 248 51
1320
Requirements
Ending
0
62
0
92
72
0
164 0
0
0
0
0
390
Inventory
Table 4.5 EOQ Total Cost Analysis
Therefore
Total Ordering Cost = 8*54 = Rs.432
Total Carrying Cost = 390*0.4 = Rs.156
Thus, the Total Cost is reduced if EOQ policy is used instead of three month
(quarterly) rule.
4.4 Deterministic Multi Item Inventory Models
When there is more than one item in the inventory is called as multi item inventory. Since this
contains more items, the inventory control requires special type of care. This type of
171
MBA-H2040 Quantitative Techniques
for Managers
inventory problems may different types of constraints like capital, cost structure, storage
space, purchasing load etc. As the number of constraints increase the problem becomes more
complex. In this section we will discuss some examples of this type of inventory.
There are two types of multi item inventory model which is based on the structure of
the cost, they are:
1. Model with Unknown Cost Structure
2. Model with Known Cost Structure
4.4.1 Unknown Cost Structure Model
In India most of the organizations do not maintain the proper inventory records which may
provide sufficient cost information to generate the two basic parameters viz. procurement cost
and carrying cost of inventory control. Some organizations in some situations have not
developed cost structure related to inventory control, but still they wish to minimize total cost
of inventory. There may be critical situations, in which an organization may need to take
immediate actions to improve the situation without considering the structure of cost.
The use of inventory models without cost information is impossible, but we will show
here that it is possible to get many of benefits of inventory techniques even when carrying
cost and ordering cost are not known. In such a problem, there are two different approaches,
they are:
1. Minimize the Total Carrying Cost while keeping the number of orders/year fixed
or
2. Minimize the Total Number of orders/year while keeping the same level of
inventory.
Minimize the Total Carrying Cost while keeping the number of orders/year fixed Model
As per the EOQ model
Q* = 2CoD
or
Q* = 2CoD * D
Ch
Ch
Suppose = 2Co = Constant, because ordering cost and carrying cost are deterministic
values.
Ch
Therefore
Q* = * D ------------ eq.24
The equation eq.24 tells that EOQ is proportional to the square root of demand for any
inventory item of control.
172
MBA-H2040 Quantitative Techniques
for Managers
For this equation, we get
= Q* = Q*D = D = Square root of Demand ----------- eq.25
D D (D/Q*) Number of orders/year
Since is constant for any single item of inventory, we take as the constant for all
the inventory items. Thus, we take
= D
or = Sum of square roots of Demand of each inventory item -
eq.26
(D/Q*) Sum of the number of orders/year for each item
This model is explained with the help of the following Example 4.11.
Example 4.11
Consider the following problem.
Problem
An organization has the following procurement pattern of six items irrespective of their
demand level. Reduce the inventory levels while keeping total number of orders/year fixed.
Item No.
Number of
Demand
Order Size
Average
Orders/Year (Rs.)
(Rs.)
Inventory
1
5
2000,000
500000
250000
2
5
700,000
260000
130000
3
5
100,000
23500
11750
4
5
9,000
10000
5000
5
5
5,000
700
350
6
5
2,700
500
250
Solution
First we find the value of = D
---------- eq.27
(D/Q*)
Therefore
D = 1414 + 836 + 316 + 95 + 70 + 52 = Rs.2783
(D/Q*) = 5 + 5 + 5 + 5 + 5 + 5 = 30
Thus, = D
= 2783 = 92.7
(D/Q*) 30
Now we will analyze the ordering quantity, which is illustrated in the following Table
4.6.
Demand
D
EOQ=
Number of Average
Item No.
(D) (Rs.)
D
orders/year Inventory =
=D/EOQ
EOQ/
Number of
orders/year
1
2000,000
1414 92.7
131077.8
15.258
4369
2
700,000
836
92.7
77497.2
9.0
2583
3
100,000
316
92.7
29293.2
3.41
976
4
9,000
95
92.7
8806.5
1.0
293
173
MBA-H2040 Quantitative Techniques
for Managers
5
5,000
70
92.7
6489
0.77
216
6
2,700
52
92.7
4820.4
0.56
160
29.998
8597
Table 4.6 Analysis of Ordering Quantity
Thus, according to the policy of the organization ordering for six items a year each
item, Total Average Inventory becomes Rs.397350.
But, as per the new schedule as obtained in the above Table , the Average Inventory is
Rs.8597 which is much less and at the same time total Number of Orders remains same.
Therefore, substantial savings can still be achieved when cost information is known.
Minimize the Total Number of orders/year (or purchasing workload) while keeping the
same level of inventory Model
We will know that
Q* = D or
= Q*
D
Now onwards, we will take = Q* ------- eq.28 for all inventory items.
D
This model is explained with the help of the previous Example 4.11.
Here we will discuss about how to minimize the number of orders or purchasing
workload.
Problem
An organization has the following procurement pattern of six items irrespective of their
demand level. Reduce the inventory levels while keeping total number of orders/year fixed.
Item No.
Number of
Demand
Order Size
Average
Orders/Year (Rs.)
(Rs.)
Inventory
1
5
2000,000
500000
250000
2
5
700,000
260000
130000
3
5
100,000
23500
11750
4
5
9,000
10000
5000
5
5
5,000
700
350
6
5
2,700
500
250
Solution
Here K= Sum of the order size
Sum of square root of Demand
174
MBA-H2040 Quantitative Techniques
for Managers
K= 50000+260000+23500+10000+700+500 = 794700 = 285.555
1414+836+316+95+70+52 2783
175
MBA-H2040 Quantitative Techniques
for Managers
Item
Demand
D
EOQ= D
Number of Average
No.
(D) (Rs.)
orders/year Inventory =
=D/EOQ
EOQ/2
1
2000,000
1414 285.555 403774.77
4.95
201887.385
2
700,000
836
285.555 238723.98
2.93
119361.99
3
100,000
316
285.555 90235.38
1.10
45117.69
4
9,000
95
285.555 27127.725
0.33
13563.8625
5
5,000
70
285.555 19988.85
0.25
9994.425
6
2,700
52
285.555 14848.86
0.18
7424.43
9.74
397349.8
EOQ=794700
Table 4.7 Analysis for Reducing Number of Orders
Thus, from the Table 4.7, it is obvious that the total purchasing workload has reduced by
35%. Therefore, there is a definite saving of cost by applying these methods for multiple
items even when cost information is not known.
4.4.2 Known Cost Structure Model
We may classify the models with known cost structures into two main types, they are:
Model without Limitations
Model with Limitations
Here we will discuss these models, if we know the complete cost structure. In this model let
us consider the following symbolic notation:
Dj
-
Demand
Coj
-
Ordering Cost
Chj
-
Holding Cost or Carrying Cost for jth item respectively.
Model without Limitations
In any situation the items may be purchased according to their individual economic order
quantities, if there are no restrictions for the items being storing.
In this case,
Total Variable Cost per annum for n-items can be represented as
n
TC = CojDj + ChjQj and -------------------eq.29
J=1 Qj
2
176
MBA-H2040 Quantitative Techniques
for Managers
The optimum Order Size for each item is
Qj* =
2CojDj
for j = 1,........., n ---------------- eq.30
Chj
Model with Limitations
We know that the different constraints on inventories are available capital, order size per year,
storage space, etc. In this case we will consider a single constraint for the discussion of this
model. Note that the constraints capital and storage space are interchangeable.
Let
Qj is the order quantity for item j
Fj is the storage space (or floor space) for one unit of item j
(or) Fj ? the capital requirement for one unit of item j, since the storage
space and
the capital are interchangeable.
F is the total available storage space (floor space) or available capital
Then the constraints are as follows:
n
F1Q1 + F2Q2 + ............ FnQn F -------------eq.31
j=1
Here the objective is to minimize the total inventory cost expressed by the equation eq.29. So
that in order to solve this problem first we have to convert the constraint problem into
unconstrained minimization problem and then the optimum result is obtained.
The problem is to minimize the function known as Lagrange Function:
n n
L(, Q
) =
1, Q2, ..., Qn
CojDj + ChjQj + ( FjQj - F) ---------eq.32
j=1 Qj 2 j=1
Then, the values of optimal sizes are
Qj* =
2CojDj
for j = 1,........., n ------------eq.33
(C +2 F
hj
j)
177
MBA-H2040 Quantitative Techniques
for Managers
The known cost structures model is explained with the following Example 4.12.
Example 4.12
Consider the following problem.
Problem
A retail shop purchases three items of inventory viz. A, B and C respectively. The shop works
on the limitations that the shop is not able to invest more than Rs.40000 at any time. The other
shop relevant information is given in the following table:
Item
A
B
C
Demand rage (units/year) ? Dj
20000
1000
2000
Purchase Cost/unit ? Cj
40
200
100
Ordering Cost/Order ? Coj
100
150
200
Holding or Carrying Cost - J
40%
40%
40%
Solution
In the absence of constraints, the Optimal Order Sizes are:
Q1 =
2*100*20000 = 500
(.40) (40)
Q2 =
2*150*1000 = 61
(.40)(200)
Q3 =
2*200*2000 = 141
(.40) (100)
Now we know the optimal sizes, so that with these optimal sizes we may determine
the maximum investment, that is
Maximum Investment = 500 * 40 + 61 * 200 + 141 * 100
= 20000 + 12200 + 14100
= Rs. 46300
178
MBA-H2040 Quantitative Techniques
for Managers
Note that the Maximum Investment Rs.46300 is greater than the allowable investment
capacity i.e. Rs.40000 in inventory.
Therefore, equation eq.33 is used with the following alteration
Qj* =
2*Coj*Dj for j = 1, 2, 3. Since, there are only three items. --- eq.34
Cj(J+2)
Note that here = is the solution of the equation.
3
Qj* =
2*Coj*Dj * Cj = 40000
j=1
Cj(J+2)
or
40*20000*100 +
150*1000*200 +
200*2000*100 = 40000
0.1+
0.1+
0.1+
or
= 0.16899919
Now we will substitute the value of in equation eq.34
Therefore, now the order size becomes
Q*1 = 368
Q*2 = 45
Q*3 = 104
That is,
If limitations are not imposed on the purchase of quantities,
The optimal Total Cost = 2*100*20000 + 2*150*1000 + 2*200*2000
500
61
141
= 8000 + 4918 + 5674
= Rs.18592
But, under the limitations,
179
MBA-H2040 Quantitative Techniques
for Managers
The optimal Total Cost = 2*100*20000 + 2*150*1000 + 2*200*2000
368+2*0.17*40 45+2*0.17*200 104+2*0.17*100
= 18934.19, which is higher than the total cost without
limitations.
Note:
Many times the application of the equation
Q*j =
2CojDj
for j = 1,........., n ------------eq.33
(Chj+2 Fj)
to find optimal order quantities under limitations does not help to obtain the result. Therefore,
we have to apply trial and error procedure in the following manner:
i)
First determine the EOQ's for all type of inventory items without considering the
limitation that is taking =0 and then find Qj. If these values satisfy the constraint
(eq.31), then this solution becomes optimal because the constraint is not active.
ii)
If the constraint is not satisfied by the values obtained under (i) above, we give
some value to (arbitrarily but institutively) for example say =h and solve for
Qjs. Qjs satisfy the constraint, these are optimal quantities. Otherwise, we
interpolate or extrapolate the value in between 0 and h or beyond h. With this
value of , the order sizes obtained wil be approximately optimal.
4.5 Probabilistic Inventory Models
In previous sections, we have discussed simple deterministic inventory models where each
and every influencing factor is completely known. Generally in actual business environment
complete certainty never occurs. Therefore, here we will discuss some practical situations of
inventory problems by relaxing the condition of certainty for some of the factors.
The major influencing factors for the inventory problems are Demand, Price and Lead
Time. There are also other factors like Ordering Cost, Carrying Cost or Holding Cost and
Stock out Costs, but their nature is not so much disturbing. Because of this their estimation
provides almost, on the average, as known as values. Even Price can also be averaged out to
reflect the condition of certainty. But there are situations where Price fluctuations are too
much in the market and hence they influence the inventory decisions. Similarly, the demand
variations or consumption variation of an item as well as the lead time variation influence the
overall inventory policy. In this section we will discuss single period probabilistic models.
4.5.1 Single Period Probabilistic Models
180
MBA-H2040 Quantitative Techniques
for Managers
Single Period Discrete Probabilistic Model deals with the inventory situation of the items like
perishable goods, seasonal goods and spare parts requiring one time purchase only. These
items demand may by discrete or continuous. In these models the lead time is very much
important because purchases are made only once.
In single period model, the problem is analyzed using incremental (or
marginal) analysis and the decision procedure consists of a sequence of steps. In such cases,
there are two types of cost involved. There are Under Stocking Cost and Over Stocking Cost.
These two costs describe opportunity losses incurred when the number of units stocked is not
exactly equal to the number of units actually demanded.
In this section we will use the following symbols:
D = Demand for each unit of item (or a random variable)
Q = Number of units stocked or to be purchased
C1 = Under Stocking Cost some times also known as over ordering cost. This is an
opportunity loss associated with each unit left unsold i.e.
C1 = S ? Ch/2 ? Cs
C2 = Over Stocking Cost some times also known as under ordering cost. This is an
opportunity loss due to not meeting the demand, i.e
C2 = C + Ch ? V
Where
C = cost/unit
Ch = carrying cost/unit for the entire period
Cs = shortage cost
V = salvage value
S = selling price
In this section we are going to discuss only discrete demand distribution.
4.5.2 Single Period Discrete Probabilistic Demand Model (Discrete Demand
Distribution)
Here we will discuss the following methods of solving the single period discrete probabilistic
demand.
a. Incremental Analysis Method
b. Payoff Matrix Method
a. Incremental Analysis Method
The Cost equation is developed as follows:
For any quantity in stock Q, only D units are consumed (or demanded). Then for
specific period of time, the cost associated with Q units in stock is either:
181
MBA-H2040 Quantitative Techniques
for Managers
i)
(Q-D)C2, where D is the number of units demanded or consumed is less that or
equal to the number of units Q in stock. That is D Q.
ii)
(D-Q)C1, where the number of units required is greater than the number of units Q
in stock. That is D > Q.
We know that D is a random variable, so its probability distribution of demand is known.
(D) represents the probability that the demand is D units, such that total probability is one.
That is (0) + (1) + ................. + (D) + ............. = (D) = 1 ------- eq. p1
D=0
The total expected cost is the sum of expected cost of under-stocking and over-stocking.
Therefore
The Total Expected Cost, say f(Q), is given by
Q
f(Q) = C2 (Q-D) (D) + C1 (D-Q) (D) -------------------- eq.p2
D=0 D=Q+1
Suppose, Q* is the optimal quantity stocked, then the total expected cost f(Q*) will be
minimum. Thus, if we stock one unit more or less than the optimal quantity, the total expected
cost will be higher than the optimal.
Thus,
Q*+1
f(Q*+1) = C2 (Q*+1-D) (D) + C1 (D-Q*-1) (D)
D=0 D=Q*+2
Q* Q*
= C2 (Q*-D) (D) + C2 (D-Q*) (D) + C2 (D) - C1
(D)
D=0 D=Q*+1 D=0
D=Q*+1
Q*
= f(Q*) + (C2+C1) (D) - C1 --------------- eq.p3
D=0
Thus,
f(Q*+1) ? f(Q*) = (C2+C1) (DQ*) ? C1 0 --------- eq.p4
Q*
where (DQ*) = (D) - is a cumulative probability.
D=0
Similarly, we obtain
f(Q*-1) ? f(Q*) = C1 ? (C2+C1) (DQ*-1) 0 ------eq.p5
182
MBA-H2040 Quantitative Techniques
for Managers
We will obtain the following from the above equations eq.p4 and eq.p5
(DQ*-1) C1 (DQ*) --------------------- eq.p6
C2+C1
Therefore, the optimal stock level Q* satisfies the relationship (eq.p6).
Note that for practical application of (eq.p6), the three step procedure is as follows:
Step 1:
From the data, prepare a table showing the probability (D), and cumulative probability
(DQ) for each reasonable value of D.
Step 2:
Calculate the ratio C1 which is called as service level.
C2+C1
Step 3:
Determine the value of Q which satisfies the inequality eq.p6.
This situation is explained with the following Example 4.13.
Example 4.13
An organization stocks seasonal products at the start of the season and cannot reorder. The
inventory item costs him Rs.35 each and he sells at Rs.50 each. For any item that cannot be
met on demand, the organization has estimated a goodwill cost of Rs.25. Any unsold item will
have a salvage value of Rs.20. Holding cost during the period is estimated to be 10% of the
price. The probability distribution of demand is as follows:
Units Stocked
2
3
4
5
6
Probability of
0.35
0.25
0.20
0.15
0.05
Demand (D=Q)
Determine the Optimum Number of Items to be stocked.
Solution
Now we have to follow the above sequence of steps.
Step 1:
We will prepare the Table 4.8 containing the data regarding demand distribution as
follows:
183
MBA-H2040 Quantitative Techniques
for Managers
Units Stocked
Probability of
Cumulative
Demand (D=Q) Probability (DQ)
2
0.35
0.35
3
0.25
0.60
4
0.20
0.80
5
0.15
0.95
6
0.05
1.00
Table 4.8 Probability Distribution of Demand
Step 2:
Calculate the ratio C1 which is called as service level.
C2+C1
We see that
S=50, C=35, Ch=0.1*35= 3.5, V=20, Cs=25
Therefore
C2 = C+Ch-V = 35+3.5-20 = 18.5
C1 = S-C-Ch + Cs = 50-35-3.5 + 25 = 38.25
2
2
Thus, C1 = 38.25 = 0.6740
C2+C1 18.5+38.25
Step 3:
Look into the Table 4.8, the ratio 0.6740 lies between cumulative probabilities of 0.60 and
0.80 which in turn reflect the values of Q as 3 and 4 (units stocked).
That is
(D3) = 0.60<0.6740<0.80 = (D4)
Therefore, the optimal number of units to stock is 4 units.
Cost of Under Stocking Estimation
Suppose, in the previous Example 4.13, the under stocking cost is not known, but the decision
maker policy is to maintain a stock level of say 5 units. We can determine for what values of
C1(under estimating cost) does Q*=5?
In this case, we have the following inequality:
(D4) C1 (D5)
C2+C1
That is
(D4) C1 (D5) or
18.5+C1
Or
184
MBA-H2040 Quantitative Techniques
for Managers
0.80 C1 0.95
18.5+C1
So that the Minimum Value of C1 is determine by letting
C1 = 0.80 or C1 = (0.80)(18.5) = Rs.74
18.5+C1 (1-0.80)
Similarly, the Maximum Value of C1 is determine by letting
C1 = 0.95 or C1 = (0.95)(18.5) = Rs.351.5
18.5+C1 (1-0.95)
Therefore
74C1351.5
Perishable Products Inventory
Many of the organization manage merchandise which contains negligible utility if it is not
sold almost immediately. The examples of such kind of products are newspaper, fresh
produce, printed programmes for special events and other perishable products. Generally such
inventory items have high mark-up. The major difference between the wholesale cost and the
retail price is due to the risk vendor faces in stocking the inventory. Vendor faces
obsolescence costs on the one hand and opportunity costs on the other.
All this kind of problems can be very easily solved with the help of the above
discussed model. This is explained in the following Example 4.14.
Example 4.14
A boy selling newspaper, he buys papers for Rs.0.45 each and sells them for Rs.0.70 each.
The condition here is the boy cannot return unsold newspapers. The following table shows the
daily demand distribution. If each days demand is independent of the pervious days demand,
how many news papers should he order each day?
Number of
240
250
260
270
280
290
300
310
320
330
Customers
Probability
0.01
0.03
0.06
0.10
0.20
0.25
0.15
0.10
0.05
0.05
Solution
Step 1:
185
MBA-H2040 Quantitative Techniques
for Managers
Prepare a following Table 4.9 showing the probability (D), and cumulative probability
(DQ) for each reasonable value of D.
Number of
240
250
260
270
280
290
300
310
320
330
Customers
Probability
0.01
0.03
0.06
0.10
0.20
0.25
0.15
0.10
0.05
0.05
Cumulative
0.01
0.04
0.10
0.20
0.40
0.65
0.80
0.90
0.95
1.00
Probability
Table 4.9 Probability Discrete Distribution of Demand
Step 2:
C1 = 0.25 = 0.25 = 0.357
C2+C1 0.45+0.25 0.70
Step 3:
Thus, the Value of Q such that
(DQ*-1) 0.357 (DQ*) this gives
Q* = 280
Therefore
The newspaper boy should buy 280 papers each day.
b. Payoff Matrix Method
The Payoff Matrix Method of single period Discrete Probabilistic Demand Model is explained
with the help of the following Example 4.15.
Example 4.15
Consider the example 4.13 i.e.
An organization stocks seasonal products at the start of the season and cannot reorder.
The inventory item costs him Rs.35 each and he sells at Rs.50 each. For any item that cannot
be met on demand, the organization has estimated a goodwill cost of Rs.25. Any unsold item
will have a salvage value of Rs.20. Holding cost during the period is estimated to be 10% of
the price. The probability distribution of demand is as follows:
Units Stocked
2
3
4
5
6
Probability of
0.35
0.25
0.20
0.15
0.05
Demand (D=Q)
Determine the Optimum Number of Items to be stocked.
In this case the organization has five reasonable courses of action. The organization
can stock the items from 2 to 6 units. There is no possible reason to stock more than 6 items
186
MBA-H2040 Quantitative Techniques
for Managers
since the organization can never sell more than 6 items and there is no possible reason for
ordering less than 2 items. Since there are five courses of action for stocking and five levels of
demand, it follows that there are 25 combinations of one course of action and one level of
demand. For these 25 combinations, we can determine the organization payoffs in the form of
payoff matrix.
As per the information of cost given in the problem, the payoffs are obtained for the
following two situations:
When demand is not more than the stock level
When demand is more than the stock quantity.
That is
Payoffs For
QD
Q<D
Item Cost
-35Q
-35Q
Sale of items
50D
50Q
Goodwill Cost
-
-25(D-Q)
Salvage Cost
20(Q-D)
-
Holding Cost
-3.5(Q-D)-3.5D/2
-3.5Q/2
Total payoff
-18.50Q+31.75D
38.25Q-25D
The payoff matrix will be 5X5. Each element of the matrix can be computed by above
total payoffs for demand less than, equal to, or greater than the order size (Q). When demand
is less than or equal to the order size, we have the following contributions to the payoff.
Here the organization buys the items for Rs.Rs.35Q and the organization sells D of
them for Rs.50D, the organization earns salvage of Rs.20 (Q-D) for unsold items, and the
organization incurs holding cost of Rs.(.10)(35)(Q-D) on unsold items an average holding
cost of (.10)(35)D/2 on the sold items during the period. Thus the total payoff becomes -
128.5Q+31.75D for demand less or equal to order size.
If demand is more than the order size, the contributory payoff will consist of the
following:
- Purchase Cost Rs.35Q
- Selling Profit of Rs.50Q
- Goodwill Cost Rs.25(D-Q) and
- Holding Cost of Rs. (0.10)(35)Q/2
Thus, the total payoff for demand more than order size is 38.25Q-25D.
The payoff matrix is as follows (Table 4.10):
Units Demanded D
2 3 4 5
6
Units stocked or 2 26.50 1.50 -23.50 -48.5 -
187
MBA-H2040 Quantitative Techniques
for Managers
73.5
Order Size (Q) 3 8.0 39.75 14.75 -10.25 -
35.25
4 -10.5 21.25 53.0 28.0
3.0
5 -29.0 2.75 34.5 66.25
41.25
6 -47.5 -15.75 16.0 47.75
79.5
Probability of Demand 0.35 0.25 0.20 0.15
0.05
Table 4.10 Payoff Matrix
Now we will determine the expected payoff for each order size or courses of action.
The procedure for computing the expected values is simple, as follows:
Procedure: For any given course of action multiply each possible payoff for that
course of action by the corresponding probability of the given level of demand and add all of
these products up.
Thus, for first course of action of order size 2 units, the expected value of payoff is:
(26.5)(0.35) + (1.5)(0.25) + (-23.5)(0.2) + (-48.5)(0.15) + (-73.5)(0.05) = Rs.-6
For order size 3 units
(8.0)(0.35) + (39.75)(0.25) + (14.75)(0.2) + (-10.25)(0.15) + (-35.25)(0.05) =
Rs.12.3875
For order size 4 units
(-10.5)(0.35) + (21.25)(0.25) + (53)(0.2) + (28)(0.15) + (3)(0.05) = Rs.20.2625
For order size 5 units
(-29)(0.35) + (2.75)(0.25) + (34.5)(0.2) + (66.25)(0.15) + (41.25)(0.05) =
Rs.9.4375
For order size 6 units
(-47.5)(0.35) + (-15.75)(0.25) + (16)(0.2) + (47.75)().15) + (79.5)(0.05) = Rs.-
6.225
Therefore, we compute all the expected values:
Order Size Q
2
3
4
5
6
Expected value
Rs.-6
Rs.12.3875 Rs.20.2625 Rs.9.4375
Rs.-
6.225
188
MBA-H2040 Quantitative Techniques
for Managers
Here the objective is to select course of action which provides the highest payoff.
Thus, the organization should order for 4 units for the highest expected payoff value of
Rs.20.2625.
Note: If we compare the two methods i.e. incremental analysis and payoff matrix, if we see
the solution that incremental analysis provides only the optimum level of purchase quantity
and does not indicate about the level of expected profit. But, the payoff matrix method
provides both the answers i.e. optimum purchase quantity as well as the optimum expected
profit.
The interesting is here, we may also convert the payoff matrix to opportunity cost
matrix, where the opportunity cost is, in short, a cost sustained because the decision taken is
not the best in terms of the level of demand which actually occurs.
The computation of opportunity cost matrix from the payoff matrix is very easy. Take
any column of the payoff matrix corresponding to a specific level of demand and select the
largest payoff if the payoffs are profits, the smallest payoff if the payoffs are costs. Then
subtract each payoff in the same column from the largest payoff to get the corresponding
opportunity costs in the case of profits. If it is cost, subtract the smallest payoff from each
payoff in the same column to get the opportunity costs.
In this Example 4.15, we may obtain the opportunity cost matrix as follows (Table
4.11):
Units Demanded D
2 3 4 5 6
Order size
2 0 38.25 76.5 114.75 153.0
3 18.5 0 38.25 76.5 114.75
4 37.0 18.5 0 38.25 76.5
5 55.5 37.0 18.5 0 38.0
6 74.0 55.5 37.0 18.5 0
Probability of
Demand 0.35 0.25 0.20 0.15 0.05
Table 4.11 Opportunity Cost Matrix
Now, we have to determine the expected opportunity costs for each alternative courses
of action. The objective is to select the course of action which provides minimum expected
opportunity costs.
Therefore,
The expected opportunity cost for the first alternative course of action of order size 2
is:
189
MBA-H2040 Quantitative Techniques
for Managers
(0)(.35) + (38.25)(.25) + (76.5)(.2) + (114.75)(0.15) + (153)(.05) = Rs.49.725
For order size 3 units
(18.5)(.35) + (0)(.25) + (38.25)(.2) + (76.5)(.15) + (114.75)(.05) = Rs.31.3375
For order size 4 units
(37)(.35) + (18.5)(.25) + (0)(.2) + (38.25)(.15) + (76.5)(.05) = Rs.27.1375
For order size 5 units
(55.5)(.35) + (37)(.25) + (18.5)(.2) + (0)(.15) + (38)(.05) = Rs.34.275
For order size 6 units
(74)(.35) + (55.5)(.25) + (37)(.2) + (18.5)(.15) + (0)(.05) = Rs.49.95
That is, the following are the obtained expected opportunity costs;
Order Size Q
2
3
4
5
6
Expected Cost
Rs.
Rs.31.3375 Rs.27.1375 Rs.34.275 Rs.49.95
49.725
Thus, the decision is to select the minimum expected cost is that, the organization
should store 4 units for the lowest cost of Rs.27.1375
Relationship between the Payoff Matrix and Opportunity Cost Matrix
Here, we may find a relationship between the payoff matrix and the opportunity cost
matrix, as follows:
Let
EOC = K ? EP
Where,
EOC = Expected Opportunity Cost
EP = Expected Payoff or profit
K = Constant or
K = (26.5)(.35) + (39.75)(.25) + (53)(.2) + (66.25)(.15) + (79.5)(.05)
= 43.735
That is K = sum of the expected value of the largest elements in each column
of the
payoff matrix.
= the expected value of the payoffs for all the best courses of action.
Or
The expected opportunity cost for a given courses of action= K(43.735)-Expected pay of for
each
courses of action -
(eq1)
190
MBA-H2040 Quantitative Techniques
for Managers
Thus, it is obvious from the above equation eq1 that the maximum value of EP will
simultaneously produce, the minimum of EOC. The two analyses namely, the payoff matrix
method and the opportunity cost matrix method produce the same result.
Suppose, if the original matrix is in terms of costs it can by similar reasoning, be
shown that the above relationship (eq1) will be of the following form:
EOC = EP ? K
In this case EP is in terms of costs.
4.6 Summary
In this lesson various deterministic inventory models have been developed for various
operating conditions. Here we discussed single item inventory and as well as multi item
inventory models. In this lesson we also discussed probabilistic discrete demand models for
single period inventory items.
4.7 Key Terms
Inventory ? stores of goods or stocks.
Ordering Cost - Cost involved in placing an order.
Procurement Cost ? Same as ordering cost.
Replenishment Cost ? Same as ordering cost.
Set up Cost ? Cost associated with the setting of machine for production.
Shortage Cost ? Costs associated with the demand when stocks have been depleted, it
is
generally called as back order costs.
Safety Stock ? Extra Stocks.
Perishable Product ? The inventories that deteriorate with time.
Deterministic Model ? An inventory model where all the factors are completely
known.
Discrete Probability Distribution ? A probability distribution in which the variable
is allowed
to take only limited number of values.
Expected Opportunity Cost ? Expected value of the variable indicating opportunity
costs.
Expected payoff ? Expected value of the variable indicating payoffs.
Expected Value ? The average value or mean.
Minimum Value ? This is also known as safety stock or the buffer stock.
Maximum Value ? Level of inventory beyond which inventory is not allowed.
Payoff ? The benefit which accrues form a given combination of decision alternative
courses of
action and state of nature.
Reorder Level ? The stock level which is sufficient for the lead time consumption,
and an order
is initiated when inventory dips to this level.
Under Stocking Cost ? Cost relating to the out of stock situation under the
probabilistic
191
MBA-H2040 Quantitative Techniques
for Managers
situation.
Over Stocking Cost ? This is the cost of keeping more units than demanded.
Opportunity Cost Matrix ? Matrix of opportunity Costs.
4.8 Self Assessment Questions
Q1. A Production unit uses Rs.10,000 worth of an item during the year. The production units
estimated the ordering cost as Rs.25 per order and holding cost as 12.5 percent of the average
inventory value. Determine the optimal order size, number of orders per year, time period per
order and total cost.
Answer
Order Size = Q* = Rs.2000
No. of orders per year = N = 5
Time period per order = t* = 73 days
Total Cost = TC* = Rs.250
Q2. The usage of an inventory item each costing Re 1, is 10000 units/year and the ordering
cost is Rs.10, carrying charge is 20% based on the average inventory per year, stock out cost
is Rs.5 per unit of shortage incurred. Determine EOQ, inventory level, shortage level, cycle
period, number of order per year and the total cost.
Answer
EOQ = Q* = 1020 units
Inventory Level = I* = 980 units
Shortage Level = 40 units
Cycle Period = t* = 37 days
No. of orders/year = 10
Total Cost = Rs.400
Q3. The demand for a unit of item is at the rate of 200 per day and can be produced at a rate
of 800 per day. It costs Rs.5000 to set up the production process and Rs.0.2 per unit per day
held in inventory based on the actual inventory any time. Assume that the shortage is not
allowed. Find out the minimum cost and the optimum number of units per production run.
Answer
Hint: D=200 P=800 Co=Rs.5000 Ch=Rs.0.2
Q* = 3651 units
TC* = Rs.547.7
Q4. The demand for an item is 2400 units per year. The ordering cost is Rs.100, inventory
holding cost is 24 percent of the purchase price per year. Determine the optimum purchase
quantity if the purchase prices are as follows:
192
MBA-H2040 Quantitative Techniques
for Managers
P1 = Rs.10 for purchasing Q1 < 500
P2 = Rs.9.25 for purchasing 500 Q2 < 750
P3 = Rs.8.75 for purchasing 750 Q3
Answer
Economic Purchase Quantity = EPQ = Q* = 750 units
Q5. A company follows the following procurement pattern of five items irrespective of their
level of demand. Reduce the inventory levels while keeping the same total number of orders
per annum.
Item
Demand/Year Number of
Order size ($)
Average
($)
orders/year
Inventory
1
1000,000
5
250,000
125,000
2
640,000
5
160,000
80,000
3
90,000
5
22,500
11,250
4
2,500
5
625
350
5
1,600
5
400
200
Answer
According to the company policy, ordering five times a year each item, total average
inventory becomes $.216800.
But after the analysis of ordering quantity the average inventory becomes $.7698.32,
which is much less, at the same time the number of orders almost remain same. Thus,
substantial savings can still be achieved when cost information is not known.
Q6. Suppose the carrying cost is 30% per unit/year, unit price is Rs.4, and the ordering cost is
Rs.30 per order for an item used in an organization in the following pattern:
Month
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Demand 200
220
150
170
210
200
180
220
170
200
160
140
Determine the ordering schedule and the total inventory cost using the following method.
1. Prescribed Rule Method
2. Fixed EOQ Method
Q7. The following numbers indicate the annual values in dollars of some thirty inventory
items of materials selected at random. Carry out an ABC analysis and list out the values of
three items viz. A-items, B-items and C-items.
1
2
3
9
75
3
4
6
13
2
3
12
193
MBA-H2040 Quantitative Techniques
for Managers
100
2
7
40
15
55
1
12
25
15
8
10
1
20
30
1
4
5
194
MBA-H2040 Quantitative Techniques
for Managers
Q8. An organization stoking two products. The organization has limited storage space and
can't store more than 40 units. The following are the demand distribution for the products:
Product1 Product2
Demand Probability of Demand Demand Probability of Demand
10 0.10 10 0.05
20 0.20 20 0.20
30 0.35 30 0.30
40 0.25 40 0.13
50 0.10 50 0.10
If the inventory holding cost is Rs.10 (product 1) and Rs15 (product 2) per unit of the ending
inventories, the shortage costs are Rs.20 and Rs.50 per unit at the ending shortage for the first
and second products respectively. Determine the economic order quantities for both the
products.
4.9 Further References
Hamdy A Taha, 1999. Introduction to Operations Research, PHI Limited, New Delhi.
Mustafi, C.K. 1988. Operations Research Methods and Practices, Wiley Eastern Limited,
New Delhi.
Levin, R and Kirkpatrick, C.A. 1978. Quantitative Approached to Management, Tata
McGraw Hill, Kogakusha Ltd., International Student Edition.
Peterson R and Silver, E. A. 1979. Decision Systems for Inventory Management and
Production Planning.
Handley, G and T.N. Whitin. 1983. Analysis of Inventory Systems, PHI.
Starr, M.K. and D.W. Miller. 1977. Inventory Control Theory and Practice, PHI.
195
MBA-H2040 Quantitative Techniques
for Managers
UNIT III
NETWORK PROBLEMS
Introduction
A network consists of several destinations or jobs which are linked with one another. A
manager will have occasions to deal with some network or other. Certain problems pertaining
to networks are taken up for consideration in this unit.
LESSON 1
SHORTEST PATH PROBLEM
LESSON OUTLINE
The description of a shortest path problem.
The determination of the shortest path.
LEARNING OBJECTIVES
After reading this lesson you should be able to
- understand a shortest path problem
- understand the algorithm for a shortest path problem
-
work out numerical problems
THE PROBLEM
Imagine a salesman or a milk vendor or a post man who has to cover certain previously earmarked places to
perform his daily routines. It is assumed that all the places to be visited by him are connected well for a suitable
mode of transport. He has to cover all the locations. While doing so, if he visits the same place again and again
on the same day, it will be a loss of several resources such as time, money, etc. Therefore he shall place a
constraint upon himself not to visit the same place again and again on the same day. He shall be in a position to
determine a route which would enable him to cover all the locations, fulfilling the constraint.
The shortest route method aims to find how a person can travel from one location to another, keeping
the total distance traveled to the minimum. In other words, it seeks to identify the shortest route to a series of
destinations.
EXAMPLE
Let us consider a real life situation involving a shortest route problem.
A leather manufacturing company has to transport the finished goods from the factory
to the store house. The path from the factory to the store house is through certain
196
MBA-H2040 Quantitative Techniques
for Managers
intermediate stations as indicated in the following diagram. The company executive wants to
identify the path with the shortest distance so as to minimize the transportation cost. The
problem is to achieve this objective.
95 Store house
2
4
Factory 40
40
35
65
70
6
40
1
100
5
20
3
Linkages from Factory to Store house
The shortest route technique can be used to minimize the total distance from a node designated as the
starting node or origin to another node designated as the final node.
In the example under consideration, the origin is the factory and the final node is the store house.
STEPS IN THE SHORTEST ROUTE TECHNIQUE
The procedure consists of starting with a set containing a node and enlarging the set by choosing a node in each
subsequent step.
Step 1:
First, locate the origin. Then, find the node nearest to the origin. Mark the distance between the origin and the
nearest node in a box by the side of that node.
In some cases, it may be necessary to check several paths to find the nearest node.
Step 2:
Repeat the above process until the nodes in the entire network have been accounted for. The last distance placed
in a box by the side of the ending node will be the distance of the shortest route. We note that the distances
indicated in the boxes by each node constitute the shortest route to that node. These distances are used as
intermediate results in determining the next nearest node.
SOLUTION FOR THE EXAMPLE PROBLEM
Looking at the diagram, we see that node 1 is the origin and the nodes 2 and 3 are neighbours
to the origin. Among the two nodes, we see that node 2 is at a distance of 40 units from node
1 whereas node 3 is at a distance of 100 units from node 1. The minimum of {40, 100} is 40.
Thus, the node nearest to the origin is node 2, with a distance of 40 units. So, out of the two
nodes 2 and 3, we select node 2. We form a set of nodes {1, 2} and construct a path
connecting the node 2 with node 1 by a thick line and mark the distance of 40 in a box by the
side of node 2. This first iteration is shown in the following diagram.
40
197
MBA-H2040 Quantitative Techniques
for Managers
95 Store house
2
4
40
40
6
Factory
35
65 70
1
40
5
100
3
20
ITERATION No. 1
Now we search for the next node nearest to the set of nodes {1, 2}. For this purpose, consider those
nodes which are neighbours of either node 1 or node 2. The nodes 3, 4 and 5 fulfill this condition. We calculate
the following distances.
The distance between nodes 1 and 3 = 100.
The distance between nodes 2 and 3 = 35.
The distance between nodes 2 and 4 = 95.
The distance between nodes 2 and 5 = 65.
Minimum of {100, 35, 95, 65} = 35.
Therefore, node 3 is the nearest one to the set {1, 2}. In view of this observation, the set of nodes is enlarged
from {1, 2} to {1, 2, 3}. For the set {1, 2, 3}, there are two possible paths, viz. Path 1 2 3 and Path 1
3 2. The Path 1 2 3 has a distance of 40 + 35 = 75 units while the Path 1 3 2 has a distance of
100 + 35 = 135 units.
Minimum of {75, 135} = 75. Hence we select the path 1 2 3 and display this path by thick edges. The
distance 75 is marked in a box by the side of node 3. We obtain the following diagram at the end of Iteration
No. 2.
40
95 Store house
2
4
Factory
40
40
6
35
65
70
1
40
5
100
3
20
75
ITERATION No. 2
REPEATING THE PROCESS
We repeat the process. The next node nearest to the set {1, 2, 3} is either node 4 or node 5.
Node 4 is at a distance of 95 units from node 2 while node 2 is at a distance of 40 units
from node 1. Thus, node 4 is at a distance of 95 + 40 = 135 units from the origin.
198
MBA-H2040 Quantitative Techniques
for Managers
As regards node 5, there are two paths viz. 2 5 and 3 5, providing a link to the origin. We already
know the shortest routes from nodes 2 and 3 to the origin. The minimum distances have been indicated in boxes
near these nodes. The path 3 5 involves the shortest distance. Thus, the distance between nodes 1 and 5 is 95
units (20 units between nodes 5 and 3 + 75 units between node 3 and the origin). Therefore, we select node 5
and enlarge the set from {1, 2, 3} to {1, 2, 3, 5}. The distance 95 is marked in a box by the side of node 5. The
following diagram is obtained at the end of Iteration No. 3.
40
95 Store house
2
4
40
40
6
35
65
70
1
40
5
Factory 100
3
20 95
75
ITERATION No. 3
Now 2 nodes remain, viz., nodes 4 and 6. Among them, node 4 is at a distance of 135 units from the
origin (95 units from node 4 to node 2 + 40 units from node 2 to the origin). Node 6 is at a distance of 135
units from the origin (40 + 95 units). Therefore, nodes 4 and 6 are at equal distances from the origin. If we
choose node 4, then travelling from node 4 to node 6 will involve an additional distance of 40 units. However,
node 6 is the ending node. Therefore, we select node 6 instead of node 4. Thus the set is enlarged from {1, 2, 3,
5} to {1, 2, 3, 5, 6}. The distance 135 is marked in a box by the side of node 6. Since we have got a path
beginning from the start node and terminating with the stop node, we see that the solution to the given problem
has been obtained. We have the following diagram at the end of Iteration No. 4.
40
95 Store house
2
4
40
40
6
35
65
70
1
40
135
5
Factory
100
3
20
95
75
ITERATION No. 4
MINIMUM DISTANCE
Referring to the above diagram, we see that the shortest route is provided by the path 1 2
3 5 6 with a minimum distance of 135 units.
199
MBA-H2040 Quantitative Techniques
for Managers
QUESTIONS
1. Explain the shortest path problem.
2. Explain the algorithm for a shortest path problem
3. Find the shortest path of the following network:
30
3
5
40
40
30 50 30
1
45
6
25
35
2
4
4. Determine the shortest path of the following network:
2
16
5
7
9 7
1
4
15
6
8 4
25
3
LESSON 2
MINIMUM SPANNING TREE PROBLEM
LESSON OUTLINE
The description of a minimum spanning tree problem.
The identification of the minimum spanning tree.
LEARNING OBJECTIVES
After reading this lesson you should be able to
-
understand a minimum spanning tree problem
-
understand the algorithm for minimum spanning tree problem
-
locate the minimum spanning tree
200
MBA-H2040 Quantitative Techniques
for Managers
-
carry out numerical problems
Tree: A minimally connected network is called a tree. If there are n nodes in a network, it
will be a tree if the number of edges = n-1.
Minimum spanning tree algorithm
Problem : Given a connected network with weights assigned to the edges, it is required to
find out a tree whose nodes are the same as those of the network.
The weight assigned to an edge may be regarded as the distance between the two
nodes with which the edge is incident.
Algorithm:
The problem can be solved with the help of the following algorithm.
The procedure consists of selection of a node at each step.
Step 1: First select any node in the network. This can be done arbitrarily. We will start with
this node.
Step 2: Connect the selected node to the nearest node.
Step 3: Consider the nodes that are now connected. Consider the remaining nodes. If there is
no node remaining, then stop. On the other hand, if some nodes remain, among them find out
which one is nearest to the nodes that are already connected. Select this node and go to Step
2.
Thus the method involves the repeated application of Steps 2 and 3. Since the number
of nodes in the given network is finite, the process will end after a finite number of steps. The
algorithm will terminate with step 3.
How to break ties:
While applying the above algorithm, if some nodes remain in step 3 and if there is a tie in the
nearest node, then the tie can be broken arbitrarily.
As a consequence of tie, we may end up with more than one optimal solution.
Problem 1:
Determine the minimum spanning tree for the following network.
60
5
2
70
60 60 80
100
7
1
3
40 120 50
8
201
4
6
MBA-H2040 Quantitative Techniques
for Managers
80 50 60
30
90
Solution:
Step 1: First select node 1. (This is done arbitrarily)
Step 2: We have to connect node 1 to the nearest node. Nodes 2, 3 and 4 are adjacent to node
1. They are at distances of 60, 40 and 80 units from node 1. Minimum of {60, 40, 80} = 40.
Hence the shortest distance is 40. This corresponds to node 3. So we connect node 1 to node
3 by a thick line. This is Iteration No. 1.
60
5
2 70
60 60 80
100
7
1
3
40 120 50
8
80 50 60
30
90
6
4
Iteration No. 1
Step 3: Now the connected nodes are 1 and 3. The remaining nodes are 2, 4, 5, 6, 7 and 8.
Among them, nodes 2 and 4 are connected to node 1. They are at distances of 60 and 80 from
node 1. Minimum of {60, 80} = 60. So the shortest distance is 60. Next, among the nodes 2,
4, 5, 6, 7 and 8, find out which nodes are connected to node 3. We find that all of them are
connected to node 3. They are at distances of 60, 50, 80, 60, 100 and 120 from node 3.
Minimum of {60, 50, 80, 60, 100, 120} = 50. Hence the shortest distance is 50.
Among these nodes, it is seen that node 4 is nearest to node 3.
Now we go to Step 2. We connect node 3 to node 4 by a thick line. This is Iteration
No.2.
60
5
2
70
60 60 80
100
7
1
3
202
MBA-H2040 Quantitative Techniques
for Managers
40 120 50
8
80 50 60
30
90
6
4
Iteration No. 2
Next go to step 3.
Now the connected nodes are 1, 3 and 4. The remaining nodes are 2, 5, 6, 7 and 8.
Node 2 is at a distance of 60 from node 1. Nodes 5, 6, 7 and 8 are not adjacent to node 1. All
of the nodes 2, 5, 6, 7 and 8 are adjacent to node 3. Among them, nodes 2 and 6 are nearer to
node 3, with equal distance of 60.
Node 6 is adjacent to node 4, at a distance of 90. Now there is a tie between nodes 2
and 6. The tie can be broken arbitrarily. So we select node 2. Connect node 3 to Node 2 by
a thick line. This is Iteration No. 3.
60
5
2
70
60 60 80
100
7
1
3
40 120 50
8
80 50 60
30
90
6
4
Iteration No. 3
We continue the above process.
Now nodes 1, 2, 3 and 4 are connected. The remaining nodes are 5, 6, 7 and 8. None
of them is adjacent to node 1. Node 5 is adjacent to node 2 at a distance of 60. Node 6 is at a
distance of 60 from node 3. Node 6 is at a distance of 90 from node 4. There is a tie between
203
MBA-H2040 Quantitative Techniques
for Managers
nodes 5 and 6. We select node 5. Connect node 2 to node 5 by a thick line. This is Iteration
No. 4.
60
5
2
70
60 60 80
100
7
1
3
40 120 50
8
80 50 60
30
90
6
4
Iteration No. 4
Now nodes 1, 2, 3, 4 and 5 are connected. The remaining nodes are 6, 7 and 8. Among them,
node 6 is at the shortest distance of 60 from node 3. So, connect node 3 to node 6 by a thick
line. This is Iteration No. 5.
60
5
2
70
60 60 80
100
7
1
3
40 120 50
8
80 50 60
30
90
6
4
Iteration No. 5
Now nodes 1, 2, 3, 4, 5 and 6 are connected. The remaining nodes are 7 and 8. Among them,
node 8 is at the shortest distance of 30 from node 6. Consequently we connect node 6 to node
8 by a thick line. This is Iteration No. 6.
204
MBA-H2040 Quantitative Techniques
for Managers
60
5
2
70
60 60 80
100
7
1
3
40 120 50
8
80 50 60
30
90
6
4
Iteration No. 6
Now nodes 1, 2, 3, 4, 5, 6 and 8 are connected. The remaining node is 7. It is at the shortest
distance of 50 from node 8. So, connect node 8 to node 7 by a thick line. This is Iteration
No.7.
60
5
2
70
60 60 80
100
7
1
3
40 120 50
8
80 50 60
30
90
6
4
Iteration No. 7
Now all the nodes 1, 2, 3, 4, 5, 6, 7 and 8 are connected by seven thick lines. Since no node
is remaining, we have reached the stopping condition. Thus we obtain the following minimum
spanning tree for the given network.
60
5
2
60
7
1
3
205
8
MBA-H2040 Quantitative Techniques
for Managers
40 50
50 60
30
Minimum Spanning Tree
QUESTIONS
1. Explain the minimum spanning tree algorithm.
2. From the following network, find the minimum spanning tree.
75
6
2
80 55 90
100
1
3
70 40
25 60
5
30
4
3. Find the minimum spanning tree of the following network:
12
5
2 15
5 8
2 10 13
8
1
3
6
9 4 10
5
7
4
4
206
MBA-H2040 Quantitative Techniques
for Managers
207
MBA-H2040 Quantitative Techniques
for Managers
LESSON 3
PROJECT NETWORK
LESSON OUTLINE
The key concepts
Construction of project network diagram
LEARNING OBJECTIVES
After reading this lesson you should be able to
- understand the definitions of important terms
-
understand the development of project network diagram
-
work out numerical problems
KEY CONCEPTS
Certain key concepts pertaining to a project network are described below:
1. Activity
An activity means a work. A project consists of several activities. An activity takes time. It
is represented by an arrow in a diagram of the network. For example, an activity in house
construction can be flooring. This is represented as follows:
flooring
Construction of a house involves various activities. Flooring is an activity in this project. We
can say that a project is completed only when all the activities in the project are completed.
2. Event
It is the beginning or the end of an activity. Events are represented by circles in a project
network diagram. The events in a network are called the nodes. Example:
Start Stop
Punching
Starting a punching machine is an activity. Stopping the punching machine is another activity.
3. Predecessor Event
The event just before another event is called the predecessor event.
4. Successor Event
The event just following another event is called the successor event.
Example: Consider the following.
3
208
1
2
4
6
MBA-H2040 Quantitative Techniques
for Managers
In this diagram, event 1 is predecessor for the event 2.
Event 2 is successor to event 1.
Event 2 is predecessor for the events 3, 4 and 5.
Event 4 is predecessor for the event 6.
Event 6 is successor to events 3, 4 and 5.
5. Network
A network is a series of related activities and events which result in an end product or service.
The activities shall follow a prescribed sequence. For example, while constructing a house,
laying the foundation should take place before the construction of walls. Fitting water tapes
will be done towards the completion of the construction. Such a sequence cannot be altered.
6. Dummy Activity
A dummy activity is an activity which does not consume any time. Sometimes, it may be
necessary to introduce a dummy activity in order to provide connectivity to a network or for
the preservation of the logical sequence of the nodes and edges.
7. Construction of a Project Network
A project network consists of a finite number of events and activities, by adhering to a certain
specified sequence. There shall be a start event and an end event (or stop event). All the
other events shall be between the start and the end events. The activities shall be marked by
directed arrows. An activity takes the project from one event to another event.
An event takes place at a point of time whereas an activity takes place from one point
of time to another point of time.
CONSTRUCTION OF PROJECT NETWORK DIAGRAMS
Problem 1:
Construct the network diagram for a project with the following activities:
Activity
Name of
Immediate
209
MBA-H2040 Quantitative Techniques
for Managers
EventEvent
Activity
Predecessor
Activity
12
A
-
13
B
-
14
C
-
25
D
A
36
E
B
46
F
C
56
G
D
Solution:
The start event is node 1.
The activities A, B, C start from node 1 and none of them has a predecessor activity. A joins
nodes1 and 2; B joins nodes 1 and 3; C joins nodes 1 and 4. So we get the following:
A
2
1
3
B
C
4
This is a part of the network diagram that is being constructed.
Next, activity D has A as the predecessor activity. D joins nodes 2 and 5. So we get
A D
1
2
5
Next, activity E has B as the predecessor activity. E joins nodes 3 and 6. So we get
B E
1
3
6
Next, activity G has D as the predecessor activity. G joins nodes 5 and 6. Thus we obtain
D G
2
5
6
Since activities E, F, G terminate in node 6, we get
5
G
3
6
E
4
210
MBA-H2040 Quantitative Techniques
for Managers
F
6 is the end event.
Combining all the pieces together, the following network diagram is obtained for the given
project:
D
5
2
A G
Start event End event
B E
1
3
6
C F
4
We validate the diagram by checking with the given data.
Problem 2:
Develop a network diagram for the project specified below:
Activity
Immediate
Predecessor Activity
A
-
B
A
C, D
B
E
C
F
D
G
E, F
Solution:
Activity A has no predecessor activity. i.e., It is the first activity. Let us suppose that activity
A takes the project from event 1 to event 2. Then we have the following representation for A:
211
MBA-H2040 Quantitative Techniques
for Managers
A
1
2
For activity B, the predecessor activity is A. Let us suppose that B joins nodes 2 and 3. Thus
we get
A B
1
2
3
Activities C and D have B as the predecessor activity. Therefore we obtain the following:
C
B
4
2
3
D
5
Activity E has D as the predecessor activity. So we get
C E
3
4
6
Activity F has D as the predecessor activity. So we get
D F
3
5
6l
"
Activity G has E and F as predecessor activities. This is possible only if nodes 6 and 6l are
one and the same. So, rename node 6l as node 6. Then we get
D F
3
5
6!
and
4
E
6
7
G
5
212
MBA-H2040 Quantitative Techniques
for Managers
F
G is the last activity.
Putting all the pieces together, we obtain the following diagram the project network:
Start event C
4 E End event
A B G
1
2
3
6
7
D F
5
The diagram is validated by referring to the given data.
Note: An important point may be observed for the above diagram. Consider the following
parts in the diagram
C E
3
4
6
and
D F
3
5
6l
"
We took nodes 6 and 6l as one and the same. Instead, we can retain them as different nodes.
Then, in order to provide connectivity to the network, we join nodes 6l and 6 by a dummy
activity. Then we arrive at the following diagram for the project network:
4
Start event C E
6
G
1
2
3
A B dummy
activity
7
D F End event
5
6l
!
213
MBA-H2040 Quantitative Techniques
for Managers
QUESTIONS:
1. Explain the terms: event, predecessor event, successor event, activity, dummy
activity, network.
2. Construct the network diagram for the following project:
Activity
Immediate
Predecessor Activity
A
-
B
-
C
A
D
B
E
A
F
C, D
G
E
H
E
I
F, G
J
H, I
214
MBA-H2040 Quantitative Techniques
for Managers
LESSON 4
CRITICAL PATH METHOD (CPM)
LESSON OUTLINE
The concepts of critical path and critical activities
Location of the critical path
Evaluation of the project completion time
LEARNING OBJECTIVES
After reading this lesson you should be able to
-
understand the definitions of critical path and critical activities
- identify critical path and critical activities
- determine the project completion time
INTRODUCTION
The critical path method (CPM) aims at the determination of the time to complete a project
and the important activities on which a manager shall focus attention.
ASSUMPTION FOR CPM
In CPM, it is assumed that precise time estimate is available for each activity.
PROJECT COMPLETION TIME
From the start event to the end event, the time required to complete all the activities of the
project in the specified sequence is known as the project completion time.
PATH IN A PROJECT
A continuous sequence, consisting of nodes and activities alternatively, beginning with the
start event and stopping at the end event of a network is called a path in the network.
CRITICAL PATH AND CRTICAL ACTIVITIES
Consider all the paths in a project, beginning with the start event and stopping at the end
event. For each path, calculate the time of execution, by adding the time for the individual
activities in that path.
The path with the largest time is called the critical path and the activities along this
path are called the critical activities or bottleneck activities. The activities are called critical
215
MBA-H2040 Quantitative Techniques
for Managers
because they cannot be delayed. However, a non-critical activity may be delayed to a certain
extent. Any delay in a critical activity will delay the completion of the whole project.
However, a certain permissible delay in a non ?critical activity will not delay the completion
of the whole project. It shall be noted that delay in a non-critical activity beyond a limit
would certainly delay the completion the whole project. Sometimes, there may be several
critical paths for a project. A project manager shall pay special attention to critical activities.
Problem 1:
The following details are available regarding a project:
Activity
Predecessor
Duration (Weeks)
Activity
A
-
3
B
A
5
C
A
7
D
B
10
E
C
5
F
D,E
4
Determine the critical path, the critical activities and the project completion time.
Solution:
First let us construct the network diagram for the given project. We mark the time estimates
along the arrows representing the activities. We obtain the following diagram:
Start event B
3 D End event
5 10
A
1
2
3 F
5
6
C 4
7 E
4 5
Consider the paths, beginning with the start node and stopping with the end node. There are
two such paths for the given project. They are as follows:
Path I
216
MBA-H2040 Quantitative Techniques
for Managers
A B D F
1
2
3
5
6
3 5 10 4
with a time of 3 + 5 + 10 + 4 = 22 weeks.
Path II
A C E F
1
2
4
5
6
3 7 5 4
with a time of 3 + 7 + 5 + 4 = 19 weeks.
Compare the times for the two paths. Maximum of {22,19} = 22. We see that path I has the
maximum time of 22 weeks. Therefore, path I is the critical path. The critical activities are A,
B, D and F. The project completion time is 22 weeks.
We notice that C and E are non- critical activities.
Time for path I - Time for path II = 22- 19 = 3 weeks.
Therefore, together the non- critical activities can be delayed upto a maximum of 3
weeks, without delaying the completion of the whole project.
Problem 2:
Find out the completion time and the critical activities for the following project:
D
5
2 20
A 8 G 8
B E H 11 K 6
1
3
6
8
10
10 16 I 14 L 5
C 7 J
9
F
7 10
4
25
Solution:
In all, we identify 4 paths, beginning with the start node of 1 and terminating at the end node
of 10. They are as follows:
Path I
A D G K
1
2
5
8
10
217
MBA-H2040 Quantitative Techniques
for Managers
8 20 8 6
Time for the path = 8 + 20 + 8 + 6 = 42 units of time.
Path II
B E H K
1
3
6
8
10
10 16 11 6
Time for the path = 10 + 16 + 11 + 6 = 43 units of time.
Path III
B E I L
1
3
6
9
10
10 16 14 5
Time for the path = 10 + 16 + 14 + 5 = 45 units of time.
Path IV
C F J L
1
4
7
9
10
7 25 10 5
Time for the path = 7 + 25 + 10 + 5 = 47 units of time.
Compare the times for the four paths. Maximum of {42, 43, 45, 47} = 47. We see that the
following path has the maximum time and so it is the critical path:
C F J L
1
4
7
9
10
7 25 10 5
The critical activities are C, F, J and L. The non-critical activities are A, B, D, E, G, H, I and
K. The project completion time is 47 units of time.
Problem 3:
Draw the network diagram and determine the critical path for the following project:
Activity
Time estimate (Weeks)
1- 2
5
1- 3
6
218
MBA-H2040 Quantitative Techniques
for Managers
1- 4
3
2 -5
5
3 -6
7
3 -7
10
4 -7
4
5 -8
2
6 -8
5
7 -9
6
8 -9
4
Solution: We have the following network diagram for the project:
D
5
2 5
2 H
A
5
1 B E
6 I
8 K
3
6 7 5 4
9
3 C F 10 J
6
7
G
4
4
Solution:
We assert that there are 4 paths, beginning with the start node of 1 and terminating at the end
node of 9. They are as follows:
Path I
A D H K
1
2
5
8
9
5 5 2 4
Time for the path = 5 + 5 + 2 + 4 = 16 weeks.
Path II
B E I K
1
3
6
8
9
219
MBA-H2040 Quantitative Techniques
for Managers
6 7 5 4
Time for the path = 6 + 7 + 5 + 4 = 22 weeks.
Path III
B F J
1
3
7
9
6 10 6
Time for the path = 6 + 10 + 6 = 16 weeks.
Path IV
C G J
1
4
7
9
3 4 6
Time for the path = 3 + 4 + 6 = 13 weeks.
Compare the times for the four paths. Maximum of {16, 22, 16, 13} = 22. We see that the
following path has the maximum time and so it is the critical path:
B E I K
1
3
6
8
9
6 7 5 4
The critical activities are B, E, I and K. The non-critical activities are A, C, D, F, G, H and J.
The project completion time is 22 weeks.
220
MBA-H2040 Quantitative Techniques
for Managers
QUESTIONS:
1. Explain the terms: critical path, critical activities.
2. The following are the time estimates and the precedence relationships of the activities
in a project network:
Activity
IMMEDIATE
time estimate
Predecessor
(weeks)
Activity
A
-
4
B
-
7
C
-
3
D
A
6
E
B
4
F
B
7
G
C
6
H
E
10
I
D
3
J
F, G
4
K
H, I
2
Draw the project network diagram. Determine the critical path and the project completion
time.
221
MBA-H2040 Quantitative Techniques
for Managers
LESSON 5
PERT
LESSON OUTLINE
The concept of PERT
Estimates of the time of an activity
Determination of critical path
Probability estimates
LEARNING OBJECTIVES
After reading this lesson you should be able to
- understand the importance of PERT
- locate the critical path
- determine the project completion time
- find out the probability of completion of a project before a stipulated time
INTRODUCTION
Programme Evaluation and Review Technique (PERT) is a tool that would help a project
manager in project planning and control. It would enable him in continuously monitoring a
project and taking corrective measures wherever necessary. This technique involves
statistical methods.
ASSUMPTIONS FOR PERT
Note that in CPM, the assumption is that precise time estimate is available for each activity in
a project. However, one finds most of the times that this is not practically possible.
In PERT, we assume that it is not possible to have precise time estimate for each
activity and instead, probabilistic estimates of time alone are possible. A multiple time
estimate approach is followed here. In probabilistic time estimate, the following 3 types of
estimate are possible:
1. Pessimistic time estimate ( t )
p
2. Optimistic time estimate ( t )
o
3. Most likely time estimate ( t )
m
The optimistic estimate of time is based on the assumption that an activity will not involve
any difficulty during execution and it can be completed within a short period. On the other
hand, a pessimistic estimate is made on the assumption that there would be unexpected
222
MBA-H2040 Quantitative Techniques
for Managers
problems during the execution of an activity and hence it would consume more time. The
most likely time estimate is made in between the optimistic and the pessimistic estimates of
time. Thus the three estimates of time have the relationship
t t t .
o
m
p
Practically speaking, neither the pessimistic nor the optimistic estimate may hold in
reality and it is the most likely time estimate that is expected to prevail in almost all cases.
Therefore, it is preferable to give more weight to the most likely time estimate.
We give a weight of 4 to most likely time estimate and a weight of 1 each to the
pessimistic and optimistic time estimates. We arrive at a time estimate ( t ) as the weighted
e
average of these estimates as follows:
t 4 t t
o
m
p
t
e
6
Since we have taken 6 units ( 1 for t , 4 for t and 1 for t ), we divide the sum by 6. With
p
m
o
this time estimate, we can determine the project completion time as applicable for CPM.
Since PERT involves the average of three estimates of time for each activity, this
method is very practical and the results from PERT will be have a reasonable amount of
reliability.
MEASURE OF CERTAINTY
The 3 estimates of time are such that
t t t .
o
m
p
Therefore the range for the time estimate is t t .
p
o
The time taken by an activity in a project network follows a distribution with a
standard deviation of one sixth of the range, approximately.
t t
i.e., The standard deviation =
p
o
6
2
t t
and the variance = 2
p
o
6
The certainty of the time estimate of an activity can be analysed with the help of the
variance. The greater the variance, the more uncertainty in the time estimate of an activity.
Problem 1:
Two experts A and B examined an activity and arrived at the following time estimates.
Expert
Time Estimate
223
MBA-H2040 Quantitative Techniques
for Managers
t
t
t
o
m
p
A
4
6
8
B
4
7
10
Determine which expert is more certain about his estimates of time:
Solution:
2
t t
Variance ( 2
) in time estimates = p o
6
2
8 4
4
In the case of expert A, the variance =
6
9
2
10 4
As regards expert B, the variance =
1
6
So, the variance is less in the case of A. Hence, it is concluded that the expert A is more
certain about his estimates of time.
Determination of Project Completion Time in PERT
Problem 2:
Find out the time required to complete the following project and the critical activities:
Activity
Predecessor
Optimistic time
Most likely time
Pessimistic time
Activity
estimate (to days)
estimate (tm days)
estimate (tp days)
A
-
2
4
6
B
A
3
6
9
C
A
8
10
12
D
B
9
12
15
E
C
8
9
10
F
D, E
16
21
26
G
D, E
19
22
25
H
F
2
5
8
I
G
1
3
5
Solution:
From the three time estimates t , t and t , calculate t for each activity. We obtain the following table:
p
m
o
e
Activity
Optimistic
4 x Most likely
Pessimistic
to+ 4tm+ tp
Time estimate
time estimate
time estimate
time estimate
t 4 t t
(t
o
m
p
o)
(tp)
t
e
6
A
2
16
6
24
4
B
3
24
9
36
6
C
8
40
12
60
10
D
9
48
15
72
12
E
8
36
10
54
9
F
16
84
26
126
21
224
MBA-H2040 Quantitative Techniques
for Managers
G
19
88
25
132
22
H
2
20
8
30
5
I
1
12
5
18
3
Using the single time estimates of the activities, we get the following network diagram for the project.
B
3 D F H 6
6 12 21 5
A
4
1
2
C 10 E G
5 I
8
9 22 3
4
Consider the paths, beginning with the start node and stopping wit
7 h the end node. There are
four such paths for the given project. They are as follows:
Path I
A B D F H
1
2
3
5
6
8
4 6 12 21 5
Time for the path: 4+6+12+21+5 = 48 days.
Path II
A B D G I
1
2
3
5
7
8
4 6 12 6 3
6
Time for the path: 4+6+12+ 6+3 = 31 days.
Path III
A C E F H
1
2
4
5
6
8
4 10 9 21 5
3
7
Time for the path: 4+10+9+ 21+5 = 49 days.
Path IV
A C E G I
1
2
4
5
7
8
4 10 9 6 3
6
Time for the path: 4+10+9+ 6+3 = 32 days.
Compare the times for the four paths.
Maximum of {48, 31, 49, 32} = 49.
We see that Path III has the maximum time.
Therefore the critical path is Path III. i.e., 1 2 4 5 6 8.
The critical activities are A, C, E, F and H.
The non-critical activities are B, D, G and I.
Project time (Also called project length) = 49 days.
Problem 3:
Find out the time, variance and standard deviation of the project with the following time estimates in weeks:
225
MBA-H2040 Quantitative Techniques
for Managers
Activity
Optimistic time
Most likely time
Pessimistic time
estimate (to)
estimate (tm)
estimate (tp)
1-2
3
6
9
1-6
2
5
8
2-3
6
12
18
2-4
4
5
6
3-5
8
11
14
4-5
3
7
11
6-7
3
9
15
5-8
2
4
6
7-8
8
16
18
Solution:
From the three time estimates t , t and t , calculate t for each activity. We obtain the following table:
p
m
o
e
Activity
Optimistic
4 x Most likely Pessimistic
to+ 4tm+ tp
Time estimate
time estimate
time estimate time estimate
t 4 t t
(t
o
m
p
o)
(tp)
t
e
6
1-2
3
24
9
36
6
1-6
2
20
8
30
5
2-3
6
48
18
72
12
2-4
4
20
6
30
5
3-5
8
44
14
66
11
4-5
3
28
11
42
7
6-7
3
36
15
54
9
5-8
2
16
6
24
4
7-8
8
64
18
90
15
With the single time estimates of the activities, we get the following network diagram for the project.
C
3 F
12 11
2
D 5 G I
5
A 6 7
4
4
1 5 B H
8
E 15
9
Consider the paths, beginn
6 ing with the start node and stopping with the end node. There are three such paths for
7
the given project. They are as follows:
Path I
A C F I
1
2
3
5
8
226
MBA-H2040 Quantitative Techniques
for Managers
6 12 11 4
Time for the path: 6+12+11+4 = 33 weeks.
Path II
A D G I
1
2
4
5
8
6 5 7
3 4
Time for the path: 6+5+7+ 4= 22 weeks.
Path III
B E H
6
7
1 5 9 15
8
3
Time for the path: 5+9+15 = 29 weeks.
Compare the times for the three paths.
Maximum of {33, 22, 29} = 33.
It is noticed that Path I has the maximum time.
Therefore the critical path is Path I. i.e., 1 2 3 5 8
The critical activities are A, C, F and I.
The non-critical activities are B, D, G and H.
Project time = 33 weeks.
Calculation of Standard Deviation and Variance for the Critical Activities:
Critical
Optimistic
Most likely Pessimistic
Range
Standard
Variance
Activity
time
time
time
(tp - to)
deviation =
2
t t
estimate
estimate
estimate
t t
2
p
o
p
o
(t
o)
(tm)
(tp)
6
6
A: 12
3
6
9
6
1
1
C: 23
6
12
18
12
2
4
F: 35
8
11
14
6
1
1
I: 58
2
4
6
4
2/3
4/9
Variance of project time (Also called Variance of project length) =
Sum of the variances for the critical activities = 1+4+1+ 4/9 = 58/9 Weeks.
Standard deviation of project time = Variance = 58/9 = 2.54 weeks.
Problem 4
A project consists of seven activities with the following time estimates. Find the probability that the project will
be completed in 30 weeks or less.
Activity
Predecessor
Optimistic time
Most likely time
Pessimistic time
Activity
estimate (to days)
estimate (tm days)
estimate (tp days)
A
-
2
5
8
B
A
2
3
4
C
A
6
8
10
D
A
2
4
6
E
B
2
6
10
F
C
6
7
8
G
D, E, F
6
8
10
Solution:
227
MBA-H2040 Quantitative Techniques
for Managers
From the three time estimates t , t and t , calculate t for each activity. The results are furnished in the
p
m
o
e
following table:
Activity
Optimistic
4 x Most
Pessimistic time
to+ 4tm+ tp
Time estimate
time estimate
likely time
estimate (tp)
t 4 t t
(t
o
m
p
o)
estimate
t
e
6
A
2
20
8
30
5
B
2
12
4
18
3
C
6
32
10
48
8
D
2
16
6
24
4
E
2
24
10
36
6
F
6
28
8
42
7
G
6
32
10
48
8
With the single time estimates of the activities, the following network diagram is constructed for the project.
3
B 3 6 E
C F
4
8 7
A D G
5 4 8
1
2
5
6
Consider the paths, beginning with the start node and stopping with the end node. There are three such paths for
the given project. They are as follows:
Path I
A B E G
1
3
2
5
6
5 3 6
4 8
8
Time for the path: 5+3+6+8 = 22 weeks.
Path II
A C F G
1
2
4
5
6
5 8
4 7 8
8
Time for the path: 5+8+7+ 8 = 28 weeks.
Path III
A D G
1
5
2
6
5 4 8
4
Time for the path: 5+4+8 = 17 weeks.
Compare the times for the three paths.
Maximum of {22, 28, 17} = 28.
It is noticed that Path II has the maximum time.
228
MBA-H2040 Quantitative Techniques
for Managers
Therefore the critical path is Path II. i.e., 1 2 4 5 6.
The critical activities are A, C, F and G.
The non-critical activities are B, D and E.
Project time = 28 weeks.
Calculation of Standard Deviation and Variance for the Critical Activities:
Critical
Optimistic
Most likely
Pessimistic
Range
Standard
Variance
Activity
time
time
time estimate
(tp -to)
deviation =
2
t t
estimate
estimate
(t
2
p
o
p)
t t
p
o
(t
o)
(tm)
6
6
A: 12
2
5
8
6
1
1
C: 24
6
8
10
2
4
3
9
4
F: 45
6
7
8
1
1
3
9
2
G: 56
6
8
10
2
4
4
3
9
Standard deviation of the critical path = 2 = 1.414
The standard normal variate is given by the formula
Given value of t Expected value of t in the critical path
Z
SD for the critical path
30 28
So we get Z
= 1.414
1.414
We refer to the Normal Probability Distribution Table.
Corresponding to Z = 1.414, we obtain the value of 0.4207
We get 0.5 + 0.4207 = 0. 9207
Therefore the required probability is 0.92
i.e., There is 92% chance that the project will be completed before 30 weeks. In other words, the chance that it
will be delayed beyond 30 weeks is 8%
QUESTIONS:
1. Explain how time of an activity is estimated in PERT.
2. Explain the measure of certainty in PERT.
3. The estimates of time in weeks of the activities of a project are as follows:
Activity
Predecessor
Optimistic
Most likely
Pessimistic
Activity
estimate of time
estimate of time
estimate of time
A
-
2
4
6
B
A
8
11
20
C
A
10
15
20
D
B
12
18
24
229
MBA-H2040 Quantitative Techniques
for Managers
E
C
8
13
24
F
C
4
7
16
G
D,F
14
18
28
H
E
10
12
14
I
G,H
7
10
19
Determine the critical activities and the project completion time.
4. Draw the network diagram for the following project. Determine the time, variance and standard deviation of
the project.:
Activity
Predecessor
Optimistic
Most likely
Pessimistic
Activity
estimate of time
estimate of time
estimate of time
A
-
12
14
22
B
-
16
17
24
C
A
14
15
16
D
A
13
18
23
E
B
16
18
20
F
D,E
13
14
21
G
C,F
6
8
10
230
MBA-H2040 Quantitative Techniques
for Managers
5. Consider the following project with the estimates of time in weeks:
Activity
Predecessor
Optimistic
Most likely
Pessimistic
Activity
estimate of time
estimate of time
estimate of time
A
-
2
4
6
B
-
3
5
7
C
A
5
6
13
D
A
4
8
12
E
B,C
5
6
13
F
D,E
6
8
14
Find the probability that the project will be completed in 27 weeks.
231
MBA-H2040 Quantitative Techniques
for Managers
NORMAL DISTRIBUTION TABLE
Area Under Standard Normal Distribution
0.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.0
0.0000 0.0040 0.0080 0.0120 0.0160 0.0199 0.0239 0.0279 0.0319 0.0359
0.1
0.0398 0.0438 0.0478 0.0517 0.0557 0.0596 0.0636 0.0675 0.0714 0.0753
0.2
0.0793 0.0832 0.0871 0.0910 0.0948 0.0987 0.1026 0.1064 0.1103 0.1141
0.3
0.1179 0.1217 0.1255 0.1293 0.1331 0.1368 0.1406 0.1443 0.1480 0.1517
0.4
0.1554 0.1591 0.1628 0.1664 0.1700 0.1736 0.1772 0.1808 0.1844 0.1879
0.5
0.1915 0.1950 0.1985 0.2019 0.2054 0.2088 0.2123 0.2157 0.2190 0.2224
0.6
0.2257 0.2291 0.2324 0.2357 0.2389 0.2422 0.2454 0.2486 0.2517 0.2549
0.7
0.2580 0.2611 0.2642 0.2673 0.2704 0.2734 0.2764 0.2794 0.2823 0.2852
0.8
0.2881 0.2910 0.2939 0.2967 0.2995 0.3023 0.3051 0.3078 0.3106 0.3133
0.9
0.3159 0.3186 0.3212 0.3238 0.3264 0.3289 0.3315 0.3340 0.3365 0.3389
1.0
0.3413 0.3438 0.3461 0.3485 0.3508 0.3531 0.3554 0.3577 0.3599 0.3621
1.1
0.3643 0.3665 0.3686 0.3708 0.3729 0.3749 0.3770 0.3790 0.3810 0.3830
1.2
0.3849 0.3869 0.3888 0.3907 0.3925 0.3944 0.3962 0.3980 0.3997 0.4015
1.3
0.4032 0.4049 0.4066 0.4082 0.4099 0.4115 0.4131 0.4147 0.4162 0.4177
1.4
0.4192 0.4207 0.4222 0.4236 0.4251 0.4265 0.4279 0.4292 0.4306 0.4319
1.5
0.4332 0.4345 0.4357 0.4370 0.4382 0.4394 0.4406 0.4418 0.4429 0.4441
1.6
0.4452 0.4463 0.4474 0.4484 0.4495 0.4505 0.4515 0.4525 0.4535 0.4545
1.7
0.4554 0.4564 0.4573 0.4582 0.4591 0.4599 0.4608 0.4616 0.4625 0.4633
1.8
0.4641 0.4649 0.4656 0.4664 0.4671 0.4678 0.4686 0.4693 0.4699 0.4706
1.9
0.4713 0.4719 0.4726 0.4732 0.4738 0.4744 0.4750 0.4756 0.4761 0.4767
2.0
0.4772 0.4778 0.4783 0.4788 0.4793 0.4798 0.4803 0.4808 0.4812 0.4817
2.1
0.4821 0.4826 0.4830 0.4834 0.4838 0.4842 0.4846 0.4850 0.4854 0.4857
2.2
0.4861 0.4864 0.4868 0.4871 0.4875 0.4878 0.4881 0.4884 0.4887 0.4890
2.3
0.4893 0.4896 0.4898 0.4901 0.4904 0.4906 0.4909 0.4911 0.4913 0.4916
2.4
0.4918 0.4920 0.4922 0.4925 0.4927 0.4929 0.4931 0.4932 0.4934 0.4936
2.5
0.4938 0.4940 0.4941 0.4943 0.4945 0.4946 0.4948 0.4949 0.4951 0.4952
2.6
0.4953 0.4955 0.4956 0.4957 0.4959 0.4960 0.4961 0.4962 0.4963 0.4964
2.7
0.4965 0.4966 0.4967 0.4968 0.4969 0.4970 0.4971 0.4972 0.4973 0.4974
2.8
0.4974 0.4975 0.4976 0.4977 0.4977 0.4978 0.4979 0.4979 0.4980 0.4981
2.9
0.4981 0.4982 0.4982 0.4983 0.4984 0.4984 0.4985 0.4985 0.4986 0.4986
3.0
0.4987 0.4987 0.4987 0.4988 0.4988 0.4989 0.4989 0.4989 0.4990 0.4990
LESSON 6
EARLIEST AND LATEST TIMES
232
MBA-H2040 Quantitative Techniques
for Managers
LESSON OUTLINE
The concepts of earliest and latest times
The concept of slack
Numerical problems
LEARNING OBJECTIVES
After reading this lesson you should be able to
-
understand the concepts of earliest and latest times
- understand the concept of slack
- calculate the earliest and latest times
-
find out the slacks
-
identify the critical activities
-
carry out numerical problems
INTRODUCTION
A project manager has the responsibility to see that a project is completed by the stipulated date, without delay.
Attention is focused on this aspect in what follows.
Key concepts
Certain key concepts are introduced below.
EARLIEST TIMES OF AN ACTIVITY
We can consider (i) Earliest Start Time of an activity and (ii) Earliest Finish Time of an activity.
Earliest Start Time of an activity is the earliest possible time of starting that activity on
the condition that all the other activities preceding to it were began at the earliest possible
times.
Earliest Finish Time of an activity is the earliest possible time of completing that activity. It is given by
the formula.
The Earliest Finish Time of an activity = The Earliest Start Time of the activity + The estimated
duration to carry out that activity.
LATEST TIMES OF AN ACTIVITY
We can consider (i) Latest Finish Time of an activity and (ii) Latest Start Time of an activity.
Latest Finish Time of an activity is the latest possible time of completing that activity
on the condition that all the other activities succeeding it are carried out as per the plan of the
management and without delaying the project beyond the stipulated time.
Latest Start Time of an activity is the latest possible time of beginning that activity. It is given by the
formula
Latest Start Time of an activity = The Latest Finish Time of the activity - The estimated
duration to carry out that activity.
233
MBA-H2040 Quantitative Techniques
for Managers
TOTAL FLOAT OF AN ACTIVITY
Float seeks to measure how much delay is acceptable. It sets up a control limit for delay.
The total float of an activity is the time by which that activity can be delayed without delaying the
whole project. It is given by the formula
Total Float of an Activity = Latest Finish Time of the activity - Earliest Finish Time of that activity.
It is also given by the formula
Total Float of an Activity = Latest Start Time of the activity - Earliest Start Time of that
activity.
Since a delay in a critical activity will delay the execution of the whole project, the total
float of a critical activity must be zero.
EXPECTED TIMES OF AN EVENT
An event occurs at a point of time. We can consider (i) Earliest Expected Time of Occurrence of an event and
(ii) Latest Allowable Time of Occurrence an event.
The Earliest Expected Time of Occurrence of an event is the earliest possible time of
expecting that event to happen on the condition that all the preceding activities have been
completed.
The Latest Allowable Time of Occurrence of an event is the latest possible time of
expecting that event to happen without delaying the project beyond the stipulated time.
PROCUDURE TO FIND THE EARLIEST EXPECTED TIME OF AN EVENT
Step 1. Take the Earliest Expected Time of Occurrence of the Start Event as zero.
Step 2. For an event other than the Start Event, find out all paths in the network which
connect the Start node with the node representing the event under consideration.
Step 3. In the "Forward Pass" (i.e., movement in the network from left to right), find out
the sum of the time durations of the activities in each path identified in Step 2.
Step 4. The path with the longest time in Step 3 gives the Earliest Expected Time of
Occurrence of the event
Working Rule for finding the earliest expected time of an event:
For an event under consideration, locate all the predecessor events and identify their earliest
expected times. With the earliest expected time of each event, add the time duration of the
activity connecting that event to the event under consideration. The maximum among all these
values gives the Earliest Expected Time of Occurrence of the event.
234
MBA-H2040 Quantitative Techniques
for Managers
PROCUDURE TO FIND THE LATEST ALLOWABLE TIME OF AN EVENT
We consider the "Backward Pass" (i.e., movement in the network from right to left).
The latest allowable time of occurrence of the End Node must be the time of completion
of the project. Therefore it shall be equal to the time of the critical path of the project.
Step 1. Identify the latest allowable time of occurrence of the End Node.
Step 2. For an event other than the End Event, find out all paths in the network which
connect the End node with the node representing the event under consideration.
Step 3. In the "Backward Pass" (i.e., movement in the network from right to left),
subtract the time durations of the activities along each such path.
Step 4. The Latest Allowable Time of Occurrence of the event is determined by the path
with the longest time in Step 3. In other words, the smallest value of time obtained in
Step 3 gives the Latest Allowable Time of Occurrence of the event.
Working Rule for finding the latest allowable time of an event:
For an event under consideration, locate all the successor events and identify their latest
allowable times. From the latest allowable time of each successor event, subtract the time
duration of the activity that begins with the event under consideration. The minimum among
all these values gives the Latest Allowable Time of Occurrence of the event.
SLACK OF AN EVENT
The allowable time gap for the occurrence of an event is known as the slack of that event. It is given by the
formula
Slack of an event = Latest Allowable Time of Occurrence of the event - Earliest Expected
Time of Occurrence of that event.
SLACK OF AN ACTIVITY
The slack of an activity is the float of the activity.
Problem 1:
The following details are available regarding a project:
Activity
Predecessor
Duration (Weeks)
Activity
A
-
12
235
MBA-H2040 Quantitative Techniques
for Managers
B
A
7
C
A
11
D
A
8
E
A
6
F
B
10
G
C
9
H
D, F
14
I
E, G
13
J
H, I
16
Determine the earliest and latest times, the total float for each activity, the critical activities and the project
completion time.
Solution:
With the given data, we construct the following network diagram for the project.
3 F
10
B 7 H
D 14
5
A 8 J
12 E 16
1
2
7
8
6 I
C 11 13
G
6
9
Consider the paths, beginning
4 with the start node and stopping with the end node. There are
four such paths for the given project. They are as follows:
Path I
A B F H J
1
2
3
5
7
1
12 7 10 14 16
Time of the path = 12 + 7 + 10 + 14 + 16 = 59 weeks.
Path II
A D H J
1
2
5
7
8
236
MBA-H2040 Quantitative Techniques
for Managers
12 8 14 16
Time of the path = 12 + 8 + 14 + 16 = 50 weeks.
Path III
A E I J
1
2
6
7
8
12 6 13 16
Time of the path = 12 + 6 + 13 + 16 = 47 weeks.
Path IV
A C G I J
1
2
4
6
7
8
12 11 9 13 16
Time of the path = 12 + 11 + 9 + 13 + 16 = 61 weeks.
Compare the times for the four paths. Maximum of {51, 50, 47, 61} = 61. We see that the maximum time of a
path is 61 weeks.
Forward pass:
Calculation of Earliest Expected Time of Occurrence of Events
Node Earliest Time of Occurrence of Node
1
0
2
Time for Node 1 + Time for Activity A = 0 + 12 = 12
3
Time for Node 2 + Time for Activity B = 12 + 7 = 19
4
Time for Node 2 + Time for Activity C = 12 + 11 = 23
5
Max {Time for Node 2 + Time for Activity D,
Time for Node 3 + Time for Activity F}
= Max {12 + 8, 19 + 10} = Max {20, 29} = 29
6
Max {Time for Node 2 + Time for Activity E,
Time for Node 4 + Time for Activity G}
= Max {12 + 6, 23 + 9} = Max {18, 32} = 32
7
Max {Time for Node 5 + Time for Activity H,
Time for Node 6 + Time for Activity I}
= Max {29 + 14, 32 + 13} = Max {43, 45} = 45
8
Time for Node 7 + Time for Activity J = 45 + 16 = 61
Using the above values, we obtain the Earliest Start Times of the activities as follows:
Activity
Earliest Start Time
(Weeks)
237
MBA-H2040 Quantitative Techniques
for Managers
A
0
B
12
C
12
D
12
E
12
F
19
G
23
H
29
I
32
J
45
Backward pass:
Calculation of Latest Allowable Time of Occurrence of Events
Node Latest Allowable Time of Occurrence of Node
8
Maximum time of a path in the network = 61
7
Time for Node 8 - Time for Activity J = 61 -16 = 45
6
Time for Node 7 - Time for Activity I = 45 -13 = 32
5
Time for Node 7 - Time for Activity H = 45 -14 = 31
4
Time for Node 6 - Time for Activity G = 32 - 9 = 23
3
Time for Node 5 - Time for Activity F = 31- 10 = 21
2
Min {Time for Node 3 - Time for Activity B,
Time for Node 4 - Time for Activity C,
Time for Node 5 - Time for Activity D,
Time for Node 6 - Time for Activity E}
= Min {21 - 7, 23 - 11, 31 - 8, 32 - 6}
= Min {14, 12, 23, 26} = 12
1
Time for Node 2 - Time for Activity A = 12- 12 = 0
Using the above values, we obtain the Latest Finish Times of the activities as follows:
Activity
Latest Finish Time
(Weeks)
J
61
I
45
H
45
G
32
F
31
238
MBA-H2040 Quantitative Techniques
for Managers
E
32
D
31
C
23
B
21
A
12
Calculation of Total Float for each activity:
Activity
Duration
Earliest Start Earliest
Latest
Latest
Total Float = Latest
(Weeks)
Time
Finish Time
Start Time Finish
Finish Time - Earliest
Time
Finish Time
A
12
0
12
0
12
0
B
7
12
19
14
21
2
C
11
12
23
12
23
0
D
8
12
20
23
31
11
E
6
12
18
26
32
14
F
10
19
29
21
31
2
G
9
23
32
23
32
0
H
14
29
43
31
45
2
I
13
32
45
32
45
0
J
16
45
61
45
61
0
The activities with total float = 0 are A, C, G, I and J. They are the critical activities.
Project completion time = 61 weeks.
Problem 2:
The following are the details of the activities in a project:
Activity
Predecessor
Duration (Weeks)
Activity
A
-
15
B
A
17
C
A
21
D
B
19
E
B
22
239
MBA-H2040 Quantitative Techniques
for Managers
F
C, D
18
G
E, F
15
Calculate the earliest and latest times, the total float for each activity and the project completion time.
Solution:
The following network diagram is obtained for the given project.
E
3
22
B
A 17 D 19 G
1 15
2 15
5
6
C 21 F 18
4
Consider the paths, beginning with the start node and stopping with the end node. There are
three such paths for the given project. They are as follows:
Path I
A B E G
1
2
3
5
6
15 17 22 15
Time of the path = 15 + 17 + 22 + 15 = 69 weeks.
Path II
A B D F G
1
2
3
4
5
6
15 17 19 18 15
Time of the path = 15 + 17 + 19 + 18 + 15 = 84 weeks.
Path III
A C F G
1
2
4
5
6
15 21 18 15
Time of the path = 15 + 21 + 18 + 15 = 69 weeks.
Compare the times for the three paths. Maximum of {69, 84, 69} = 84. We see that the maximum time of a path
is 84 weeks.
240
MBA-H2040 Quantitative Techniques
for Managers
Forward pass:
Calculation of Earliest Time of Occurrence of Events
Node
Earliest Time of Occurrence of Node
1
0
2
Time for Node 1 + Time for Activity A = 0 + 15 = 15
3
Time for Node 2 + Time for Activity B = 15 + 17 = 32
4
Max {Time for Node 2 + Time for Activity C,
Time for Node 3 + Time for Activity D}
= Max {15 + 21, 32 + 19} = Max {36, 51} = 51
5
Max {Time for Node 3 + Time for Activity E,
Time for Node 4 + Time for Activity F}
= Max {32 + 22, 51 + 18} = Max {54, 69} = 69
6
Time for Node 5 + Time for Activity G = 69 + 15 = 84
Calculation of Earliest Time for Activities
Activity
Earliest Start Time
(Weeks)
A
0
B
15
C
15
D
32
E
32
F
51
G
69
Backward pass:
Calculation of the Latest Allowable Time of Occurrence of Events
Node
Latest Allowable Time of Occurrence of Node
6
Maximum time of a path in the network = 84
5
Time for Node 6 - Time for Activity G = 84 -15 = 69
4
Time for Node 5 - Time for Activity F = 69 - 18 = 51
3
Min {Time for Node 4 - Time for Activity D,
Time for Node 5 - Time for Activity E}
= Min {51 - 19, 69 - 22} = Min {32, 47} = 32
2
Min {Time for Node 3 - Time for Activity B,
Time for Node 4 - Time for Activity C}
= Min {32 - 17, 51 - 21} = Min {15, 30} = 15
1
Time for Node 2 - Time for Activity A = 15 - 15 = 0
Calculation of the Latest Finish Times of the activities
Activity
Latest Finish Time (Weeks)
G
84
241
MBA-H2040 Quantitative Techniques
for Managers
F
69
E
69
D
51
C
51
B
32
A
15
Calculation of Total Float for each activity:
Activity
Duration
Earliest Start Earliest
Latest
Latest
Total Float = Latest
(Weeks)
Time
Finish Time
Start Time Finish
Finish Time - Earliest
Time
Finish Time
A
15
0
15
0
15
0
B
17
15
32
15
32
0
C
21
15
36
30
51
15
D
19
32
51
32
51
0
E
22
32
54
47
69
15
F
18
51
69
51
69
0
G
15
69
84
69
84
0
The activities with total float = 0 are A, B, D, F and G. They are the critical activities.
Project completion time = 84 weeks.
Problem 3:
Consider a project with the following details:
Name of
Predecessor
Duration
Activity
Activity
(Weeks)
A
-
8
B
A
13
C
A
9
D
A
12
E
B
14
F
B
8
G
D
7
H
C, F, G
12
I
C, F, G
9
J
E, H
10
K
I, J
7
Determine the earliest and latest times, the total float for each activity, the critical activities, the slacks of the
events and the project completion time.
242
MBA-H2040 Quantitative Techniques
for Managers
Solution:
The following network diagram is got for the given project:
E
3
6
14
B 13 H 12
8 J 10
F
A C I K
9 9 7
8
1
2
5
7
8
D 12 G 7
4
Path I
A B E J K
1
2
3
6
7
8
8 13 14 10 7
Time of the path = 8 + 13 + 14 + 10 + 7 = 52 weeks.
Path II
A B F H J K
1
2
3
5
6
7
8
8 13 8 12 10 7
Time of the path = 8 + 13 + 8 + 12 + 10 + 7 = 58 weeks.
Path III
A B F I K
1
2
3
5
7
8
8 13 8 9 7
Time of the path = 8 + 13 + 8 + 9 + 7 = 45 weeks.
Path IV
A C H J K
1
2
5
6
7
8
8 9 12 10
5 7
Time of the path = 8 + 9 + 12 + 10 + 7 = 46 weeks.
243
MBA-H2040 Quantitative Techniques
for Managers
Path V
A C I K
1
2
5
7
8
8 9 9 7
5
6
Time of the path = 8 + 9 + 9 + 7 = 33 weeks.
Path VI
A D G H J K
1
2
4
5
6
7
8
8 12 7 12 10 7
Time of the path = 8 + 12 + 7 + 12 + 10 + 7 = 56 weeks.
Path VII
A D G I K
1
2
4
5
7
8
8 12 7 9 7
Time of the path = 8 + 12 + 7 + 9 + 7 = 43 weeks.
Compare the times for the three paths. Maximum of {52, 58, 45, 46, 33, 56, 43} = 58.
We see that the maximum time of a path is 58 weeks.
244
MBA-H2040 Quantitative Techniques
for Managers
Forward pass:
Calculation of Earliest Time of Occurrence of Events
Node Earliest Time of Occurrence of Node
1
0
2
Time for Node 1 + Time for Activity A = 0 + 8 = 8
3
Time for Node 2 + Time for Activity B = 8 + 13 = 21
4
Time for Node 2 + Time for Activity D = 8+ 12 = 20
5
Max {Time for Node 2 + Time for Activity C,
Time for Node 3 + Time for Activity F,
Time for Node 4 + Time for Activity G}
= Max { 8 + 9, 21 + 8 , 20 + 7 } = Max {17, 29, 27} = 29
6
Max {Time for Node 3 + Time for Activity E,
Time for Node 5 + Time for Activity H}
= Max {21 + 14 , 29 + 12} = Max {35, 41} = 41
7
Max {Time for Node 5 + Time for Activity I,
Time for Node 6 + Time for Activity J}
= Max {29 + 9, 41 + 10} = Max {38, 51} = 51
8
Time for Node 7 + Time for Activity J = 51+ 7 = 58
Earliest Start Times of the activities
Activity
Earliest Start Time
(Weeks)
A
0
B
8
C
8
D
8
E
21
F
21
G
20
H
29
I
29
J
41
K
51
Backward pass:
Calculation of Latest Allowable Time of Occurrence of Events
Node Latest Allowable Time of Occurrence of Node
8
Maximum time of a path in the network = 58
7
Time for Node 8 - Time for Activity K = 58 -7 = 51
6
Time for Node 7 - Time for Activity J = 51 -10 = 41
245
MBA-H2040 Quantitative Techniques
for Managers
5
Min {Time for Node 6 - Time for Activity H,
Time for Node 7 - Time for Activity I}
= Min {41 - 12, 51 - 9} = Min {29, 42} = 29
4
Time for Node 5 - Time for Activity G = 29 - 7 = 22
3
Min {Time for Node 5 - Time for Activity F,
Time for Node 6 - Time for Activity E}
= Min {29 - 8, 41 - 14} = Min {21, 27} = 21
2
Min {Time for Node 3 - Time for Activity B,
Time for Node 4 - Time for Activity D,
Time for Node 5 - Time for Activity C}
= Min {21 - 13, 22 - 12, 29 - 9}
= Min {8, 10, 20} = 8
1
Time for Node 2 - Time for Activity A = 8 - 8 = 0
Latest Finish Times of the activities
Activity
Latest Finish Time
(Weeks)
K
58
J
51
I
51
H
41
G
29
F
29
E
41
D
22
C
29
B
21
A
8
Calculation of Total Float for each activity:
Activity
Duration
Earliest Start Earliest
Latest
Latest
Total Float = Latest
(Weeks)
Time
Finish Time
Start Time Finish
Finish Time - Earliest
Time
Finish Time
A
8
0
8
0
8
0
B
13
8
21
8
21
0
C
9
8
17
20
29
12
D
12
8
20
10
22
2
E
14
21
35
27
41
6
F
8
21
29
21
29
0
246
MBA-H2040 Quantitative Techniques
for Managers
G
7
20
27
22
29
2
H
12
29
41
29
41
0
I
9
29
38
42
51
13
J
10
41
51
41
51
0
K
7
51
58
51
58
0
The activities with total float = 0 are A, B, F, H, J and K. They are the critical activities.
Project completion time = 58 weeks.
Calculation of slacks of the events
Slack of an event = Latest Allowable Time of Occurrence of the event - Earliest Expected
Time of Occurrence of that event.
Event
Earliest Expected Time
Latest Allowable Time
Slack of the
(Node)
of Occurrence of Event
of Occurrence of Event
Event
1
0
0
0
2
8
8
0
3
21
21
0
4
20
22
2
5
29
29
0
6
41
41
0
7
51
51
0
8
58
58
0
Interpretation:
On the basis of the slacks of the events, it is concluded that the occurrence of event 4 may be delayed upto a
maximum period of 2 weeks while no other event cannot be delayed.
QUESTIONS
1. Explain the terms: The earliest and latest times of the activities of a project.
2. Explain the procedure to find the earliest expected time of an event.
3. Explain the procedure to find the latest allowable time of an event.
4. What is meant by the slack of an activity? How will you determine it?
5. Consider the project with the following details:
activity
Duration (weeks)
12
1
23
3
247
MBA-H2040 Quantitative Techniques
for Managers
24
7
34
5
35
8
45
4
56
1
Determine the earliest and the latest times of the activities. Calculate the total float for
each activity and the slacks of the events.
248
MBA-H2040 Quantitative Techniques
for Managers
LESSON 7
CRASHING OF A PROJECT
LESSON OUTLINE
The idea of crashing of a project
The criterion of selection of an activity for crashing
Numerical problems
LEARNING OBJECTIVES
After reading this lesson you should be able to
- understand the concept of crashing of a project
- choose an activity for crashing
- work out numerical problems
THE MEANING OF CRASHING:
The process of shortening the time to complete a project is called crashing and is usually achieved by putting
into service additional labour or machines to one activity or more activities. Crashing involves more costs. A
project manager would like to speed up a project by spending as minimum extra cost as possible. Project
crashing seeks to minimize the extra cost for completion of a project before the stipulated time.
STEPS IN PROJECT CRASHING:
Assumption: It is assumed that there is a linear relationship between time and cost.
Let us consider project crashing by the critical path method. The following four-step
procedure is adopted.
Step 1: Find the critical path with the normal times and normal costs for the activities and
identify the critical activities.
Step 2: Find out the crash cost per unit time for each activity in the network. This is
calculated by means of the following formula.
Crash cos t
Crash cos t Normal cos t
Time period
Normal time Crash time
249
MBA-H2040 Quantitative Techniques
for Managers
Activity Cost
Crash time & Cost
Crash Cost
Normal Cost Normal time & Cost
Activity Time
Crash Time Normal Time
Step 3: Select an activity for crashing. The criteria for the selection is as follows:
Select the activity on the critical path with the smallest crash cost per unit time. Crash this
activity to the maximum units of time as may be permissible by the given data.
Crashing an activity requires extra amount to be spent. However, even if the company
is prepared to spend extra money, the activity time cannot be reduced beyond a certain limit in
view of several other factors.
In step 1, we have to note that reducing the time of on activity along the critical path
alone will reduce the completion time of a project. Because of this reason, we select an
activity along the critical path for crashing.
In step 3, we have to consider the following question:
If we want to reduce the project completion time by one unit, which critical activity
will involve the least additional cost?
On the basis of the least additional cost, a critical activity is chosen for crashing. If
there is a tie between two critical activities, the tie can be resolved arbitrarily.
Step 4: After crashing an activity, find out which is the critical path with the changed
conditions. Sometimes, a reduction in the time of an activity in the critical path may cause a
non-critical path to become critical. If the critical path with which we started is still the
longest path, then go to Step 3. Otherwise, determine the new critical path and then go to
Step 3.
Problem 1: A project has activities with the following normal and crash times and cost:
250
MBA-H2040 Quantitative Techniques
for Managers
Activity
Predecessor
Normal Time
Crash Time
Normal Cost Crash Cost
Activity
(Weeks)
(Weeks) (Rs.)
(Rs.)
A
-
4
3
8,000
9,000
B
A
5
3
16,000
20,000
C
A
4
3
12,000
13,000
D
B
6
5
34,000
35,000
E
C
6
4
42,000
44,000
F
D
5
4
16,000
16,500
G
E
7
4
66,000
72,000
H
G
4
3
2,000
5,000
Determine a crashing scheme for the above project so that the total project time is reduced by 3 weeks.
Solution:
We have the following network diagram for the given project with normal costs:
E G
C 6
3 7
5 H
7
A 4
4 4 4 3
4
2
1
5
B
8
D F
5
5
6
4
6
4
Beginning from the Start Node and terminating with the End Node, there are two paths for the network as
detailed below:
Path I:
A B D F
1
2
5
7
8
4 5 6 5
The time for the path = 4 + 5 + 6 + 5 = 20 weeks.
Path II:
A C E G H
1
2
3
5
7
8
4 4 6 7 4
The time for the path = 4 + 4 + 6 + 7 + 4 = 25 weeks.
Maximum of {20, 25} = 25.
251
MBA-H2040 Quantitative Techniques
for Managers
Therefore Path II is the critical path and the critical activities are A, C, E, G and H. The non-critical activities are
B, D and F.
Given that the normal time of activity A is 4 weeks while its crash time is 3 weeks. Hence the time of
this activity can be reduced by one week if the management is prepared to spend an additional amount. However,
the time cannot be reduced by more than one week even if the management may be prepared to spend more
money. The normal cost of this activity is Rs. 8,000 whereas the crash cost is Rs. 9,000. From this, we see that
crashing of activity A by one week will cost the management an extra amount of Rs. 1,000. In a similar fashion,
we can work out the crash cost per unit time for the other activities also. The results are provided in the
following table.
Activity
Normal
Crash
Normal
Crash Cost
Crash cost
Normal
Crash Cost
Time
Time Cost
-
Time -
per unit
Normal
Crash
time
Cost
Time
A
4
3
8,000
9,000
1,000
1
1,000
B
5
3
16,000
20,000
4,000
2
2,000
C
4
3
12,000
13,000
1,000
1
1,000
D
6
5
34,000
35,000
1,000
1
1,000
E
6
4
42,000
44,000
2,000
2
1,000
F
5
4
16,000
16,500
500
1
500
G
7
4
66,000
72,000
6,000
1
6,000
H
4
3
2,000
5,000
3,000
1
3,000
A non-critical activity can be delayed without delaying the execution of the whole project. But, if a
critical activity is delayed, it will delay the whole project. Because of this reason, we have to select a critical
activity for crashing. Here we have to choose one of the activities A, C, E, G and H The crash cost per unit time
works out as follows:
Rs. 1,000 for A; Rs. 1,000 for C; Rs. 1,000 for E; Rs. 6,000 for G; Rs. 3,000 for H.
The maximum among them is Rs. 1,000. So we have to choose an activity with Rs.1,000 as the crash cost
per unit time. However, there is a tie among A, C and E. The tie can be resolved arbitrarily. Let us select A for
crashing. We reduce the time of A by one week by spending an extra amount of Rs. 1,000.
After this step, we have the following network with the revised times for the activities:
E G
C 6
3 7
5 H
7
A 4
4 4 4 3
3
2
1
5
B
8
D F
5
5
6
4
6
4
The revised time for Path I = 3 + 5 + 6 + 5 = 19 weeks.
The time for Path II = 3 + 4 + 6 + 7 + 4 = 24 weeks.
Maximum of {19, 24} = 24.
252
MBA-H2040 Quantitative Techniques
for Managers
Therefore Path II is the critical path and the critical activities are A, C, E, G and H. However, the time for A
cannot be reduced further. Therefore, we have to consider C, E, G and H for crashing. Among them, C and E
have the least crash cost per unit time. The tie between C and E can be resolved arbitrarily. Suppose we reduce
the time of C by one week with an extra cost of Rs. 1,000.
After this step, we have the following network with the revised times for the activities:
253
MBA-H2040 Quantitative Techniques
for Managers
E G
C 6
3 7
5 H
7
A 3
4 4 4 3
3
2
1
5
B
8
D F
5
5
6
4
6
4
The time for Path I = 3 + 5 + 6 + 5 = 19 weeks.
The time for Path II = 3 + 3 + 6 + 7 + 4 = 23 weeks.
Maximum of {19, 23} = 23.
Therefore Path II is the critical path and the critical activities are A, C, E, G and H. Now the time for A or C
cannot be reduced further. Therefore, we have to consider E, G and H for crashing. Among them, E has the least
crash cost per unit time. Hence we reduce the time of E by one week with an extra cost of Rs. 1,000.
By the given condition, we have to reduce the project time by 3 weeks. Since this has been accomplished, we
stop with this step.
Result: We have arrived at the following crashing scheme for the given project:
Reduce the time of A, C and E by one week each.
Project time after crashing is 22 weeks.
Extra amount required = 1,000 + 1,000 + 1,000 = Rs. 3,000.
Problem 2:
The management of a company is interested in crashing of the following project by spending an additional
amount not exceeding Rs. 2,000. Suggest how this can be accomplished.
Activity
Predecessor
Normal Time
Crash Time
Normal Cost Crash Cost
Activity
(Weeks)
(Weeks) (Rs.)
(Rs.)
A
-
7
6
15,000
18,000
B
A
12
9
11,000
14,000
C
A
22
21
18,500
19,000
D
B
11
10
8,000
9,000
E
C, D
6
5
4,000
4,500
Solution:
We have the following network diagram for the given project with normal costs:
3
B D 11
12
A C E
1
2
4
5
7 22 6
254
MBA-H2040 Quantitative Techniques
for Managers
There are two paths for this project as detailed below:
Path I:
A B D E
1 7 2 12 11
3 6 4
5
The time for the path = 7 + 12 + 11 + 6 = 36 weeks.
Path II:
A C E
1
4
2
5
7 22 6
The time for the path = 7 + 22 + 6 = 35 weeks.
Maximum of {36, 35} = 36.
Therefore Path I is the critical path and the critical activities are A, B, D and E. The non-critical activity is C.
The crash cost per unit time for the activities in the project are provided in the following table.
255
MBA-H2040 Quantitative Techniques
for Managers
Activity
Normal
Crash
Normal
Crash Cost
Crash cost
Normal
Crash Cost
Time
Time Cost
-
Time -
per unit
Normal
Crash
time
Cost
Time
A
7
6
15,000
18,000
3,000
1
3,000
B
12
9
11,000
14,000
3,000
3
1,000
C
22
21
18,500
19,000
500
1
500
D
11
10
8,000
9,000
1,000
1
1,000
E
6
5
4,000
4,500
500
1
500
We have to choose one of the activities A, B, D and E for crashing. The crash cost per unit time is as
follows:
Rs. 3,000 for A; Rs. 1,000 for B; Rs. 1,000 for D; Rs. 500 for E.
The least among them is Rs. 500. So we have to choose the activity E for crashing. We reduce the time of E by
one week by spending an extra amount of Rs. 500.
After this step, we have the following network with the revised times for the activities:
B D
3 11
12
A C E
1
2
4
5
7 22 5
The revised time for Path I = 7 + 12 + 11 + 5 = 35 weeks.
The time for Path II = 7 + 22 + 5 = 34 weeks.
Maximum of {35, 34} = 35.
Therefore Path I is the critical path and the critical activities are A, B, D and E. The non-critical activity is C.
The time of E cannot be reduced further. So we cannot select it for crashing. Next B and have the smallest crash
cost per unit time. Let us select B for crashing. Let us reduce the time of E by one week at an extra cost of Rs.
1,000.
After this step, we have the following network with the revised times for the activities:
3
B D 11
11
A C E
1
2
4
5
7 22 5
The revised time for Path I = 7 + 11 + 11 + 5 = 34 weeks.
256
MBA-H2040 Quantitative Techniques
for Managers
The time for Path II = 7 + 22 + 5 = 34 weeks.
Maximum of {34, 34} = 34.
Since both paths have equal times, both are critical paths. So, we can choose an activity for crashing from either
of them depending on the least crash cost per unit time. In path I, the activities are A, B, D and E. In path II, the
activities are A, C and E.
The crash cost per unit time is the least for activity C. So we select C for crashing. Reduce the time of C by one
week at an extra cost of Rs. 500.
By the given condition, the extra amount cannot exceed Rs. 2,000. Since this state has been met, we stop with
this step.
Result: The following crashing scheme is suggested for the given project:
Reduce the time of E, B and C by one week each.
Project time after crashing is 33 weeks.
Extra amount required = 500 + 1,000 + 500 = Rs. 2,000.
257
MBA-H2040 Quantitative Techniques
for Managers
Problem 3:
The manager of a company wants to apply crashing for the following project by spending an additional amount
not exceeding Rs. 2,000. Offer your suggestion to the manager.
Activity
Predecessor
Normal Time
Crash Time
Normal Cost Crash Cost
Activity
(Weeks)
(Weeks) (Rs.)
(Rs.)
A
-
20
19
8,000
10,000
B
-
15
14
16,000
19,000
C
A
22
20
13,000
14,000
D
A
17
15
7,500
9,000
E
B
19
18
4,000
5,000
F
C
28
27
3,000
4,000
G
D, E
25
24
12,000
13,000
Solution:
We have the following network diagram for the given project with normal costs:
C
3
5
A 22 28
20 F
17
1
D
15
7
B E G 25
19
2
4
There are three paths for this project as detailed below:
Path I:
A C F
2
4
1 20 22 28
6
The time for the path = 20 + 22 + 28= 70 weeks.
258
MBA-H2040 Quantitative Techniques
for Managers
Path II:
A D G
1
2
5
6
20 17 25
The time for the path = 20 + 17 + 25= 62 weeks.
Path III:
B E G
1
3
5
6
15 19 25
The time for the path = 15+19 +25 = 69 weeks.
Maximum of {70, 62, 69} = 70.
Therefore Path I is the critical path and the critical activities are A, C and F. The non-critical activities are B, D,
E and G.
The crash cost per unit time for the activities in the project are provided in the following table
Activity
Normal
Crash
Normal
Crash Cost
Crash cost
Normal
Crash Cost
Time
Time Cost
-
Time -
per unit
Normal
Crash
time
Cost
Time
A
20
19
8,000
10,000
2,000
1
2,000
B
15
14
16,000
19,000
3,000
1
3,000
C
22
20
13,000
14,000
1,000
2
500
D
17
15
7,500
9,000
1,500
2
750
E
19
18
4,000
5,000
1,000
1
1,000
F
28
27
3,000
4,000
1,000
1
1,000
G
25
24
12,000
13,000
1,000
1
1,000
We have to choose one of the activities A, C and F for crashing. The crash cost per unit time is as
follows:
Rs. 2,000 for A; Rs. 500 for C; Rs. 1,000 for F.
The least among them is Rs. 500. So we have to choose the activity C for crashing. We reduce the time of C by
one week by spending an extra amount of Rs. 500.
After this step, we have the following network with the revised times for the activities:
C
3
5
A 21 28
20 F
17
1
D
15
7
B E G 25
19
2
4
259
MBA-H2040 Quantitative Techniques
for Managers
The revised time for Path I = 20 + 21 + 28= 69 weeks.
The time for Path II = 20 + 17 + 25= 62 weeks.
The time for Path III = 15+19 +25 = 69 weeks.
Maximum of {69, 62, 69} = 69.
Since paths I and III have equal times, both are critical paths. So, we can choose an activity for crashing from
either of them depending on the least crash cost per unit time.
In path I, the activities are A, C and F. In path III, the activities are B, E and G.
The crash cost per unit time is the least for activity C. So we select C for crashing. Reduce the time of C by one
week at an extra cost of Rs. 500.
After this step, we have the following network with the revised times for the activities:
C
3
5
A 20 28
20 F
17
1
D
15
7
B E G 25
19
2
4
The revised time for Path I = 20 + 20 + 28= 68 weeks.
The time for Path II = 20 + 17 + 25= 62 weeks.
The time for Path III = 15+19 +25 = 69 weeks.
Maximum of {68, 62, 69} = 69.
Therefore path III is the critical activities. Hence we have to select an activity from Path III for crashing. We see
that the crash cost per unit time is as follows:
Rs. 3,000 for B; Rs. 1,000 for E; Rs. 1,000 for G.
The least among them is Rs. 1,000. So we can select either E or G for crashing. Let us select E for crashing. We
reduce the time of E by one week by spending an extra amount of Rs. 1,000.
By the given condition, the extra amount cannot exceed Rs. 2,000. Since this condition has been reached, we
stop with this step.
Result: The following crashing scheme is suggested for the given project:
Reduce the time of C by 2 weeks and that of E by one week.
Project time after crashing is 67 weeks.
Extra amount required = 2 x 500 + 1,000 = Rs. 2,000.
QUESTIONS
1. Explain the concept of crashing of a project.
2. Explain the criterion for selection of an activity for crashing.
260
MBA-H2040 Quantitative Techniques
for Managers
261
MBA-H2040 Quantitative Techniques
for Managers
UNIT IV
GAME THEORY & GOAL PROGRAMMING
LESSON 1 BASIC CONCEPTS IN GAME THEORY
LESSON OUTLINE
Introduction to the theory of games
The definition of a game
Competitive game
Managerial applications of the theory of games
Key concepts in the theory of games
Types of games
LEARNING OBJECTIVES
After reading this lesson you should be able to
- understand the concept of a game
- grasp the assumptions in the theory of games
- appreciate the managerial applications of the theory of games
- understand the key concepts in the theory of games
- distinguish between different types of games
Introduction to game theory
Game theory seeks to analyse competing situations which arise out of conflicts of interest.
Abraham Maslow's hierarchical model of human needs lays emphasis on fulfilling the basic
needs such as food, water, clothes, shelter, air, safety and security. There is conflict of interest
between animals and plants in the consumption of natural resources. Animals compete among
themselves for securing food. Man competes with animals to earn his food. A man also
competes with another man. In the past, nations waged wars to expand the territory of their
rule. In the present day world, business organizations compete with each other in getting the
market share. The conflicts of interests of human beings are not confined to the basic needs
alone. Again considering Abraham Maslow's model of human needs, one can realize that
conflicts also arise due to the higher levels of human needs such as love, affection, affiliation,
recognition, status, dominance, power, esteem, ego, self-respect, etc. Sometimes one
witnesses clashes of ideas of intellectuals also. Every intelligent and rational participant in a
conflict wants to be a winner but not all participants can be the winners at a time. The
situations of conflict gave birth to Darwin's theory of the `survival of the fittest'. Nowadays
262
MBA-H2040 Quantitative Techniques
for Managers
the concepts of conciliation, co-existence, co-operation, coalition and consensus are gaining
ground. Game theory is another tool to examine situations of conflict so as to identify the
courses of action to be followed and to take appropriate decisions in the long run. Thus this
theory assumes importance from managerial perspectives. The pioneering work on the theory
of games was done by von Neumann and Morgenstern through their publication entitled `The
Theory of Games and Economic Behaviour' and subsequently the subject was developed by
several experts. This theory can offer valuable guidelines to a manager in `strategic
management' which can be used in the decision making process for merger, take-over, joint
venture, etc. The results obtained by the application of this theory can serve as an early
warning to the top level management in meeting the threats from the competing business
organizations and for the conversion of the internal weaknesses and external threats into
opportunities and strengths, thereby achieving the goal of maximization of profits. While this
theory does not describe any procedure to play a game, it will enable a participant to select the
appropriate strategies to be followed in the pursuit of his goals. The situation of failure in a
game would activate a participant in the analysis of the relevance of the existing strategies and
lead him to identify better, novel strategies for the future occasions.
Definitions of game theory
There are several definitions of game theory. A few standard definitions are presented below.
In the perception of Robert Mockler, "Game theory is a mathematical technique
helpful in making decisions in situations of conflicts, where the success of one part depends at
the expense of others, and where the individual decision maker is not in complete control of
the factors influencing the outcome".
The definition given by William G. Nelson runs as follows: "Game theory, more
properly the theory of games of strategy, is a mathematical method of analyzing a conflict.
The alternative is not between this decision or that decision, but between this strategy or that
strategy to be used against the conflicting interest".
In the opinion of Matrin Shubik, "Game theory is a method of the study of decision
making in situation of conflict. It deals with human processes in which the individual
decision-unit is not in complete control of other decision-units entering into the environment".
According to von Neumann and Morgenstern, "The `Game' is simply the totality of
the rules which describe it. Every particular instance at which the game is played ? in a
particular way ? from beginning to end is a `play'. The game consists of a sequence of
moves, and the play of a sequence of choices".
263
MBA-H2040 Quantitative Techniques
for Managers
J.C.C McKinsey points out a valid distinction between two words, namely `game' and
`play'. According to him, "game refers to a particular realization of the rules".
In the words of O.T. Bartos, "The theory of games can be used for `prescribing' how
an intelligent person should go about resolving social conflicts, ranging all the way from open
warfare between nations to disagreements between husband and wife".
Martin K Starr gave the following definition: "Management models in the competitive
sphere are usually termed game models. By studying game theory, we can obtain substantial
information into management's role under competitive conditions, even though much of the
game theory is neither directly operational nor implementable".
According to Edwin Mansfield, "A game is a competitive situation where two or more
persons pursue their own interests and no person can dictate the outcome. Each player, an
entity with the same interests, make his own decisions. A player can be an individual or a
group".
Assumptions for a Competitive Game
Game theory helps in finding out the best course of action for a firm in view of the
anticipated countermoves from the competing organizations. A competitive situation is a
competitive game if the following properties hold:
1. The number of competitors is finite, say N.
2. A finite set of possible courses of action is available to each of the N competitors.
3. A play of the game results when each competitor selects a course of action from the
set of courses available to him. In game theory we make an important assumption
that al the players select their courses of action simultaneously. As a result, no
competitor will be in a position to know the choices of his competitors.
4. The outcome of a play consists of the particular courses of action chosen by the
individual players. Each outcome leads to a set of payments, one to each player,
which may be either positive, or negative, or zero.
Managerial Applications of the Theory of Games
The techniques of game theory can be effectively applied to various managerial
problems as detailed below:
1) Analysis of the market strategies of a business organization in the long run.
2) Evaluation of the responses of the consumers to a new product.
3) Resolving the conflict between two groups in a business organization.
4) Decision making on the techniques to increase market share.
264
MBA-H2040 Quantitative Techniques
for Managers
5) Material procurement process.
6) Decision making for transportation problem.
7) Evaluation of the distribution system.
8) Evaluation of the location of the facilities.
9) Examination of new business ventures and
10) Competitive economic environment.
Key concepts in the Theory of Games
Several of the key concepts used in the theory of games are described below:
Players:
The competitors or decision makers in a game are called the players of the game.
Strategies:
The alternative courses of action available to a player are referred to as his strategies.
Pay off:
The outcome of playing a game is called the pay off to the concerned player.
Optimal Strategy:
A strategy by which a player can achieve the best pay off is called the optimal strategy for
him.
Zero-sum game:
A game in which the total payoffs to all the players at the end of the game is zero is referred
to as a zero-sum game.
Non-zero sum game:
Games with "less than complete conflict of interest" are called non-zero sum games. The
problems faced by a large number of business organizations come under this category. In
such games, the gain of one player in terms of his success need not be completely at the
expense of the other player.
Payoff matrix:
The tabular display of the payoffs to players under various alternatives is called the payoff
matrix of the game.
Pure strategy:
If the game is such that each player can identify one and only one strategy as the optimal
strategy in each play of the game, then that strategy is referred to as the best strategy for that
player and the game is referred to as a game of pure strategy or a pure game.
265
MBA-H2040 Quantitative Techniques
for Managers
Mixed strategy:
If there is no one specific strategy as the `best strategy' for any player in a game, then the
game is referred to as a game of mixed strategy or a mixed game. In such a game, each player
has to choose different alternative courses of action from time to time.
N-person game:
A game in which N-players take part is called an N-person game.
Maximin-Minimax Principle :
The maximum of the minimum gains is called the maximin value of the game and the
corresponding strategy is called the maximin strategy. Similarly the minimum of the
maximum losses is called the minimax value of the game and the corresponding strategy is
called the minimax strategy. If both the values are equal, then that would guarantee the best
of the worst results.
Negotiable or cooperative game:
If the game is such that the players are taken to cooperate on any or every action which may
increase the payoff of either player, then we call it a negotiable or cooperative game.
Non-negotiable or non-cooperative game:
If the players are not permitted for coalition then we refer to the game as a non-negotiable or
non-cooperative game.
Saddle point:
A saddle point of a game is that place in the payoff matrix where the maximum of the row
minima is equal to the minimum of the column maxima. The payoff at the saddle point is
called the value of the game and the corresponding strategies are called the pure strategies.
Dominance:
One of the strategies of either player may be inferior to at least one of the remaining ones.
The superior strategies are said to dominate the inferior ones.
Types of Games:
There are several classifications of a game. The classification may be based on various
factors such as the number of participants, the gain or loss to each participant, the number of
strategies available to each participant, etc. Some of the important types of games are
enumerated below.
Two person games and n-person games:
In two person games, there are exactly two players and each competitor will have a finite
number of strategies. If the number of players in a game exceeds two, then we refer to the
game as n-person game.
266
MBA-H2040 Quantitative Techniques
for Managers
Zero sum game and non-zero sum game:
If the sum of the payments to all the players in a game is zero for every possible outcome of
the game, then we refer to the game as a zero sum game. If the sum of the payoffs from any
play of the game is either positive or negative but not zero, then the game is called a non-zero
sum game
Games of perfect information and games of imperfect information:
A game of perfect information is the one in which each player can find out the strategy that
would be followed by his opponent. On the other hand, a game of imperfect information is the
one in which no player can know in advance what strategy would be adopted by the
competitor and a player has to proceed in his game with his guess works only.
Games with finite number of moves / players and games with unlimited number of
moves:
A game with a finite number of moves is the one in which the number of moves for each
player is limited before the start of the play. On the other hand, if the game can be continued
over an extended period of time and the number of moves for any player has no restriction,
then we call it a game with unlimited number of moves.
Constant-sum games:
If the sum of the game is not zero but the sum of the payoffs to both players in each case is
constant, then we call it a constant sum game. It is possible to reduce such a game to a zero-
sum game.
2x2 two person game and 2xn and mx2 games:
When the number of players in a game is two and each player has exactly two strategies, the
game is referred to as 2x2 two person game.
A game in which the first player has precisely two strategies and the second player has
three or more strategies is called an 2xn game.
A game in which the first player has three or more strategies and the second player has
exactly two strategies is called an mx2 game.
3x3 and large games:
When the number of players in a game is two and each player has exactly three strategies, we
call it a 3x3 two person game.
Two-person zero sum games are said to be larger if each of the two players has 3 or
more choices.
267
MBA-H2040 Quantitative Techniques
for Managers
The examination of 3x3 and larger games is involves difficulties. For such games, the
technique of linear programming can be used as a method of solution to identify the optimum
strategies for the two players.
Non-constant games :
Consider a game with two players. If the sum of the payoffs to the two players is not constant
in all the plays of the game, then we call it a non-constant game.
Such games are divided into negotiable or cooperative games and non-negotiable or
non-cooperative games.
QUESTIONS
1. Explain the concept of a game.
2. Define a game.
3. State the assumptions for a competitive game.
4. State the managerial applications of the theory of games.
5. Explain the following terms: strategy, pay-off matrix, saddle point, pure strategy and
mixed strategy.
6. Explain the following terms: two person game, two person zero sum game, value of a
game, 2xn game and mx2 game.
268
MBA-H2040 Quantitative Techniques
for Managers
LESSON 2
TWO-PERSON ZERO SUM GAMES
LESSON OUTLINE
The concept of a two-person zero sum game
The assumptions for a two-person zero sum game
Minimax and Maximin principles
LEARNING OBJECTIVES
After reading this lesson you should be able to
- understand the concept of a two-person zero sum game
- have an idea of the assumptions for a two-person zero sum game
- understand Minimax and Maximin principles
- solve a two-person zero sum game
- interpret the results from the payoff matrix of a two-person zero sum game
Definition of two-person zero sum game
A game with only two players, say player A and player B, is called a two-person zero sum
game if the gain of the player A is equal to the loss of the player B, so that the total sum is
zero.
Payoff matrix:
When players select their particular strategies, the payoffs (gains or losses) can be represented
in the form of a payoff matrix.
Since the game is zero sum, the gain of one player is equal to the loss of other and
vice-versa. Suppose A has m strategies and B has n strategies. Consider the following payoff
matrix.
Player B's strategies
B
B
B
1
2
n
A a
a
a
1
11
12
1n
Player A's strategies
A
a
a
a
1
21
22
2n
Am a
a
a
1
m
m2
mn
Player A wishes to gain as large a payoff a as possible while player B will do his best to
ij
reach as small a value a as possible where the gain to player B and loss to player A be (-
ij
a ).
ij
269
MBA-H2040 Quantitative Techniques
for Managers
Assumptions for two-person zero sum game:
For building any model, certain reasonable assumptions are quite necessary. Some
assumptions for building a model of two-person zero sum game are listed below.
a) Each player has available to him a finite number of possible courses of action.
Sometimes the set of courses of action may be the same for each player. Or, certain
courses of action may be available to both players while each player may have certain
specific courses of action which are not available to the other player.
b) Player A attempts to maximize gains to himself. Player B tries to minimize losses to
himself.
c) The decisions of both players are made individually prior to the play with no
communication between them.
d) The decisions are made and announced simultaneously so that neither player has an
advantage resulting from direct knowledge of the other player's decision.
e) Both players know the possible payoffs of themselves and their opponents.
Minimax and Maximin Principles
The selection of an optimal strategy by each player without the knowledge of the competitor's
strategy is the basic problem of playing games.
The objective of game theory is to know how these players must select their respective
strategies, so that they may optimize their payoffs. Such a criterion of decision making is
referred to as minimax-maximin principle. This principle in games of pure strategies leads to
the best possible selection of a strategy for both players.
For example, if player A chooses his ith strategy, then he gains at least the payoff min
a , which is minimum of the ith row elements in the payoff matrix. Since his objective is to
ij
maximize his payoff, he can choose strategy i so as to make his payoff as large as possible.
i.e., a payoff which is not less than max min a .
ij
1im
1 jn
Similarly player B can choose jth column elements so as to make his loss not greater than
min max a .
ij
1 jn 1im
If the maximin value for a player is equal to the minimax value for another player, i.e.
max min a V min max a
ij
ij
1im
1 jn
1 jn
1im
then the game is said to have a saddle point (equilibrium point) and the corresponding
strategies are called optimal strategies. If there are two or more saddle points, they must be
equal.
270
MBA-H2040 Quantitative Techniques
for Managers
The amount of payoff, i.e., V at an equilibrium point is known as the value of the
game.
The optimal strategies can be identified by the players in the long run.
Fair game:
The game is said to be fair if the value of the game V = 0.
Problem 1:
Solve the game with the following pay-off matrix.
Player B
Strategies
I
II
III
IV
V
1
2
5
3
6
7
Player A Strategies 2
4
6
8
1
6
3
8
2
3
5
4
4 15 14
18
12
20
Solution:
First consider the minimum of each row.
Row
Minimum Value
1
-3
2
-1
3
2
4
12
Maximum of {-3, -1, 2, 12} = 12
Next consider the maximum of each column.
Column
Maximum Value
1
15
2
14
3
18
4
12
5
20
Minimum of {15, 14, 18, 12, 20}= 12
We see that the maximum of row minima = the minimum of the column maxima. So the
game has a saddle point. The common value is 12. Therefore the value V of the game = 12.
Interpretation:
In the long run, the following best strategies will be identified by the two players:
271
MBA-H2040 Quantitative Techniques
for Managers
The best strategy for player A is strategy 4.
The best strategy for player B is strategy IV.
The game is favourable to player A.
Problem 2:
Solve the game with the following pay-off matrix
Player Y
Strategies
I
II
III
IV
V
1 9 12
7
14
26
Player X Strategies 2 25 35 20
28 30
3 7
6
8
3
2
4 8
11 13
2
1
Solution:
First consider the minimum of each row.
Row
Minimum Value
1
7
2
20
3
-8
4
-2
Maximum of {7, 20, ?8, -2} = 20
Next consider the maximum of each column.
Column
Maximum Value
1
25
2
35
3
20
4
28
5
30
Minimum of {25, 35, 20, 28, 30}= 20
It is observed that the maximum of row minima and the minimum of the column
maxima are equal. Hence the given the game has a saddle point. The common value is 20.
This indicates that the value V of the game is 20.
Interpretation.
272
MBA-H2040 Quantitative Techniques
for Managers
The best strategy for player X is strategy 2.
The best strategy for player Y is strategy III.
The game is favourable to player A.
Problem 3:
Solve the following game:
Player B
Strategies
I
II
III
IV
1 1 6
8
4
Player A Strategies 2 3 7
2 8
3 5
5 1
0
4 3 4
5
7
Solution
First consider the minimum of each row.
Row
Minimum Value
1
-6
2
-8
3
-5
4
-4
Maximum of {-6, -8, -5, -4} = -4
Next consider the maximum of each column.
Column
Maximum Value
1
5
2
-4
3
8
4
7
Minimum of {5, -4, 8, 7}= - 4
Since the max {row minima} = min {column maxima}, the game under consideration has a
saddle point. The common value is ?4. Hence the value of the game is ?4.
Interpretation.
The best strategy for player A is strategy 4.
The best strategy for player B is strategy II.
273
MBA-H2040 Quantitative Techniques
for Managers
Since the value of the game is negative, it is concluded that the game is favourable to
player B.
QUESTIONS
1. What is meant by a two-person zero sum game? Explain.
2. State the assumptions for a two-person zero sum game.
3. Explain Minimax and Maximin principles.
4. How will you interpret the results from the payoff matrix of a two-person zero sum
game? Explain.
5. What is a fair game? Explain.
6. Solve the game with the following pay-off matrix.
Player B
Strategies
I
II
III
IV
V
1
7
5
2
3
9
Player A Strategies 2 10
8
7
4
5
3
9
12
0
2
1
4 11
2 1
3
4
Answer: Best strategy for A: 2
Best strategy for B: IV
V = 4
The game is favourable to player A
7. Solve the game with the following pay-off matrix.
Player B
Strategies
I
II
III
IV
V
1
2
3
8
7
0
Player A Strategies 2
1
7 5 2
3
3
4
2
3
5
1
4
6
4
5
4
7
Answer: Best strategy for A: 3
Best strategy for B: II
V = -2
The game is favourable to player B
8. Solve the game with the following pay-off matrix.
274
MBA-H2040 Quantitative Techniques
for Managers
Player B
6
4
Player A
7 5
Answer: Best strategy for A: 2
Best strategy for B: II
V = 5
The game is favourable to player A
9. Solve the following game and interpret the result.
Player B
Strategies
I
II
III
IV
1 3
7
1
3
Player A Strategies 2 1
2
3
1
3 0
4
2
6
4 2
1
5
1
Answer: Best strategy for A: 3
Best strategy for B: I
V = 0
The value V = 0 indicates that the game is a fair one.
10. Solve the following game:
Player B
Strategies
I
II
III
1 1
8
2
Player A Strategies 2 3
5
6
3 2
2
1
Answer: Best strategy for A: 2
Best strategy for B: I
V = 3
The game is favourable to player A
11. Solve the game
275
MBA-H2040 Quantitative Techniques
for Managers
Player B
I
II
III
IV
1 4
1
2
0
Player A
2 3
5
9
2
3 2
8
0
11
Answer : V = -1
12. Solve the game
Player Y
I
II
III
IV
V
1 4
0
1
7
1
2 0
3
5 7
5
Player X
3 3
2
3
4
3
4 6
4
1
0
5
5 0
0
6
0
0
Answer : V = 2
13. Solve the game
Player B
I
II
III
IV
V
1 9
3
4
4
2
2 8
6
8
5 12
Player A
3 10 7 19 18 14
4 8
6
8 11
6
5 3 5 16 10
8
Answer : V = 7
276
MBA-H2040 Quantitative Techniques
for Managers
14. Solve the game
Player Y
17 10 12
5
4
8
2
5
6
7
6
9
7
6
9
2
3
1
Player X
10 11 14
8 13
8
20 18 17 10 15 17
12 11 15
9
5 11
Answer : V = 10
15. Solve the game
Player B
12 14
8
7
4
9
2 13
6
7
9
8
13
6
8
6
3
1
Player A
14
9 10
8
9
6
20 18 17 11 14 16
8 12 16
9
6 13
Answer : V = 11
16. Examine whether the following game is fair.
Player Y
6
4
3
2
Player X 3
5
0
8
7
2
6
5
Answer : V = 0. Therefore, it is a fair game.
277
MBA-H2040 Quantitative Techniques
for Managers
LESSON 3
GAMES WITH NO SADDLE POINT
LESSON OUTLINE
The concept of a 2x2 game with no saddle point
The method of solution
LEARNING OBJECTIVES
After reading this lesson you should be able to
- understand the concept of a 2x2 game with no saddle point
- know the method of solution of a 2x2 game without saddle point
- solve a game with a given payoff matrix
- interpret the results obtained from the payoff matrix
2 x 2 zero-sum game
When each one of the first player A and the second player B has exactly two strategies,
we have a 2 x 2 game.
Motivating point
First let us consider an illustrative example.
Problem 1:
Examine whether the following 2 x 2 game has a saddle point
Player B
3 5
Player A
4
2
Solution:
First consider the minimum of each row.
Row
Minimum Value
1
3
2
2
Maximum of {3, 2} = 3
Next consider the maximum of each column.
Column
Maximum Value
1
4
2
5
Minimum of {4, 5}= 4
278
MBA-H2040 Quantitative Techniques
for Managers
We see that max {row minima} and min {column maxima} are not equal. Hence the
game has no saddle point.
Method of solution of a 2x2 zero-sum game without saddle point
Suppose that a 2x2 game has no saddle point. Suppose the game has the following pay-off
matrix.
Player B
Strategy
a
b
Player A Strategy
c
d
Since this game has no saddle point, the following condition shall hold:
Max{Min{a, }
b , Min{c, d}} Min{Max{a, }
c , Max{b, d}}
In this case, the game is called a mixed game. No strategy of Player A can be called the
best strategy for him. Therefore A has to use both of his strategies. Similarly no strategy
of Player B can be called the best strategy for him and he has to use both of his
strategies.
Let p be the probability that Player A will use his first strategy. Then the
probability that Player A will use his second strategy is 1-p.
If Player B follows his first strategy
Expected value of the pay-off to Player A
Expected value of the pay-off to Player A
Expected value of the pay-off to Player A
{
} {
}
arising from his first strategy
arising from his second strategy
ap (
c 1 )
p
(1)
In the above equation, note that the expected value is got as the product of the corresponding
values of the pay-off and the probability.
If Player B follows his second strategy
Expected value of the bp d(1 p)
(2)
pay-off to Player A
If the expected values in equations (1) and (2) are different, Player B will prefer the minimum
of the two expected values that he has to give to player A. Thus B will have a pure strategy.
This contradicts our assumption that the game is a mixed one. Therefore the expected values
of the pay-offs to Player A in equations (1) and (2) should be equal. Thus we have the
condition
279
MBA-H2040 Quantitative Techniques
for Managers
ap c(1 p)
bp d (1 p)
ap bp
(1 p)[d c]
p(a b)
(d c) p(d c)
p(a b) p(d c)
d c
p(a b d c)
d c
d c
p
(a d ) (b c)
a d b c d c
1 p
(a d ) (b c)
a b
(a d ) (b c)
The number of times A
The number of times A
d c
a b
{
}:{
}
:
will use first strategy
will use second strategy
(a d ) (b c) (a d ) (b c)
The expected pay-off to Player A
ap c(1 p)
c p(a c)
(d c)(a c)
c (ad)(bc)
c(a d) (b c
) (d c)(a c)
(a d ) (b c)
2
2
ac cd bc c ad cd ac c )
(a d ) (b c)
ad bc
(ad)(bc)
Therefore, the value V of the game is
ad bc
(a d ) (b c)
To find the number of times that B will use his first strategy and second strategy:
Let the probability that B will use his first strategy be r. Then the probability that B will use
his second strategy is 1-r.
When A use his first strategy
The expected value of loss to Player B with his first strategy = ar
The expected value of loss to Player B with his second strategy = b(1-r)
Therefore the expected value of loss to B = ar + b(1-r)
(3)
When A use his second strategy
The expected value of loss to Player B with his first strategy = cr
280
MBA-H2040 Quantitative Techniques
for Managers
The expected value of loss to Player B with his second strategy = d(1-r)
Therefore the expected value of loss to B = cr + d(1-r)
(4)
If the two expected values are different then it results in a pure game, which is a contradiction.
Therefore the expected values of loss to Player B in equations (3) and (4) should be equal.
Hence we have the condition
ar b(1 r) cr d (1 r)
ar b br cr d dr
ar br cr dr d b
r(a b c d ) d b
d b
r a bc d
d b
(ad)(bc)
Problem 2:
Solve the following game
Y
2 5
X
4 1
Solution:
First consider the row minima.
Row
Minimum Value
1
2
2
1
Maximum of {2, 1} = 2
Next consider the maximum of each column.
Column
Maximum Value
1
4
2
5
Minimum of {4, 5}= 4
We see that
281
MBA-H2040 Quantitative Techniques
for Managers
Max {row minima} min {column maxima}
So the game has no saddle point. Therefore it is a mixed game.
We have a = 2, b = 5, c = 4 and d = 1.
Let p be the probability that player X will use his first strategy. We have
d c
p (a d) (b c)
1 4
(21)(54)
3
3 9
3
6
1
2
1
1
The probability that player X will use his second strategy is 1-p = 1-
=
.
2
2
ad bc
2 20
18
Value of the game V =
3 .
(a d ) (b c)
3 9
6
Let r be the probability that Player Y will use his first strategy. Then the probability that Y
will use his second strategy is (1-r). We have
d b
r (a d)(b c)
1 5
(21)(54)
4
39
4
6
2
3
2
1
1 r 1
3
3
Interpretation.
1
1
p : (1-p) =
:
2
2
Therefore, out of 2 trials, player X will use his first strategy once and his second strategy
once.
282
MBA-H2040 Quantitative Techniques
for Managers
2
1
r : (1-r) =
:
3
3
Therefore, out of 3 trials, player Y will use his first strategy twice and his second strategy
once.
QUESTIONS
1. What is a 2x2 game with no saddle point? Explain.
2. Explain the method of solution of a 2x2 game without saddle point.
3. Solve the following game
Y
1
2 4
X
3 7
1
1
Answer: p = , r =
, V = 6
3
4
4. Solve the following game
Y
5
4
X
9
3
4
1
Answer: p =
, r = , V = -1
7
3
5. Solve the following game
Y
10 4
X
6
8
1
1
Answer: p =
, r =
, V = 7
4
2
6. Solve the following game
Y
20
8
X
2 10
1
1
Answer: p =
, r =
, V = 9
2
12
7. Solve the following game
Y
283
MBA-H2040 Quantitative Techniques
for Managers
10
2
X
1
5
1
1
Answer: p = , r =
, V = 4
3
4
8. Solve the following game
Y
12
6
X
6
9
1
1
Answer: p = , r = , V = 8
3
3
9. Solve the following game
Y
10
8
X
8
10
1
1
Answer: p =
, r =
, V = 9
2
2
10. Solve the following game
Y
16
4
X
4
8
1
1
Answer: p =
, r =
, V = 7
4
4
11. Solve the following game
Y
11
5
X
7
9
1
7
Answer: p =
, r =
, V = - 2
2
16
12. Solve the following game Y
9
3
X
5
7
1
5
Answer: p =
, r =
, V = - 2
2
12
284
MBA-H2040 Quantitative Techniques
for Managers
LESSON 4
THE PRINCIPLE OF DOMINANCE
LESSON OUTLINE
The principle of dominance
Dividing a game into sub games
LEARNING OBJECTIVES
After reading this lesson you should be able to
- understand the principle of dominance
-
solve a game using the principle of dominance
- solve a game by dividing a game into sub games
The principle of dominance
In the previous lesson, we have discussed the method of solution of a game without a saddle
point. While solving a game without a saddle point, one comes across the phenomenon of the
dominance of a row over another row or a column over another column in the pay-off matrix
of the game. Such a situation is discussed in the sequel.
In a given pay-off matrix A, we say that the ith row dominates the kth row if
a a for all j = 1,2,...,n
ij
kj
and
a a for at least one j.
ij
kj
In such a situation player A will never use the strategy corresponding to kth row,
because he will gain less for choosing such a strategy.
Similarly, we say the pth column in the matrix dominates the qth column if
a a for all i = 1,2,...,m
ip
iq
and
a a for at least one i.
ip
iq
In this case, the player B will loose more by choosing the strategy for the qth column than by
choosing the strategy for the pth column. So he will never use the strategy corresponding to
the qth column. When dominance of a row ( or a column) in the pay-off matrix occurs, we can
delete a row (or a column) from that matrix and arrive at a reduced matrix. This principle of
dominance can be used in the determination of the solution for a given game.
285
MBA-H2040 Quantitative Techniques
for Managers
Let us consider an illustrative example involving the phenomenon of dominance in a game.
Problem 1:
Solve the game with the following pay-off matrix:
Player B
I
II
III
IV
1 4
2
3
6
Player A
2 3
4
7
5
3 6
3
5
4
Solution:
First consider the minimum of each row.
Row
Minimum Value
1
2
2
3
3
3
Maximum of {2, 3, 3} = 3
Next consider the maximum of each column.
Column
Maximum Value
1
6
2
4
3
7
4
6
Minimum of {6, 4, 7, 6}= 4
The following condition holds:
Max {row minima} min {column maxima}
Therefore we see that there is no saddle point for the game under consideration.
Compare columns II and III.
Column II
Column III
2
3
4
7
3
5
286
MBA-H2040 Quantitative Techniques
for Managers
We see that each element in column III is greater than the corresponding element in column
II. The choice is for player B. Since column II dominates column III, player B will discard
his strategy 3.
Now we have the reduced game
I
II
IV
1 4
2
6
2 3
4
5
3 6
3
4
For this matrix again, there is no saddle point. Column II dominates column IV. The choice
is for player B. So player B will give up his strategy 4
The game reduces to the following:
I
II
1 4 2
2 3
4
3 6 3
This matrix has no saddle point.
The third row dominates the first row. The choice is for player A. He will give up his
strategy 1 and retain strategy 3. The game reduces to the following:
3 4
6 3
a b
Again, there is no saddle point. We have a 2x2 matrix. Take this matrix as
c d
Then we have a = 3, b = 4, c = 6 and d = 3. Use the formulae for p, 1-p, r, 1-r and V.
d c
p (a d) (b c)
3 6
(33)(64)
3
610
3
4
3
4
3
1
1 p 1
4
4
287
MBA-H2040 Quantitative Techniques
for Managers
d b
r (a d)(b c)
3 4
(33)(64)
1
610
1
4
1
4
1
3
1 r 1
4
4
The value of the game
ad bc
V (a d)(b c)
3x3 4x6
4
15
4
15
4
3 1
1 3
Thus, X =
, , 0, 0
and Y =
, , 0,0
are the optimal strategies.
4 4
4 4
Method of convex linear combination
A strategy, say s, can also be dominated if it is inferior to a convex linear combination of
several other pure strategies. In this case if the domination is strict, then the strategy s can be
deleted. If strategy s dominates the convex linear combination of some other pure strategies,
then one of the pure strategies involved in the combination may be deleted. The domination
will be decided as per the above rules. Let us consider an example to illustrate this case.
Problem 2:
Solve the game with the following pay-off matrix for firm A:
Firm B
B
B
B
B
B
1
2
3
4
5
A 4
8
2
5
6
1
A
4
0
6
8
5
Firm A 2
A 2
6
4
4
2
3
A
4
3
5
6
3
4
A 4 1
5 7
3
5
288
MBA-H2040 Quantitative Techniques
for Managers
Solution:
First consider the minimum of each row.
Row
Minimum Value
1
-2
2
0
3
-6
4
-3
5
-1
Maximum of {-2, 0, -6, -3, -1} = 0
Next consider the maximum of each column.
Column
Maximum Value
1
4
2
8
3
6
4
8
5
6
Minimum of { 4, 8, 6, 8, 6}= 4
Hence,
Maximum of {row minima} minimum of {column maxima}.
So we see that there is no saddle point. Compare the second row with the fifth row. Each
element in the second row exceeds the corresponding element in the fifth row. Therefore, A
2
dominates A . The choice is for firm A. It will retain strategy A and give up strategy A .
5
2
5
Therefore the game reduces to the following.
B
B
B
B
B
1
2
3
4
5
A 4
8
2 5
6
1
A
4
0
6 8
5
2
A 2
6 4 4
2
3
A 4
3
5 6
3
4
Compare the second and fourth rows. We see that A dominates A . So, firm A will retain
2
4
the strategy A and give up the strategy A . Thus the game reduces to the following:
2
4
289
MBA-H2040 Quantitative Techniques
for Managers
B
B
B
B
B
1
2
3
4
5
A 4
8
2
5 6
1
A
4
0
6
8
5
2
A 2
6
4
4 2
3
Compare the first and fifth columns. It is observed that B1 dominates B5. The choice is for
firm B. It will retain the strategy B and give up the strategy B . Thus the game reduces to the
1
5
following
B
B
B
B
1
2
3
4
A 4
8
2 5
1
A
4
0
6 8
2
A 2
6 4 4
3
Compare the first and fourth columns. We notice that B1 dominates B4. So firm B will discard
the strategy B and retain the strategy B . Thus the game reduces to the following:
4
1
B
B
B
1
2
3
A 4
8
2
1
A
4
0
6
2
A 2
6
4
3
For this reduced game, we check that there is no saddle point.
Now none of the pure strategies of firms A and B is inferior to any of their other
strategies. But, we observe that convex linear combination of the strategies B and B
2
3
dominates B , i.e. the averages of payoffs due to strategies B and B ,
1
2
3
8 2 0 6 6
4
,
,
3,3,
5
2
2
2
dominate B . Thus B may be omitted from consideration. So we have the reduced matrix
1
1
B
B
2
3
A 8
2
1
A
0
6
2
A 6
4
3
Here, the average of the pay-offs due to strategies A and A of firm A, i.e.
1
2
8 0 2
6
,
4,
2 dominates the pay-off due to A . So we get a new reduced 2x2 pay-
2
2
3
off matrix.
290
MBA-H2040 Quantitative Techniques
for Managers
Firm B's strategy
B
B
2
3
Firm A's strategy A 8
2
1
A 0
6
2
We have a = 8, b = -2, c = 0 and d = 6.
d c
p (a d) (b c)
6 0
(68)(20)
6
16
3
8
3
5
1 p 1
8
8
d b
r (a d)(b c)
6 (2)
16
8
16
1
2
1
1
1 r 1
2
2
Value of the game:
ad bc
V (a d)(b c)
6 8
x 0x(2)
16
48
3
16
So the optimal strategies are
3 5
1 1
A = , , 0, 0, 0 and B = 0, , , 0, 0 .
8 8
2 2
The value of the game = 3. Thus the game is favourable to firm A.
291
MBA-H2040 Quantitative Techniques
for Managers
Problem 3:
For the game with the following pay-off matrix, determine the saddle point
Player B
I
II
III
IV
1 2 1
0
3
Player A
2
1
0
3
2
3 3
2
1
4
Solution:
Column II
Column III
1
1
0
0 1
2 0
3
3 0
3 2
1
1 2
The choice is with the player B. He has to choose between strategies II and III. He will lose
more in strategy III than in strategy II, irrespective of what strategy is followed by A. So he
will drop strategy III and retain strategy II. Now the given game reduces to the following
game.
I
II
IV
1 2
1
3
2
1
0
2
3 3
2
4
Consider the rows and columns of this matrix.
Row minimum:
I Row
:
-3
II Row
:
0
Maximum of {-3, 0, -3} = 0
III Row
:
-3
Column maximum:
I Column
:
2
II Column
:
0
Minimum of {2, 0, 4} = 0
III Column
:
4
We see that
Maximum of row minimum = Minimum of column maximum = 0.
So, a saddle point exists for the given game and the value of the game is 0.
292
MBA-H2040 Quantitative Techniques
for Managers
Interpretation:
No player gains and no player loses. i.e., The game is not favourable to any player. i.e. It is a
fair game.
Problem 4:
Solve the game
Player B
4 8 6
Player A 6
2 10
4 5 7
Solution:
First consider the minimum of each row.
Row
Minimum
1
4
2
2
3
4
Maximum of {4, 2, 4} = 4
Next, consider the maximum of each column.
Column
Maximum
1
6
2
8
3
10
Minimum of {6, 8, 10} = 6
Since Maximum of { Row Minima} and Minimum of { Column Maxima } are different, it
follows that the given game has no saddle point.
Denote the strategies of player A by A , A , A . Denote the strategies of player B by B , B , B .
1
2
3
1
2
3
Compare the first and third columns of the given matrix.
B
B
1
3
4
6
6
10
7
7
The pay-offs in B are greater than or equal to the corresponding pay-offs in B . The
3
1
player B has to make a choice between his strategies 1 and 3. He will lose more if he follows
293
MBA-H2040 Quantitative Techniques
for Managers
strategy 3 rather than strategy 1. Therefore he will give up strategy 3 and retain strategy 1.
Consequently, the given game is transformed into the following game:
B
B
1
2
A 4
8
1
A 6
2
2
A 4
5
3
Compare the first and third rows of the above matrix.
B
B
1
2
A 4
8
1
A 4
5
3
The pay-offs in A are greater than or equal to the corresponding pay-offs in A . The player
1
3
A has to make a choice between his strategies 1 and 3. He will gain more if he follows
strategy 1 rather than strategy 3. Therefore he will retain strategy 1 and give up strategy 3.
Now the given game is transformed into the following game.
B
B
1
2
A 4
8
1
A 6
2
2
It is a 2x2 game. Consider the row minima.
Row
Minimum
1
4
2
2
Maximum of {4, 2} = 4
Next, consider the maximum of each column.
Column
Maximum
1
6
2
8
Minimum of {6, 8} = 6
Maximum {row minima} and Minimum {column maxima } are not equal
Therefore, the reduced game has no saddle point. So, it is a mixed game
a b 4 8
Take
. We have a = 4, b = 8, c = 6 and d = 2.
c d 6 2
The probability that player A will use his first strategy is p. This is calculated as
294
MBA-H2040 Quantitative Techniques
for Managers
d c
p (a d) (b c)
2 6
(42)(86)
4
614
4 1
8 2
The probability that player B will use his first strategy is r. This is calculated as
d b
r (a d)(b c)
2 8
8
6
8
3
4
Value of the game is V. This is calculated as
ad bc
V (a d)(b c)
4x2 8x6
8
8 48
8
40
5
8
Interpretation
Out of 3 trials, player A will use strategy 1 once and strategy 2 once. Out of 4 trials, player B
will use strategy 1 thrice and strategy 2 once. The game is favourable to player A.
Problem 5: Dividing a game into sub-games
Solve the game with the following pay-off matrix.
Player B
1
2
3
I 4
6
3
Player A
II 3
3
4
III 2
3
4
Solution:
First, consider the row mimima.
295
MBA-H2040 Quantitative Techniques
for Managers
Row
Minimum
1
-4
2
-3
3
-3
Maximum of {-4, -3, -3} = -3
Next, consider the column maxima.
Column
Maximum
1
2
2
6
3
4
Minimum of {2, 6, 4} = 2
We see that Maximum of { row minima} Minimum of { column maxima}.
So the game has no saddle point. Hence it is a mixed game. Compare the first and third
columns.
I Column
III Column
4
3
4 3
3
4
3 4
2
4
2 4
We assert that Player B will retain the first strategy and give up the third strategy. We get the
following reduced matrix.
4
6
3
3
2
3
We check that it is a game with no saddle point.
Sub games
Let us consider the 2x2 sub games. They are:
4 6 4
6 3
3
3
3 2 3 2 3
First, take the sub game
4 6
3
3
296
MBA-H2040 Quantitative Techniques
for Managers
Compare the first and second columns. We see that 4 6 and 3 3 . Therefore, the
4
game reduces to . Since 4 3, it further reduces to ?3.
3
Next, consider the sub game
4
6
2 3
We see that it is a game with no saddle point. Take a = -4, b = 6, c = 2, d = -3. Then the
value of the game is
ad bc
V (a d)(b c)
(4)(3) (6)(2)
(4 3) (6 2)
0
3
3
Next, take the sub game
. In this case we have a = -3, b = 3, c = 2 and d = -3. The
2 3
value of the game is obtained as
ad bc
V (a d)(b c)
(3)( 3
) (3)(2)
(3 3) (3 2)
9 6
3
6
5
11
Let us tabulate the results as follows:
Sub game
Value
4 6
3
3
-3
4
6
0
2 3
3
3
3
2 3
-
11
The value of 0 will be preferred by the player A. For this value, the first and third
strategies of A correspond while the first and second strategies of the player B correspond to
the value 0 of the game. So it is a fair game.
297
MBA-H2040 Quantitative Techniques
for Managers
QUESTIONS
1. Explain the principle of dominance in the theory of games.
2. Explain how a game can be solved through sub games.
3. Solve the following game by the principle of dominance:
Player B
Strategies
I
II
III
IV
1 8
10
9
14
Player A Strategies 2 10
11
8
12
3 13
12
14
13
Answer: V = 12
4. Solve the game by the principle of dominance:
1 7 2
6
2 7
5 2 6
Answer: V = 4
5. Solve the game with the following pay-off matrix
6 3 1 0
3
3 2 4 2 1
3
2
11
Answer : p
, r
, V
5
5
5
6. Solve the game
8
7
6
1
2
12 10 12
0
4
1
4
6
8 14 16
4
7
70
Answer : p
, r
, V
9
9
9
298
MBA-H2040 Quantitative Techniques
for Managers
LESSON 5
GRAPHICAL SOLUTION OF A 2x2 GAME WITH NO SADDLE POINT
LESSON OUTLINE
The principle of graphical solution
Numerical example
LEARNING OBJECTIVES
After reading this lesson you should be able to
-
understand the principle of graphical solution
-
derive the equations involving probability and expected value
- solve numerical problems
Example: Consider the game with the following pay-off matrix.
Player B
2 5
Player A
4 1
First consider the row minima.
Row
Minimum
1
2
2
1
Maximum of {2, 1} = 2.
Next, consider the column maxima.
Column
Maximum
1
4
2
5
Minimum of {4, 5} = 4.
We see that Maximum { row minima} Minimum { column maxima }
So, the game has no saddle point. It is a mixed game.
Equations involving probability and expected value:
Let p be the probability that player A will use his first strategy.
Then the probability that A will use his second strategy is 1-p.
299
MBA-H2040 Quantitative Techniques
for Managers
Let E be the expected value of pay-off to player A.
When B uses his first strategy
The expected value of pay-off to player A is given by
E 2 p 4(1 p)
2 p 4 4 p
(1)
4 2 p
When B uses his second strategy
The expected value of pay-off to player A is given by
E 5 p 1(1 p)
5 p 1 p
(2)
4 p 1
300
MBA-H2040 Quantitative Techniques for Managers
Consider equations (1) and (2). For plotting the two equations on a graph sheet, get some points on them
as follows:
E = -2p+4
p
0
1
0.5
E
4
2
3
E = 4p+1
p
0
1
0.5
E
1
5
3
Graphical solution:
Procedure:
Take probability and expected value along two rectangular axes in a graph sheet. Draw two straight
lines given by the two equations (1) and (2). Determine the point of intersection of the two straight lines
in the graph. This will give the common solution of the two equations (1) and (2). Thus we would obtain
the value of the game.
Represent the two equations by the two straight lines AB and CD on the graph sheet. Take the
point of intersection of AB and CD as T. For this point, we have p = 0.5 and E = 3. Therefore, the
value V of the game is 3.
301
MBA-H2040 Quantitative Techniques for Managers
E
E=4P+1
D
A
4
T
3
2 C B
E=-2P+4
1
P
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Problem 1:
Solve the following game by graphical method.
Player B
18
2
Player A
6
4
Solution:
First consider the row minima.
Row
Minimum
1
- 18
2
- 4
302
MBA-H2040 Quantitative Techniques for Managers
Maximum of {-18, - 4} = - 4.
Next, consider the column maxima.
Column
Maximum
1
6
2
2
Minimum of {6, 2} = 2.
We see that Maximum { row minima} Minimum { column maxima }
So, the game has no saddle point. It is a mixed game.
Let p be the probability that player A will use his first strategy.
Then the probability that A will use his second strategy is 1-p.
When B uses his first strategy
The expected value of pay-off to player A is given by
E 18
p 6 (1 p)
18
p 6 6 p
(I)
24
p 6
When B uses his second strategy
The expected value of pay-off to player A is given by
E 2 p 4 (1 p)
2 p 4 4 p (II)
6 p 4
Consider equations (I) and (II). For plotting the two equations on a graph sheet, get some points on them
as follows:
E = -24 p + 6
p
0
1
0.5
E
6
-18 -6
E = 6p-4
p
0
1
0.5
E
-4
2
-1
Graphical solution:
Take probability and expected value along two rectangular axes in a graph sheet. Draw two straight lines
given by the two equations (1) and (2). Determine the point of intersection of the two straight lines in the
303
MBA-H2040 Quantitative Techniques for Managers
graph. This will provide the common solution of the two equations (1) and (2). Thus we would get the
value of the game.
Represent the two equations by the two straight lines AB and CD on the graph sheet. Take the
1
point of intersection of AB and CD as T. For this point, we have p =
and E = -2. Therefore, the
3
value V of the game is -2.
E
8
6
A
4
E=6P-4
2
0 D
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.
-2
C
-4
-6
-8
B
E=-24P+6
304
MBA-H2040 Quantitative Techniques for Managers
QUESTIONS
1. Explain the method of graphical solution of a 2x2 game.
2. Obtain the graphical solution of the game
10
6
8 12
1
Answer: p =
, V = 9
2
3. Graphically solve the game
4
10
8
6
1
Answer: p =
, V = 7
4
4. Find the graphical solution of the game
12
12
2
6
1
3
Answer: p =
, V =
4
2
5. Obtain the graphical solution of the game
10
6
8 12
1
Answer: p =
, V = 9
2
6. Graphically solve the game
3
5
5
1
3
7
Answer: p =
, V =
4
2
305
MBA-H2040 Quantitative Techniques for Managers
LESSON 6
2 x n ZERO-SUM GAMES
LESSON OUTLINE
A 2 x n zero-sum game
Method of solution
Sub game approach and graphical method
Numerical example
LEARNING OBJECTIVES
After reading this lesson you should be able to
- understand the concept of a 2 x n zero-sum game
- solve numerical problems
The concept of a 2 x n zero-sum game
When the first player A has exactly two strategies and the second player B has n (where n is three
or more) strategies, there results a 2 x n game. It is also called a rectangular game. Since A has
two strategies only, he cannot try to give up any one of them. However, since B has many
strategies, he can make out some choice among them. He can retain some of the advantageous
strategies and discard some disadvantageous strategies. The intention of B is to give as minimum
payoff to A as possible. In other words, B will always try to minimize the loss to himself.
Therefore, if some strategies are available to B by which he can minimize the payoff to A, then B
will retain such strategies and give such strategies by which the payoff will be very high to A.
Approaches for 2 x n zero-sum game
There are two approaches for such games: (1) Sub game approach and (2) Graphical approach.
Sub game approach
The given 2 x n game is divided into 2 x 2 sub games. For this purpose, consider all possible 2 x 2 sub
matrices of the payoff matrix of the given game. Solve each sub game and have a list of the values of
each sub game. Since B can make out a choice of his strategies, he will discard such of those sub games
which result in more payoff to A. On the basis of this consideration, in the long run, he will retain two
strategies only and give up the other strategies.
Problem
Solve the following game
306
MBA-H2040 Quantitative Techniques for Managers
Player B
8
2
6
9
Player A
3 5 10 2
Solution:
Let us consider all possible 2x2 sub games of the given game. We have the following sub games:
8
2
1.
3
5
8
6
2.
3 10
8
9
3.
3 2
2
6
4.
5 10
2
9
5.
5 2
6
9
6.
10 2
Let E be the expected value of the pay off to player A. Let p be the probability that player A will use his
first strategy. Then the probability that he will use his second strategy is 1-p. We form the equations for
E in all the sub games as follows:
Sub game (1)
Equation 1: E 8 p 3(1 p) 5 p 3
Equation 2: E 2
p 5(1 p) 7
p 5
Sub game (2)
Equation 1: E 8 p 3(1 p) 5 p 3
Equation 2: E 6
p 10(1 p) 1
6 p 10
Sub game (3)
Equation 1: E 8 p 3(1 p) 5 p 3
Equation 2: E 9 p 2(1 p) 7 p 2
Sub game (4)
Equation 1: E 2
p 5(1 p) 7
p 5
Equation 2: E 6
p 10(1 p) 1
6 p 10
307
MBA-H2040 Quantitative Techniques for Managers
Sub game (5)
Equation 1: E 2
p 5(1 p) 7
p 5
Equation 2: E 9 p 2(1 p) 7 p 2
Sub game (6)
Equation 1: E 6
p 10(1 p) 1
6 p 10
Equation 2: E 9 p 2(1 p) 7 p 2
Solve the equations for each sub game. Let us tabulate the results for the various sub games. We have
the following:
Sub game
p
Expected value E
1
1
23
6
6
2
1
14
3
3
3
1
11
2
2
4
5
10
9
9
5
3
7
14
2
6
8
102
23
23
Interpretation:
Since player A has only 2 strategies, he cannot make any choice on the strategies. On the other hand,
player B has 4 strategies. Therefore he can retain any 2 strategies and give up the other 2 strategies.
This he will do in such a way that the pay-off to player A is at the minimum. The pay-off to A is the
2
6
minimum in the case of sub game 4. i.e., the sub game with the matrix
.
5 10
Therefore, in the long run, player B will retain his strategies 2 and 3 and give up his strategies 1
5
and 4. In that case, the probability that A will use his first strategy is p =
and the probability that he
9
4
will use his second strategy is 1-p =
. i.e., Out of a total of 9 trials, he will use his first strategy five
9
308
MBA-H2040 Quantitative Techniques for Managers
10
times and the second strategy four times. The value of the game is
. The positive sign of V shows
9
that the game is favourable to player A.
GRAPHICAL SOLUTION:
Now we consider the graphical method of solution to the given game.
Draw two vertical lines MN and RS. Note that they are parallel to each other. Draw UV
perpendicular to MN as well as RS. Take U as the origin on the line MN. Take V as the origin on the
line RS.
Mark units on MN and RS with equal scale. The units on the two lines MN and RS are taken as
the payoff numbers. The payoffs in the first row of the given matrix are taken along the line MN while
the payoffs in the second row are taken along the line RS.
We have to plot the following points: (8, 3), (-2, 5), (-6, 10), (9, 2). The points 8, -2, -6, 9 are
marked on MN. The points 3, 5, 10, 2 are marked on RS.
Join a point on MN with the corresponding point on RS by a straight line. For example, join the
point 8 on MN with the point 3 on RS. We have 4 such straight lines. They represent the 4 moves of the
second player. They intersect in 6 points. Take the lowermost point of intersection of the straight lines. It
is called the Maximin point. With the help of this point, identify the optimal strategies for the second
player. This point corresponds to the points ?2 and ?6 on MN and 5 and 10 on RS. They correspond to
2
6
the sub game with the matrix
.
5 10
The points ?2 and ?6 on MN correspond to the second and third strategies of the second player.
Therefore, the graphical method implies that, in the long run, the second player will retain his strategies
2 and 3 and give up his strategies 1 and 4.
We graphically solve the sub game with the above matrix. We have to solve the two equations E
= -7 p + 5 and E = - 16 p + 10. Represent the two equations by two straight lines AB and CD on the
5
graph sheet. Take the point of intersection of AB and CD as T. For this point, we have p =
and E =
9
10
10
. Therefore, the value V of the game is
. We see that the probability that first player will use his
9
9
5
4
first strategy is p = and the probability that he will use his second strategy is 1-p =
.
9
9
M R
10 10
309
MBA-H2040 Quantitative Techniques for Managers
9 9
8 8
7
7
6
5 6
4 5
3
4
2
3
1
2
1
Maximin point
U 0 0 V
-1 -1
-2 -2
-3 -3
N S
E
10 C
9
8
7
6
A
5
4
3
2
T
1
P
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
B
D
310
MBA-H2040 Quantitative Techniques for Managers
E=-7P+5
E=-16P+10
E = - 7p+5 E = - 16p+10
p
0
1
0.5
p
0
1
0.5
E
5
- 2 1.5
E 10
- 6
2
QUESTIONS
1. Explain a 2 x n zero-sum game.
2. Describe the method of solution of a 2 x n zero-sum game.
3. Solve the following game:
Player B
10
2 6
Player A
1 5 8
1
Answer: p = , V = 4
3
LESSON 7
m x 2 ZERO-SUM GAMES
LESSON OUTLINE
An m x 2 zero-sum game
Method of solution
Sub game approach and graphical method
Numerical example
LEARNING OBJECTIVES
After reading this lesson you should be able to
- understand the concept of an m x 2 zero-sum game
-
solve numerical problems
The concept of an m x 2 zero-sum game
When the second player B has exactly two strategies and the first player A has m (where m is
three or more) strategies, there results an m x 2 game. It is also called a rectangular game. Since B
has two strategies only, he will find it difficult to discard any one of them. However, since A has
311
MBA-H2040 Quantitative Techniques for Managers
more strategies, he will be in a position to make out some choice among them. He can retain some
of the most advantageous strategies and give up some other strategies. The motive of A is to get as
maximum payoff as possible. Therefore, if some strategies are available to A by which he can get
more payoff to himself, then he will retain such strategies and discard some other strategies which
result in relatively less payoff.
Approaches for m x 2 zero-sum game
There are two approaches for such games: (1) Sub game approach and (2) Graphical approach.
Sub game approach
The given m x 2 game is divided into 2 x 2 sub games. For this purpose, consider all possible 2 x 2 sub
matrices of the payoff matrix of the given game. Solve each sub game and have a list of the values of
each sub game. Since A can make out a choice of his strategies, he will be interested in such of those
sub games which result in more payoff to himself. On the basis of this consideration, in the long run, he
will retain two strategies only and give up the other strategies.
Problem
Solve the following game:
Player B
Strategies
I
II
1
5
8
Player A Strategies
2
2
10
3
12
4
4
6
5
Solution:
Let us consider all possible 2x2 sub games of the given game. We have the following sub games:
5
8
7.
2
10
5
8
8.
12
4
5
8
9.
6
5
2
10
10.
12
4
312
MBA-H2040 Quantitative Techniques for Managers
2
10
11.
6
5
12
4
12.
6
5
Let E be the expected value of the payoff to player A. i.e., the loss to player B. Let r be the probability
that player B will use his first strategy. Then the probability that he will use his second strategy is 1-r.
We form the equations for E in all the sub games as follows:
Sub game (1)
Equation 1: E 5r 8(1 r) 3r 8
Equation 2: E 2
r 10(1 r) 1
2r 10
Sub game (2)
Equation 1: E 5r 8(1 r) 3
r 8
Equation 2: E 12r 4(1 r) 8r 4
Sub game (3)
Equation 1: E 5r 8(1 r) 3
r 8
Equation 2: E 6r 5(1 r) r 5
Sub game (4)
Equation 1: E 2
r 10(1 r) 1
2r 10
Equation 2: E 12r 4(1 r) 8r 4
Sub game (5)
Equation 1: E 2
r 10(1 r) 1
2r 10
Equation 2: E 6r 5(1 r) r 5
Sub game (6)
Equation 1: E 12r 4(1 r) 8r 4
Equation 2: E 6r 5(1 r) r 5
Solve the equations for each 2x2 sub game. Let us tabulate the results for the various sub games. We
have the following:
Sub game
R
Expected value E
1
2
22
9
3
313
MBA-H2040 Quantitative Techniques for Managers
2
4
76
11
11
3
3
23
4
4
4
3
32
10
5
5
5
70
13
13
6
1
36
7
7
Interpretation:
Since player B has only 2 strategies, he cannot make any choice on his strategies. On the other
hand, player A has 4 strategies and so he can retain any 2 strategies and give up the other 2 strategies.
Since the choice is with A, he will try to maximize the payoff to himself. The pay-off to A is the
5
8
maximum in the case of sub game 1. i.e., the sub game with the matrix
.
2
10
Therefore, player A will retain his strategies 1 and 2 and discard his strategies 3 and 4, in the
2
long run. In that case, the probability that B will use his first strategy is r =
and the probability that
9
7
he will use his second strategy is 1-r =
. i.e., Out of a total of 9 trials, he will use his first strategy two
9
times and the second strategy seven times.
22
The value of the game is
. The positive sign of V shows that the game is favourable to player A.
3
GRAPHICAL SOLUTION:
Now we consider the graphical method of solution to the given game.
Draw two vertical lines MN and RS. Note that they are parallel to each other. Draw UV
perpendicular to MN as well as RS. Take U as the origin on the line MN. Take V as the origin on the
line RS.
Mark units on MN and RS with equal scale. The units on the two lines MN and RS are taken as the
payoff numbers. The payoffs in the first row of the given matrix are taken along the line MN while
the payoffs in the second row are taken along the line RS.
We have to plot the following points: (5, 8), (-2, 10), (12, 4), (6, 5).The points 5, -2, 12, 6 are marked
on MN. The points 8, 10, 4, 5 are marked on RS.
314
MBA-H2040 Quantitative Techniques for Managers
Join a point on MN with the corresponding point on RS by a straight line. For example, join the
point 5 on MN with the point 8 on RS. We have 4 such straight lines. They represent the 4 moves of the
first player. They intersect in 6 points. Take the uppermost point of intersection of the straight lines. It is
called the Minimax point. With the help of this point, identify the optimal strategies for the first player.
This point corresponds to the points 5 and -2 on MN and 8 and 10 on RS. They correspond to the sub
5
8
game with the matrix
. The points 5 and -2 on MN correspond to the first and second
2
10
strategies of the first player. Therefore, the graphical method implies that the first player will retain his
strategies 1 and 2 and give up his strategies 3 and 4, in the long run.
We graphically solve the sub game with the above matrix. We have to solve the two equations E
= - 3 r + 8 and E = - 12 r + 10. Represent the two equations by two straight lines AB and CD on the
2
graph sheet. Take the point of intersection of AB and CD as T. For this point, we have r =
and E =
9
22
22
. Therefore, the value V of the game is
. We see that the probability that the second player will
3
3
2
7
use his first strategy is r =
and the probability that he will use his second strategy is 1-r =
.
9
9
315
MBA-H2040 Quantitative Techniques for Managers
M R
12 12
11 11
10 10
9 9
8 minimax 8
7 point 7
6 6
5 5
4 4
3 3
2 2
1 1
U 0 0 V
-1 -1
-2 -2
-3 -3
-4 -4
-5 -5
-6 -6
-7 -7
-8 -8
-9 -9
-10 -10
N S
316
MBA-H2040 Quantitative Techniques for Managers
E
10
9 C
8
A T
7
6
5
4 B
E= -3r+8
3
2
1
r
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
D
E= -12r+10
E = - 3r+8 E = - 12r+10
p
0
1
0.5
p
0
1
0.5
E
8
5
6.5
E
10
- 2
4
317
MBA-H2040 Quantitative Techniques for Managers
QUESTIONS
1.
What is an m x 2 zero-sum game? Explain.
2.
How will you solve an m x 2 zero-sum game? Explain.
3.
Solve the following game:
Player B
Strategies
I
II
1
20
8
Player A Strategies
2
5
2
3
8
12
1
Answer: r =
, V = 11
4
318
MBA-H2040 Quantitative Techniques for Managers
LESSON 8
LINEAR PROGRAMMING APPROACH TO GAME THEORY
LESSON OUTLINE
4.
How to solve a game with LPP?
5.
Formulation of LPP
6.
Solution by simplex method
LEARNING OBJECTIVES
After reading this lesson you should be able to
- understand the transformation of a game into LPP
- solve a game by simplex method
Introduction
When there is neither saddle point nor dominance in a problem of game theory and the payoff matrix is
of order 3x3 or higher, the probability and graphical methods cannot be employed. In such a case, linear
programming approach may be followed to solve the game.
Linear programming technique:
A general approach to solve a game by linear programming technique is presented below. Consider the
following m n game:
Player B
Y
Y
Y
1
2
n
X a
a
a
1
11
12
1n
Player A X
a
a
a
2
21
22
2n
X a
a
a
m
1
m
m 2
mn
It is required to determine the optimal strategy for A = X , X ,...X
and B = Y ,Y ,...Y . First we
1
2
n
1
2
m
shall determine the optimal strategies of player B.
If player A adopts strategy X , then the expected value of loss to B is
1
a Y a Y ... a Y V ,
11 1
12 2
1n n
where V is the value of game. If A adopts strategy X , then the expected value of loss to B is
2
319
MBA-H2040 Quantitative Techniques for Managers
a Y a Y ... a Y V
21 1
22 2
2n n
and so on. Also we have
Y Y ... Y 1
1
2
n
and
Y 0 for all j.
j
Without loss of generality, we can assume that V > 0. Divide each of the above relation by V and let
Y
'
j
Y
.
j
V
Then we have
Yj
1
'
Y
j
V
V
From this we obtain
'
a Y a Y ' ... a Y ' 1,
11
1
12
2
1n
n
a Y ' a Y ' ... a Y ' 1,
21
1
22
2
2n
n
a Y ' a Y ' ... a Y ' 1
m1
1
m2
2
mn
n
and
1
'
Y Y ' ... Y '
1
2
n
V
with '
Y
0 for all j.
j
The objective of player B is to minimise the loss to himself . Thus the problem is to minimize V, or
1
equivalently to maximise
. Therefore, the objective of player B is to maximise the value of
V
'
Y Y ' ... Y ' subject to the m linear constraints provided above.
1
2
n
Statement of the problem:
Maximise: '
Y Y ' ... Y ' , subject to
1
2
n
'
a Y a Y ' ... a Y ' 1,
11
1
12
2
1n
n
a Y ' a Y ' ... a Y ' 1,
21
1
22
2
2n
n
a Y ' a Y ' ... a Y ' 1
m1
1
m2
2
mn
n
'
Y , Y ' ,...,Y ' 0.
1
2
n
320
MBA-H2040 Quantitative Techniques for Managers
We can use simplex method to solve the above problem. For this purpose, we have to introduce non-
negative slack variables s , s ,..., s to each of the inequalities. So the problem can be restated as
1
2
m
follows:
Restatement of the problem:
Maximise: '
Y Y ' ... Y ' 0s 0s ... 0s subject to
1
2
n
1
2
m
'
a Y a Y ' ... a Y ' s 1,
11
1
12
2
1n
n
1
a Y ' a Y ' ... a Y ' s 1,
21
1
22
2
2n
n
2
a Y ' a Y ' ... a Y ' s 1
m1
1
m2
2
mn
n
m
with Y Y ' V for all j and s 0, s 0,..., s 0 .
j
j
1
2
m
Thus we get the optimal strategy for player B to be (Y ,Y ,...,Y ).
1
2
n
In a similar manner we can determine the optimal strategy for player A.
Application:
We illustrate the method for a 2X2 zero sum game.
Problem 1:
Solve the following game by simplex method for LPP:
Player B
3 6
Player A
5 2
Solution :
Row minima :
I row
: 3
II row
: 2
Maximum of {3,2} = 3
Column maxima:
I column
: 5
II column
: 6
Minimum of {5,6} = 5
So, Maximum of {Row minima} Minimum of {Column maxima}.
Therefore the given game has no saddle point. It is a mixed game. Let us convert the given game into a
LPP.
Problem formulation:
321
MBA-H2040 Quantitative Techniques for Managers
Let V denote the value of the game. Let the probability that the player B will use his first strategy be r
and second strategy be s. Let V denote the value of the game.
When A follows his first strategy:
The expected payoff to A (i.e., the expected loss to B) = 3 r + 6 s.
This pay-off cannot exceed V. So we have
3 r + 6 s V
(1)
When A follows his second strategy:
The expected pay-off to A (i.e., expected loss to B) = 5 r + 2 s.
This cannot exceed V. Hence we obtain the condition
5 r + 2 s V
(2)
From (1) and (2) we have
r
s
3
6 1
V
V
r
s
and 5
2 1
V
V
r
s
Substitute
x,
y .
V
V
Then we have
3x 6 y 1
and 5x 2 y
1
where r and s are connected by the relation
r s 1.
r
s
1
i.e.,
V
V
V
1
i.e., x y
V
1
B will try to minimise V. i.e., He will try to maximise
. Thus we have the following LPP.
V
1
Maximize
x y ,
V
subject to the restrictions
3x 6 y 1,
5x 2 y 1,
x 0, y 0
Solution of LPP:
322
MBA-H2040 Quantitative Techniques for Managers
Introduce two slack variables s , s . Then the problem is transformed into the
1
2
following one:
1
Maximize
x y 0.s 0.s
1
2
V
subject to the constraints
3x 6 y 1.s 0.s 1,
1
2
5x 2 y 0.s 1.s 1,
1
2
x 0, y 0, s 0, s 0
1
2
Let us note that the above equations can be written in the form of a single matrix equation as
A X = B
x
3 6 1
0
y
1
where A =
, X = , B = .
5
2 0 1
s
1
1
s
2
The entries in B are referred to as the b ? values. Initially, the basic variables are s , s . We have the
1
2
following simplex tableau:
x
y
s
s
1
2
b ? value
s - row
1
3
6
1
0
1
s - row
2
5
2
0
1
1
Objective
-1
-1
0
0
0
function row
Consider the negative elements in the objective function row. They are ?1, -1. The absolute values are
1, 1. There is a tie between these coefficients. To resolve the tie, we select the variable x. We take the
new basic variable as x. Consider the ratio of b-value to x-value. We have the following ratios:
1
s - row
:
1
3
1
s - row
:
2
5
1 1
1
Minimum of {
, } = .
3 5
5
Hence select s as the leaving variable. Thus the pivotal element is 5. We obtain the following tableau
2
at the end of Iteration No. 1.
323
MBA-H2040 Quantitative Techniques for Managers
x
y
s
s
1
2
b ? value
s - row
24
3
2
1
0
1
5
5
5
x - row
2
1
1
1
0
5
5
5
Objective
3
1
1
0
0
function row
5
5
5
3
Now, the negative element in the objective function row is
. This corresponds to y. We take the
5
new basic variable as y. Consider the ratio of b-value to y-value. We have the following ratios:
2
5
1
s - row
:
1
24
12
5
1
5
1
x - row
:
2
2
5
1 1
1
Minimum of
,
12
2 12
24
Hence select s as the leaving variable. The pivotal element is
. We get the following tableau at the
1
5
end of Iteration No. 2.
x
y
s
s
1
2
b ? value
y - row
5
1
1
0
1
-
24
8
12
x - row
1
1
1
1
0 -
12
4
6
Objective
1
1
1
0
0
function row
8
8
4
Since both x and y have been made basic variables, we have reached the stopping condition.
324
MBA-H2040 Quantitative Techniques for Managers
1
1
1
1
The optimum value of
is
. This is provided by x =
and y =
. Thus the optimum
V
4
6
12
r
s
4
2
value of the game is obtained as V = 4. Using the relations
x,
y , we obtain r and
V
V
6
3
4
1
s
.
12
3
Problem 2:
Solve the following game:
Player B
2 5
Player A
4 1
Solution:
The game has no saddle point. It is a mixed game. Let the probability that B will use his first strategy
be r. Let the probability that B will use his second strategy be s. Let V be the value of the game.
When A follows his first strategy:
The expected payoff to A (i.e., the expected loss to B) = 2 r +5 s.
The pay-off to A cannot exceed V. So we have
2 r + 5 s V
(I)
When A follows his second strategy:
The expected pay-off to A (i.e., expected loss to B) = 4 r + s.
The pay-off to A cannot exceed V. Hence we obtain the condition
4 r + s V
(II)
From (I) and (II) we have
r
s
2
5 1
V
V
r
s
and 4
1
V
V
Substitute
r
s
x and
.
y
V
V
Thus we have
2x 5 y 1
and 4x y
1
where r and s are connected by the relation
r s 1.
325
MBA-H2040 Quantitative Techniques for Managers
r
s
1
i.e.,
V
V
V
1
i.e., x y
V
1
The objective of B is to minimise V. i.e., He will try to maximise
.
V
Thus we are led to the following linear programming problem:
1
Maximize
= x + y
V
subject to the constraints
2x 5y 1,
4x y 1,
x 0, y 0.
To solve this linear programming problem, we use simplex method as detailed below.
Introduce two slack variables s , s . Then the problem is transformed into the following one:
1
2
1
Maximize
x y 0.s 0.s
1
2
V
subject to the constraints
2x 5 y 1.s 0.s 1,
1
2
4x y 0.s 1.s 1,
1
2
x 0, y 0, s 0, s 0
1
2
We have the following simplex tableau:
x
y
s
s
b ? value
1
2
s - row
2
5
1
0
1
1
s - row
2
4
1
0
1
1
Objective
-1
-1
0
0
0
function row
Consider the negative elements in the objective function row. They are ?1, -1. The absolute value are 1,
1. There is a tie between these coefficients. To resolve the tie, we select the variable x. We take the
new basic variable as x. Consider the ratio of b-value to x-value. We have the following ratios:
1
s - row
:
1
2
326
MBA-H2040 Quantitative Techniques for Managers
1
s - row :
2
4
1 1
1
Minimum of {
,
} =
2 4
4
Hence select s as the leaving variable. Thus the pivotal element is 4. We obtain the following tableau
2
at the end of Iteration No. 1.
x
y
s
s
1
2
b ? value
s - row
9
1
1
1
0
1
2
2
2
x - row
1
1
1
1
0
4
4
4
Objective
3
1
1
0
0
function row
4
4
4
3
Now, the negative element in the objective function row is
. This corresponds to y. We take the
4
new basic variable as y. Consider the ratio of b-value to y-value. We have the following ratios:
1
2
1
s - row
:
1
9
9
2
1
x - row
: 4
1
1
4
1 1
Minimum of , 1
9 9
9
Hence select s as the leaving variable. The pivotal element is
. We get the following tableau at the
1
2
end of Iteration No. 2.
x
y
s
s
b ? value
1
2
y - row
2
1
1
0
1
9
9
9
327
MBA-H2040 Quantitative Techniques for Managers
x - row
1
5
2
1
0
18
18
9
Objective
1
1
1
0
0
function row
6
6
3
Since both x and y have been made basic variables, we have reached the stopping condition.
1
1
The optimum value of
is .
V
3
2
1
This is provided by x =
and y =
. Thus the optimum value of the game is got as V = 3.
9
9
r
s
6
2
3
1
Using the relations
,
x
y , we obtain r and s .
V
V
9
3
9
3
Problem 3:
Solve the following game by simplex method for LPP:
Player B
48
2
Player A
6
4
Solution :
Row minima :
I row
: -48
II row
: -4
Maximum of {-48, -4} = -4
Column maxima:
I column
: 6
II column
: 2
Minimum of (6, 2} = 2
So, Maximum of {Row minima} Minimum of {Column maxima}.
Therefore the given game has no saddle point. It is a mixed game. Let us convert the given game into a
LPP.
Problem formulation:
Let V denote the value of the game. Let the probability that the player B will use his first strategy be r
and second strategy be s. Let V denote the value of the game.
When A follows his first strategy:
The expected payoff to A (i.e., the expected loss to B) = - 48 r + 2 s.
This pay-off cannot exceed V. So we have
328
MBA-H2040 Quantitative Techniques for Managers
- 48 r + 2 s V
(1)'
When A follows his second strategy:
The expected pay-off to A (i.e., expected loss to B) = 6 r - 4 s.
This cannot exceed V. Hence we obtain the condition
6 r - 4 s V
(2)'
From (1)' and (2)' we have
r
s
48 2 1
V
V
and
r
s
6
4 1
V
V
r
s
Substitute
x,
y .
V
V
Thus we have
48x 2y 1
and 6x 4 y
1
where r and s are connected by the relation
r s 1.
r
s
1
i.e.,
V
V
V
1
i.e., x y
V
1
B will try to minimise V. i.e., He will try to maximise
. Thus we have the following LPP.
V
1
Maximize
x y ,
V
subject to the restrictions
48
x 2y 1,
6x 4 y 1,
x 0, y 0
Solution of LPP:
Introduce two slack variables s , s . Then the problem is transformed into the following one:
1
2
1
Maximize
x y 0.s 0.s
1
2
V
subject to the constraints
329
MBA-H2040 Quantitative Techniques for Managers
48
x 2y 1.s 0.s 1,
1
2
6x 4 y 0.s 1.s 1,
1
2
x 0, y 0, s 0, s 0
1
2
Initially, the basic variables are s , s . We have the following simplex tableau:
1
2
x
y
s
s
1
2
b ? value
s - row
1
-48
2
1
0
1
s - row
6
-4
0
1
1
2
Objective
-1
-1
0
0
0
function row
Consider the negative elements in the objective function row. They are ?1, -1. The absolute value are 1,
1. There is a tie between these coefficients. To resolve the tie, we select the variable x. We take the
new basic variable as x. Consider the ratio of b-value to x-value. We have the following ratios:
1
s - row :
1
48
1
s - row:
2
6
1
1
1
Minimum of {
,
} =
.
48
6
48
Hence select s as the leaving variable. Thus the pivotal element is - 48. We obtain the following
1
tableau at the end of Iteration No. 1.
x
y
s
s
1
2
b ? value
x - row
1
1
1
1
0
24
48
48
s - row
15
1
9
2
0
1
4
8
8
Objective
25
1
1
0
0
function row
24
48
48
330
MBA-H2040 Quantitative Techniques for Managers
25
Now, the negative element in the objective function row is
. This corresponds to y. We take the
24
new basic variable as y. Consider the ratio of b-value to y-value. We have the following ratios:
1
48
1
x - row
:
1
2
24
9
82
3
s - row
:
2
15
10
4
1
3
3
Minimum of ,
2 10
10
15
Hence select s as the leaving variable. The pivotal element is
. We get the following tableau at
2
4
the end of Iteration No. 2.
x
y
s
s
1
2
b ? value
x - row
1
1
1
1
0
480
90
30
y - row
1
4
3
0
1
30
15
10
Objective
1
5
1
0
0
function row
18
18
3
Since both x and y have been made basic variables, we have reached the stopping condition.
1
1
1
3
The optimum value of
is
. This is provided by x =
and y =
.
V
3
30
10
r
s
Thus the optimum value of the game is got as V = - 3. Using the relations
x,
y , we obtain
V
V
1
9
r
and s
.
10
10
331
MBA-H2040 Quantitative Techniques for Managers
Problem 4:
Transform the following game into an LPP:
1 8 3
6
4 5
0 1
2
Solution:
We have to determine the optimal strategy for player B. Using the entries of the given matrix, we obtain
the inequalities
r 8 s 3t V ,
6r 4 s 5t V ,
s 2t V
Dividing by V, we get
r
s
t
8
3
1,
V
V
V
r
s
t
6
4
5
1,
V
V
V
s
t
2
1
V
V
subject to the condition
r + s + t = V.
Consequently, we have
r
s
t
1
.
V
V
V
V
Substitute
r
s
t
= x,
= y,
= w.
V
V
V
Then we have the relations
x + 8 y + 3 w 1,
6 x + 4 y + 5 w 1,
y + 2 w 1.
1
r
s
t
We have to minimise V. i.e, We have to maximise
=
. i.e, We have to maximise x + y +
V
V
V
V
w.
Thus, the given game is transformed into the following equivalent LPP:
maximise x + y + w
332
MBA-H2040 Quantitative Techniques for Managers
subject to the restrictions
x + 8 y + 3 w 1,
6 x + 4 y + 5 w 1,
y + 2 w 1,
x 0, y 0, w 0.
QUESTIONS
1. Explain how a game theory problem can be solved as an LPP.
2. Transform the game
Y
Y
1
2
X a
a
1
11
12
X a
a
2
21
22
into an LPP.
3. Using simplex method for LPP, solve the following game:
6 2
3 5
1
1
Answer: r
, s
, V = 4
2
2
4. Solve the following game with LPP approach:
10
6
4 8
1
3
Answer: r
, s
, V = 7
4
4
333
MBA-H2040 Quantitative Techniques for Managers
LESSON 9
GOAL PROGRAMMING FORMULATION
LESSON OUTLINE
7.
Introduction to goal programming
8.
Formulation of goal programming problems
LEARNING OBJECTIVES
After reading this lesson you should be able to
- understand the importance of goal programming
- formulate goal programming problems
Introduction
Generally speaking, the objective of a business organization is to maximize profits and to minimize
expenditure, loss and wastage. However, a company may not always attempt at profit maximization
only. At times, a necessity may arise to pay attention to other objectives also. We describe some
such situations in the sequel.
A manufacturing organization may like to ensure uninterrupted supply of its products even if it
means additional expenditure for the procurement of raw materials or personal delivery of goods
during truckers' strike, etc. with the objective of assuring the good will of the customers.
A company may be interested in the full utilization of the capacity of the machines and therefore
mechanics may be recruited for attending to break downs of the machines even though the
occurrence of such break downs may be very rare.
A company, driven by social consciousness, may spend a portion of its profits on the maintenance of
trees, parks, public roads, etc. to ensure the safety of the environment, with the objective earning the
support of the society.
Another organization may have the objective of establishing brand name by providing high quality
products to the consumers and for this purpose it may introduce rigorous measures of quality checks
even though it may involve an increased expenditure.
While all the sales persons in a company are formally trained and highly experienced, the
management may still pursue a policy to depute them for periodical training in reputed institutes so
as to maximize their capability, without minding the extra expenditure incurred for their training.
A travel agency may be interested to ensure customer satisfaction of the highest order and as a
consequence it may come forward to operate bus services even to remote places at the normal rates,
so as to retain the customers in its fold.
A bank may offer services beyond normal working hours or on holidays even if it means payment of
overtime to the staff, in order to adhere to the policy of customer satisfaction on priority basis.
A business organization may accord priority for the welfare of the employees and so a major part of
the earnings may be apportioned on employee welfare measures.
A garment designer would like to be always known for the latest fashion and hence may spend more
money on fashion design but sell the products at the normal rates, so as to earn the maximum
reputation.
334
MBA-H2040 Quantitative Techniques for Managers
A newspaper may be interested in earning the unique distinction of `Reporter of Remote Rural
Areas' and so it may spend more money on journalists and advanced technology for communication.
The above typical instances go to show that the top level management of a business organization
may embark upon different goals in addition to profit maximization. Such goals may be necessitated
by external events or through internal discussion. At times, one such goal may be in conflict with
another goal.
`Goal programming' seeks to deal with the process of decision making in a situation of multiple
goals set forth by a business organization. A management may accord equal priority to different
goals or sometimes a hierarchy of goals may be prescribed on their importance. One has to strive to
achieve the goals in accordance with the priorities specified by the management. Sometimes the
goals may be classified as higher level goals and lower level goals as perceived by the
management and one would be interested in first achieving the higher order goals and afterwards
considering lower order goals.
Some of the goals that may be preferred by a business organization are : maximum customer
satisfaction, maximum good will of the customers, maximum utilization of the machine capacity,
maximum reliability of the products, maximum support of the society, maximum utilization of the
work force, maximum welfare of the employees, etc.
Since different goals of an organization are based on different units, the goal programming has a
multi-dimensional objective function. This is in contrast with a linear programming problem in
which the objective function is uni-dimensional.
Given a goal of an organization, one has to determine the conditions under which there will be
under-achievement and over-achievement of the goal. The ideal situation will be the one with neither
under-achievement nor over-achievement of the goal.
Formulation of Goal Programming Problems
In the sequel, we consider illustrative situations so as to explain the process of problem formulation
in goal programming.
Notations
If there is a single goal, we have the following notations:
Let Du denote the under-achievement of the goal.
Let Do denote the over-achievement of the goal.
If there are two goals, we have the following notations:
Denote the under-achievement and the over-achievement of one goal by Du1 and Do1
respectively.
Denote the under-achievement and the over-achievement of another goal by Du2 and Do2
respectively.
PROBLEM 1:
Alpha company is known for the manufacture of tables and chairs. There is a profit of Rs. 200 per table
and Rs. 80 per chair. Production of a table requires 5 hours of assembly and 3 hours in finishing. In
order to produce a chair, the requirements are 3 hours of assembly and 2 hours of finishing. The
company has 105 hours of assembly time and 65 hours of finishing. The company manager is
335
MBA-H2040 Quantitative Techniques for Managers
interested to find out the optimal production of tables and chairs so as to have a maximum profit of Rs.
4000. Formulate a goal programming problem for this situation.
Solution:
The manager is interested not only in the maximization of profit but he has also fixed a target of Rs.
4000 as profit. Thus, the problem involves a single goal of achieving the specified amount of profit.
Let D denote the under achievement of the target profit and let D be the over achievement.
u
o
The objective in the given situation is to minimize under achievement. Let Z be the objective
function. Then the problem is the minimization of Z D .
u
Formulation of the constraints:
Let the number of tables to be produced be x and let the number of chairs to be produced be Y.
Profit from x tables = Rs. 200 x
Profit from y chairs = Rs. 80 y
The total profit =
Profit from x tables and y chairs
+ under achievement of the profit target
? over achievement of the profit target
So we have the relationship 200 x + 80 y + D - D = 4000.
u
o
Assembly time:
To produce x tables, the requirement of assembly time = 5 x hours. To produce y
chairs, the requirement is 3 y hours. So, the total requirement is 5 x 3 y hours. But the available time
for assembly is 105 hours. Therefore constraint
5 x + 3 y 105
must be fulfilled.
Finishing time:
To produce x tables, the requirement of finishing time = 3 x. To produce y chairs, the requirement is 2 y .
So, the total requirement is 3 x+ 2 y. But the availability is 65 hours. Hence we have the restriction
3x 2 y 65 .
Non-negativity restrictions:
The number of tables and chairs produced, the under achievement of the profit target and the
over achievement cannot be negative. Thus we have the restrictions
x 0, y 0, D 0, D 0 .
u
o
336
MBA-H2040 Quantitative Techniques for Managers
Statement of the problem:
Minimize Z D
u
subject to the constraints
200 x 80 y D D 4000
u
o
5 x 3 y 105
3 x 2 y 65
x 0, y 0, D 0, D 0
u
o
Problem 2:
Sweet Bakery Ltd. produces two recipes A and B. Both recipes are made of two food stuffs I and II.
Production of one Kg of A requires 7 units of food stuff I and 4 units of food stuff II whereas for
producing one Kg of B, 4 units of food stuff I and 3 units of food stuff II are required. The company
has 145 units of food stuff I and 90 units of food stuff II. The profit per Kg of A is Rs. 120 while that of
B is Rs. 90. The manager wants to earn a maximum profit of Rs. 2700 and to fulfil the demand of 12
Kgs of A. Formulate a goal programming problem for this situation.
Solution:
The management has two goals.
1. To reach a profit of Rs. 2700
2. Production of 12 Kgs of recipe A.
Let D denote the under achievement of the profit target.
up
Let D denote the over achievement of the profit target.
op
Let D denote the under achievement of the production target for recipe A.
uA
Let D denote the over achievement of the production target for recipe A.
oA
The objective in this problem is to minimize the under achievement of the profit target and to
minimize the under achievement of the production target for recipe A.
Let Z be the objective function. Then the problem is the minimization of
Z D D .
up
uA
Constraints
Suppose the company has to produce x kgs of recipe A and y kgs of recipe B in order to achieve the two
goals.
337
MBA-H2040 Quantitative Techniques for Managers
Condition on profit:
Profit from x kgs of A = 120 x
Profit from y kgs of B = 90 y
The total profit
=
Profit from x kgs of A + Profit from y kgs of B
+ under achievement of the profit target
? over achievement of the profit target
=
120 x 90 y D D
up
op
Thus we have the restriction
120x 90 y D D 2700 .
up
op
Constraint for food stuff I:
7 x 4 y 145
Constraint for food stuff II:
4 x 3 y 90
Non-negativity restrictions:
x, y, D , D , D , D 0
up
op
uA
oA
Condition on recipe A:
The target production of A =
optimal production of A
+ under achievement in production target of A
? over achievement of the production target of A.
Thus we have the condition
x D D
12
uA
oA
Statement of the problem:
Minimize Z D D
up
uA
subject to the constraints
120 x 90 y D D 2700
up
op
x D D
12
uA
oA
7 x 4 y 145
4 x 3 y 90
x, y, D , D , D , D 0
up
op
uA
oA
QUESTIONS
1. Explain the necessity of a goal programming.
338
MBA-H2040 Quantitative Techniques for Managers
2. Describe some instances of goal programming.
3. Explain the formulation of a goal programming problem.
339
MBA-H2040 Quantitative Techniques for Managers
UNIT ? V
Lesson Outline
Introduction
Terminologies of Queueing System
Empirical Queueing Models
Simulation ? Introduction
Types of Simulation
Major Steps of Simulation
Replacement and Maintenance Analysis ? Introduction
Types of Maintenance
Types of Replacement Problem
Determination of Economic life of an Asset
Learning Objectives
After reading this lesson you should be able to
Understand the nature and scope of Queneing System.
Queueing models and the solution to queueing model problems.
Importance of Simulation
Need for Replacement and maintenance
Solution to problems involving economic life of an Asset.
340
MBA-H2040 Quantitative Techniques for Managers
5-1 ? Queueing Theory
5.1.1 Introduction:
A flow of customers from finite or infinite population towards the service facility forms a queue (waiting line) an account of
lack of capability to serve them all at a time. In the absence of a perfect balance between the service facilities and the
customers, waiting time is required either for the service facilities or for the customers arrival. In general, the queueing
system consists of one or more queues and one or more servers and operates under a set of procedures. Depending upon the
server status, the incoming customer either waits at the queue or gets the turn to be served. If the server is free at the time of
arrival of a customer, the customer can directly enter into the counter for getting service and then leave the system. In this
process, over a period of time, the system may experience " Customer waiting" and /or "Server idle time"
5.1.2 Queueing System:
A queueing system can be completely described by
(1) the input (arrival pattern)
(2) the service mechanism (service pattern)
(3) The queue discipline and
(4) Customer's behaviour
5.1.3. The input (arrival pattern)
The input described the way in which the customers arrive and join the system. Generally, customers arrive in a more or less
random manner which is not possible for prediction. Thus the arrival pattern can be described in terms of probabilities and
consequently the probability distribution for inter-arrival times (the time between two successive arrivals) must be defined.
We deal with those Queueing system in which the customers arrive in poisson process. The mean arrival rate is denoted by
.
341
MBA-H2040 Quantitative Techniques for Managers
5.1.4 The Service Mechanism:-
This means the arrangement of service facility to serve customers. If there is infinite number of servers, then all the
customers are served instantaneously or arrival and there will be no queue. If the number of servers is finite then the
customers are served according to a specific order with service time a constant or a random variable. Distribution of service
time follows `Exponential distribution' defined by
f(t) = e -t , t > 0
The mean Service rate is E(t) = 1/
5.1.5 Queueing Discipline:-
It is a rule according to which the customers are selected for service when a queue has been formed. The most common
disciplines are
1. First come first served ? (FCFS)
2. First in first out ? (FIFO)
3. Last in first out ? (LIFO)
4. Selection for service in random order (SIRO)
5.1.6 Customer's behaviour
1. Generally, it is assumed that the customers arrive into the system one by one. But in some cases, customers
may arrive in groups. Such arrival is called Bulk arrival.
2. If there is more than one queue, the customers from one queue may be tempted to join another queue because of
its smaller size. This behaviour of customers is known as jockeying.
3. If the queue length appears very large to a customer, he/she may not join the queue. This property is known as
Balking of customers.
4. Sometimes, a customer who is already in a queue will leave the queue in anticipation of longer waiting line.
This kind of departare is known as reneging.
5.1.7 List of Variables
The list of variable used in queueing models is give below:
n - No of customers in the system
C - No of servers in the system
Pn (t) ? Probability of having n customers in the system at time t.
Pn - Steady state probability of having customers in the
system
P0 - Probability of having zero customer in the system
Lq - Average number of customers waiting in the queue.
Ls - Average number of customers waiting in the system
(in the queue and in the service counters)
Wq - Average waiting time of customers in the queue.
Ws - Average waiting time of customers in the system
(in the queue and in the service counters)
- Arrival rate of customers
- Service rate of server
- Utilization factor of the server
eff - Effective rate of arrival of customers
M - Poisson distribution
N - Maximum numbers of customers permitted in the system. Also, it denotes the size of the calling source of the
customers.
GD - General discipline for service. This may be first in first ? serve (FIFS), last-in-first serve (LIFS) random order
(Ro) etc.
342
MBA-H2040 Quantitative Techniques for Managers
5.1.8 Traffic intensity (or utilization factor)
An important measure of a simple queue is its traffic intensity given by
Traffic intensity = Mean arrival time
=
(< 1)
Mean service time
and the unit of traffic intensity is Erlang
5.1.9 Classification of Queueing models
Generally, queueing models can be classified into six categories using Kendall's notation with six parameters to define a
model. The parameters of this notation are
P- Arrival rate distribution ie probability law for the arrival /inter ? arrival time.
Q - Service rate distribution, ie probability law according to
which the customers are being served.
R - Number of Servers (ie number of service stations)
X - Service discipline
Y - Maximum number of customers permitted in the system.
Z - Size of the calling source of the customers.
A queuing model with the above parameters is written as
(P/Q/R : X/Y/Z)
5.1.10 Model 1 : (M/M/1) : (GD/ / ) Model
In this model
(i)
the arrival rate follows poisson (M) distribution.
(ii)
Service rate follows poisson distribution (M)
(iii)
Number of servers is 1
(iv)
Service discipline is general disciple (ie GD)
(v)
Maximum number of customers permitted in the system is infinite ()
(vi)
Size of the calling source is infinite ()
The steady state equations to obtain, Pn the probability of having customers in the system and the values for Ls, Lq, Ws and
Wq are given below.
n= 0,1,2,---- where = <1
Ls ? Average number of customers waiting in the system
(ie waiting in the queue and in the service station)
P
n
n =
(1-)
Ls =
1-
Lq
=
Ls ?
=
-
1 -
343
MBA-H2040 Quantitative Techniques for Managers
=
- (1 - )
1 -
Lq
=
2
Average
waiting
time of
1 -c
u stomers in the system
(in the que
ue an
d in the service station)
= Ws
= Ws
= Ls =
(1 - )
=
x
1
=
1
1 -
(1 - )
(Since = )
= 1
-
=
1
-
Ws = 1
-
Wq = Average waiting time of customers in the
queue.
=
Lq / = [1 / ] [ 2 / [1- ]]
= 1 / [ 2 / [1- ]]
=
Since =
-
Wq =
-
Example 1:
The arrival rate of customers at a banking counter follows a poisson distibution with a mean of 30 per hours. The service rate
of the counter clerk also follows poisson distribution with mean of 45 per hour.
a) What is the probability of having zero customer in the system ?
b) What is the probability of having 8 customer in the system ?
c) What is the probability of having 12 customer in the system ?
d) Find Ls, Lq, Ws and Wq
Solution
Given arrival rate follows poisson distribution with
mean =30
= 30 per hour
Given service rate follows poisson distribution with
mean = 45
= 45 Per hour
Utilization factor = /
= 30/45
= 2/3
= 0.67
a) The probability of having zero customer in the system
P0
= 0 (1- )
= 1-
= 1-0.67
344
MBA-H2040 Quantitative Techniques for Managers
= 0.33
b) The probability of having 8 customers in the system
P8
= 8 (1- )
= (0.67)8 (1-0.67)
= 0.0406 x 0.33
= 0.0134
Probability of having 12 customers in the system is
P12
= 12 (1- )
= (0.67)12 (1-0.67)
= 0.0082 x 0.33 = 0.002706
Ls
=
=
0.67
1 -
1-0.67
=
0.67
= 2.03
0.33
= 2 customers
Lq
= 2
=
(0.67)2
=
0.4489
1-
1-0.67
0.33
=
1.36
=
1 Customer
Ws
=
1
=
1
=
1
-
45-30
15
=
0.0666 hour
Wq
=
=
0.67
=
0.67
-
45-30
15
=
0.4467 hour
Example 2 :
At one-man barbar shop, customers arrive according to poisson dist with mean arrival rate of 5 per hour
and the hair cutting time was exponentially distributed with an average hair cut taking 10 minutes. It is
assumed that because of his excellent reputation, customers were always willing to wait. Calculate the
following:
(i)
Average number of customers in the shop and the average numbers waiting for a haircut.
(ii)
The percentage of time arrival can walk in straight without having to wait.
(iii)
The percentage of customers who have to wait before getting into the barber's chair.
Solution:-
Given mean arrival of customer = 5/60 =1/12
and mean time for server
=
1/10
= /
= [1/12] x 10
=
10 /12
=
0.833
(i)
Average number of customers in the system (numbers in the queue and in the service station)
Ls
= / 1-
= 0.83 / 1- 0.83
= 0.83 / 0.17
= 4.88
= 5 Customers
(ii)
The percentage of time arrival can walk straight into barber's chair without waiting is
Service utilization
= %
= / %
= 0.833 x 100
=83.3
345
MBA-H2040 Quantitative Techniques for Managers
(iii)
The percentage of customers who have to wait before getting into the barber's chair = (1-)%
(1-0.833)%
=
0.167 x 100
=
16.7%
Example 3 :
Vehicles are passing through a toll gate at the rate of 70 per hour. The average time to pass through the
gate is 45 seconds. The arrival rate and service rate follow poisson distibution. There is a complaint
that the vehicles wait for a long duration. The authorities are willing to install one more gate to reduce
the average time to pass through the toll gate to 35 seconds if the idle time of the toll gate is less than
9% and the average queue length at the gate is more than 8 vehicle, check whether the installation of the
second gate is justified?
Solutions:-
Arrival rate of vehicles at the toll gate
=
70 per hour
Time taken to pass through the gate =
45 Seconds
Service rate
= 1 hours
45 seconds
= 3600/45
= 80
= 80 Vehicles per hour
Utilization factor = /
= 70 / 80
= 0.875
(a) Waiting no. of vehicles in the queue is Lq
Lq
= 2 / 1 -
=
(0.875)2
1-0.875
=
0.7656
0.125
=
6.125
= 6 Vehicles
(b) Revised time taken to pass through the gate
=30 seconds
The new service rate after installation of an
additional gate = 1 hour/35 Seconds
= 3600/35
=
102.68 Vehicles / hour
Utilization factor = / = 70 / 102.86
= 0.681
Percentage of idle time of the gate = (1-)%
= (1-0.681)%
= 0.319%
= 31.9
= 32%
This idle time is not less than 9% which is expected.
Therefore
the installation of the second gate is not justified since the average waiting number of vehicles in
the queue is more than 8 but the idle time is not less than 32%. Hence idle time is far greater than the number. of vehicles
waiting in the queue.
5.1.11 Second Model (M/M/C) : (GD/ / )Model
The parameters of this model are as follows:
(i)
Arrival rate follows poisson distribution
(ii)
Service rate follows poisson distribution
(iii)
No of servers is C'.
(iv)
Service discipline is general discipline.
346
MBA-H2040 Quantitative Techniques for Managers
(v)
Maximum number of customers permitted in the system is infinite
Then the steady state equation to obtain the probability of having n customers in the system is
Pn
= n Po
,
o n C
n!
= n Po
for n > c Where / c < 1
C n-c C!
Where [ / c] < 1 as = /
C-1
P0
={[ n/n!] + c / (c! [1 - /c])}-1
n = 0
where c! = 1 x 2 x 3 x ................. upto C
Lq
= [ c+1 / [c-1! (c - )] ] x P0
= (c Pc) / (c - )2
Ls
= Lq + and Ws = Wq + 1 /
Wq
= Lq /
Under special conditions Po
= 1 - and Lq = C+! / c 2 Where <1 and
Po = (C-) (c ? 1)! / c c
and L
q =
/ (c- ), where / c < 1
Example 1:
At a central warehouse, vehicles are at the rate of 24 per hour and the arrival rate follows poisson distribution. The unloading
time of the vehicles follows exponential distribution and the unloading rate is 18 vehicles per hour. There are 4 unloading
crews. Find
(i)
Po and P3
(ii)
Lq, Ls, Wq and Ws
Solution:
Arrival rate = 24 per hour
Unloading rate = 18 Per hour
No. of unloading crews C=4
= /
= 24 / 18=1.33
C-1
(i) P0
={[ n/n!] + c / (c! [1 - /c])}-1
n = 0
3
={[ (1.33)n/n!]+ (1.33)4 /(4! [1 - (1.33)/ 4])}-1
n = 0
={ (1.33)0 / = 0! + (1.33)1 / 1! + (1.33)2 / 2! + (1.33)3 / 3! +
(1.33)4 / 24! [1 - (1.33)/ 4] }-1
=[1 + 1.33 + 0.88 + 0.39 + 3.129/16.62] -1
=[3.60 + 0.19]-1 =
[3.79]-1
= 0.264
We know Pn
= ( n / n!) Po
for
0 n c
P3
= ( 3 / 3!) Po
Since 0 3 4
=
[(1.33)3 / 6 ] x 0.264
=
2.353 x 0.044
=
0.1035
(ii) Lq
=
C+1 X P0
(C ? 1)! (C-)2
=
(1.33)5
X
0.264
3! X (4 ? 1.33)2
=
(4.1616) X
0.264
6 X (2.77)2
347
MBA-H2040 Quantitative Techniques for Managers
=
(4.1616) X
0.264
46 .0374
=
1.099 / 46.0374
=
0.0239
=
0.0239 Vehicles
Ls
=
Lq +
=
0.0239 + 1.33
=
1.3539 Vehicles
Wq
=
Lq /
=
0.0239 /24
=
0.000996 hrs
Ws
=
Wq + 1 /
=
0.000996 + 1/18
=
0.000996 + 0.055555
=
0.056551 hours.
Example 2 :-
A supermarket has two girls ringing up sales at the counters. If the service time for each customer is exponential with mean 4
minutes and if the people arrive in poisson fashion at the rate of 10 per hour
a) What is the probability of having to wait for service?
b) What is the expected percentage of idle time for each girl?
c) If a customer has to wait, what is the expected length of his waiting time?
Solution:-
C-1
P0
={[ n/n!] + c / (c! [1 - /c])}-1
n = 0
Where = / given arrival rate = 10 per hour
= 10 / 60 = 1 / 6 per minute
Service rate = 4 minutes
= 1 / 4
person per minute
Hence = /
= (1 / 6) x 4
= 2 / 3
= 0.67
1
P0 ={[ n/n!]+(0.67)2 / (2! [1 - 0.67/2])}-1
n = 0
=[1 + ( / 1!) ] + 0.4489 / (2 ? 0.67)]-1
=[1 + 0.67 + 0.4489 / (1.33)]-1
=[1 + 0.67 + 0.34]-1
=[ 2.01]-1
= 1 / 2
The Probability of having to wait for the service is
P (w > 0)
=
c
X P0
c! [1 - /c]
=
0.67 2 X (1 / 2)
2! [1 ? 0.67 /2]
=
0.4489 / 2.66
=
0.168
b) The probability of idle time for each girl is
= 1- P (w > 0)
348
MBA-H2040 Quantitative Techniques for Managers
= 1-1/3
= 2/3
Percentage of time the service remains idle = 67% approximately
c) The expected length of waiting time (w/w>0)
=
1 / (c - )
=
1 / [(1 / 2) ? (1 / 6) ]
=
3 minutes
Examples 3 :
A petrol station has two pumps. The service time follows the exponential distribution with mean 4 minutes and cars arrive
for service in a poisson process at the rate of 10 cars per hour. Find the probability that a customer has to wait for service.
What proportion of time the pump remains idle?
Solution:
Given C=2
The arrival rate = 10 cars per hour.
= 10 / 60
= 1 / 6 car per minute
Service rate = 4 minute per cars.
Ie
= ? car per minute.
= /
= (1/6) / (1/4)
= 2 / 3
= 0.67
Proportion of time the pumps remain busy
= / c
= 0.67 / 2
= 0.33
= 1 / 3
The proportion of time, the pumps remain idle
=1 ? proportion of the pumps remain busy
=
1-1 / 3
= 2 / 3
C-1
P0
={[ n/n!] + c / (c! [1 - /c])}-1
n = 0
=[ ( 0.67)0 / 0!) + ( 0.67)1 / 1!) + ( 0.67)2 / 2!)[1- ( 0.67 / 2)1 ]-1
=[1 + 0.67 + 0.4489 / (1.33)]-1
=[1 + 0.67 + 0.33]-1
=[ 2]-1
=1 / 2
Probability that a customer has to wait for service
=
p [w>0]
=
c
x P0 =
(0.67)2 x 1/2
[c [1 - / c]
[2![1 ? 0.67/2]
=
0.4489
=
0.4489
1.33x2
2.66
=
0.1688
5.2 Simulation :
Simulation is an experiment conducted on a model of some system to collect necessary information on the behaviour of that
system.
5.2.1 Introduction :
The representation of reality in some physical form or in some form of Mathematical equations
are called Simulations .
Simulations are imitation of reality.
For example :
349
MBA-H2040 Quantitative Techniques for Managers
1. Children cycling park with various signals and crossing is a simulation of a read model traffic system
2. Planetarium
3. Testing an air craft model in a wind tunnel.
5.2.2. Need for simulation :
Consider an example of the queueing system, namely the reservation system of a transport corporation.
The elements of the system are booking counters (servers) and waiting customers (queue). Generally
the arrival rate of customers follow a Poisson distribution and the service time follows exponential
distribution. Then the queueing model (M/M/1) : (GD/ / ) can be used to find the standard results.
But in reality, the following combinations of distributions my exist.
1. Arrived rate does not follow Poisson distribution, but the service rate follows an exponential distribution.
2. Arrival rate follows a Poisson distribution and the service rate does not follow exponential distribution.
3. Arrival rate does not follows poisson distribution and the service time also does not follow exponential distribution.
In each of the above cases, the standard model (M/M/1) : (G/D/ / ) cannot be used. The last resort to find the
solution for such a queueing problem is to use simulation.
5.2.3. Some advantage of simulation :
1. Simulation is Mathematically less complicated
2. Simulation is flexible
3. It can be modified to suit the changing environments.
4. It can be used for training purpose
5. It may be less expensive and less time consuming in a quite a few real world situations.
5.2.4. Some Limitations of Simulation :
1. Quantification or Enlarging of the variables maybe difficult.
2. Large number of variables make simulations unwieldy and more difficult.
3. Simulation may not. Yield optimum or accurate results.
4. Simulation are most expensive and time consuming model.
5. We cannot relay too much on the results obtained from simulation models.
5.2.5. Steps in simulation :
1. Identify the measure of effectiveness.
2. Decide the variables which influence the measure of effectiveness and choose those variables, which affects the
measure of effectiveness significantly.
3. Determine the probability distribution for each variable in step 2 and construct the cumulative probability
distribution.
4. Choose an appropriate set of random numbers.
5. Consider each random number as decimal value of the cumulative probability distribution.
6. Use the simulated values so generated into the formula derived from the measure of effectiveness.
7. Repeat steps 5 and 6 until the sample is large enough to arrive at a satisfactory and reliable decision.
5.2.6. Uses of Simulation
Simulation is used for solving
1.Inventory Problem
2. Queueing Problem
3. Training Programmes etc.
Example :
350
MBA-H2040 Quantitative Techniques for Managers
Customers arrive at a milk booth for the required service. Assume that inter ? arrival and service time
are constants and given by 1.5 and 4 minutes respectively. Simulate the system by hand computations
for 14 minutes.
(i) What is the waiting time per customer?
(ii) What is the percentage idle time for the facility?
(Assume that the system starts at t = 0)
Solution :
First customer arrives at the service center at t = 0
His departure time after getting service = 0 + 4 = 4 minutes.
Second customer arrives at time t = 1.5 minutes
he has to wait = 4 ? 1.5 = 2.5 minutes.
Third customer arrives at time t = 3 minutes
he has to wait for = 8-3 = 5 minutes
Fourth customer arrives at time t = 4.5 minutes and he has to wait for 12 ? 4.5 = 7.5 minutes.
During this 4.5 minutes, the first customer leaves in 4 minutes after getting service and the second customer is getting
service.
Fifth customer arrives at t = 6 minutes
he has to wait 14 ? 6 = 8 minutes
Sixth customer arrives at t = 7.5 minutes
he has to wait 14 ? 7.5 = 6.5 minutes
Seventh customer arrives at t = 9 minutes
he has to wait 14 ? 9 = 5 minutes
During this 9 minutes the second customer leaves the service in 8th minute and third person is to get service in 9th minute.
Eighth customer arrives at t = 10.5 minutes
he has to wait 14 ? 10.5 = 3.5 minutes
Nineth customer arrives at t = 12 minutes
he has to wait 14 ? 12 = 2 minutes
But by 12th minute the third customer leaves the Service
10th Customer arrives at t = 13.5 minutes
he has to wait 14-13.5 = 0.5 minute
From this simulation table it is clear that
(i) Average waiting time for 10 customers
= 2.5+5+7.5+8+6.5+5.0+3.5+2+0.5
10
= 40.5 = 4.05
10
(ii) Average waiting time for 9 customers who are in waiting for service
40.5
= 4.5 minutes.
9
But the average service time is 4 minutes which is less that the average waiting time, the percentage of idle time for service =
0%
Exercise :
1. The arrival rate of customers at a banking counter follows a poisson distribution with a mean of 45
per hors. The service rate of the counter clerk also poisson distribution with a mean of 60 per hours.
(a) What is the probability of having Zero customer in the system (P0).
(b) What is the probability of having 5 customer in the system (P5).
(c) What is the probability of having 10 customer in the system (P10).
(d) Find Ls, Lq, Ws and Wq
351
MBA-H2040 Quantitative Techniques for Managers
2. Vehicles pass through a toll gate at a rate of 90 per hour. The average time to pass through the gate is
36 seconds. The arrival rate and service rate follow poisson distribution. There is a complaint the
vehicles wait for long duration. The authorities are willing to install one more gate to reduce the
average time to pass through the toll gate to 30 seconds if the idle time of the toll gate is less than 10%
and the average queue length at the gate is more than 5 vehicles. Vehicle whether the installation of
second gate is justified?
3. At a central ware house, vehicles arrive at the rate of 24 per hours and the arrival rate follows poisson
distribution. The unloading time of the vehicles follows exponentional distribution and the unloading
rate is 18 vehicles per hour. There are 4 unloading crews. Find the following.
a) P0 and P3
b) Lq, Ls, Wq and Ws
4.
Explain Queneing Discipline
5.
Describe the Queueing models (M/M/1) : (GD/ /)
and (M/M/C) : (GD/ / )
6.
Cars arrive at a drive-in restaurant with mean arrival rate of 30 cars per hors and the service rate
of the cars is 22 per hors. The arrival rate and the service rate follow poisson distribution. The number
parking space for cars is only 5. Find the standard results.
(Ans Lq = 2.38 cars, Ls = 3.3133 Cars, Wq = 0.116 hors and
Ws = 0.1615 hors)
7.
In a harbour, ship arrive with a mean rate of 24 per week. The harbour has 3 docks to handle
unloading and loading of ships. The service rate of individual dock is 12 per week. The arrival rate and
the service rate follow poisson distribution. At any point of time, the maximum No. of ships permitted
in the harbour is 8. Find P0, Lq, Ls, Wq, Ws
(Ans P0 = 0.1998, Lq = 1.0371 ships, Ls = 2.9671 ships,
Wq = 0.04478 week and Ws = 0.1281 week)
8.
Define simulation and its advantages.
9.
Discuss the steps of simulation.
5.3. Replacement models
5.3.1. Introduction:
The replacement problems are concerned with the situations that arise when some items such as men,
machines and usable things etc need replacement due to their decreased efficiency, failure or
breakdown. Such decreased efficiency or complete breakdown may either be gradual or all of a sudden.
If a firm wants to survive the competition it has to decide on whether to replace the out dated
equipment or to retain it, by taking the cost of maintenance and operation into account. There are two
basic reasons for considering the replacement of an equipment.
They are
352
MBA-H2040 Quantitative Techniques for Managers
(i)
Physical impairment or malfunctioning of various parts.
(ii)
Obsolescence of the equipment.
The physical impairment refers only to changes in the physical condition of the equipment itself.
This will lead to decline in the value of service rendered by the equipment, increased operating cost of
the equipments, increased maintenance cost of the equipment or the combination of these costs.
Obsolescence is caused due to improvement in the existing Tools and machinery mainly when the
technology becomes advanced therefore, it becomes uneconomical to continue production with the same
equipment under any of the above situations. Hence the equipments are to be periodically replaced.
Some times, the capacity of existing facilities may be in adequate to meet the current demand. Under
such cases, the following two alternatives will be considered.
1. Replacement of the existing equipment with a new one
2. Argument the existing one with an additional equipments.
5.3.2 Type of Maintenance
Maintenance activity can be classified into two types
i)
Preventive Maintenance
ii)
Breakdown Maintenance
Preventive maintenance (PN) is the periodical inspection and service which are aimed to detect
potential failures and perform minor adjustments a requires which will prevent major operating problem
in future. Breakdown maintenance is the repair which is generally done after the equipment breaks
down. It is offer an emergency which will have an associated penalty in terms of increasing the cost of
maintenance and downtime cost of equipment, Preventive maintenance will reduce such costs up-to a
certain extent . Beyond that the cost of preventive maintenance will be more when compared to the cost
of the breakdown maintenance.
Total cost = Preventive maintenance cost + Breakdown maintenance cost.
This total cost will go on decreasing up-to P with an increase in the level of maintenance up-to a
point, beyond which the total cost will start increasing from P. The level of maintenance corresponding
to the minimum total cost at P is the Optional level of maintenance this concept is illustrated in the
follows diagram
353
MBA-H2040 Quantitative Techniques for Managers
N
M
The points M and N denote optimal level of maintenance and optimal cost respectively
5.3.3 Types of replacement problem
The replacement problem can be classified into two categories.
i)
Replacement of assets that deteriorate with time (replacement due to gradual failure, due to wear
and tear of the components of the machines) This can be further classified into the following
types.
a)
Determination of economic type of an asset.
b)
Replacement of an existing asset with a new asset.
ii)
simple probabilistic model for assets which will fail completely (replacement due to sudden
failure).
5.3.4. Determination of Economic Life of an asset
Any asset will have the following cost components
i)
Capital recovery cost (average first cost), Computed form the first cost (Purchase price) of the
asset.
ii)
Average operating and maintenance cost.
iii)
Total cost which is the sum of capital recovery cost (average first cost) and average operating
and maintenance cost.
A typical shape of each of the above cost with respect to
life of the
asset is shown below
354
MBA-H2040 Quantitative Techniques for Managers
From figure, when the life of the machine increases, it is clear that the capital recovery cost
(average first cost) goes on decreasing and the average operating and maintenance cost goes on
increasing. From the beginning the total cost goes on decreasing upto a particular life of the asset and
then it starts increasing. The point P were the total cost in the minimum is called the Economic life of
the asset. To solve problems under replacement, we consider the basics of interest formula.
Present worth factor denoted by (P/F, i,n). If an amount P is invested now with amount earning
interest at the rate i per year, then the future sum (F) accumulated after n years can be obtained.
P
-
Principal sum at year Zero
F
-
Future sum of P at the end of the nth year
i
-
Annual interest rate
n
-
Number of interest periods.
Then the formula for future sum F = P ( 1 + i ) n
P = F/(1 +i)n = Fx (present worth factor)
If A is the annual equivalent amount which occurs at the end of every year from year one through n
years is given by
A
=
P x i (1 +i)n
(1 +i)n - 1
=
P ( A / P, i, n )
=
P x equal payment series capital recovery factor
Example:
A firm is considering replacement of an equipment whose first cost is Rs. 1750 and the scrap value is
negligible at any year. Based on experience, it is found that maintenance cost is zero during the first
year and it increases by Rs. 100 every year thereafter.
(i) When should be the equipment replaced if
a) i = 0%
b) i = 12%
Solution :
Given the first cost = Rs 1750 and the maintenance cost is Rs. Zero during the first years and
then increases by Rs. 100 every year thereafter. Then the following table shows the calculation.
Calculations to determine Economic life
(a) First cost Rs. 1750
Interest rate = 0%
Summation Average cost Average first Average
Maintenan
End
of
of
of
cost
if total
cost
ce cost at
year (n)
maintenanc
maintenance
replaced
at through the
end of year
e
through the the
given given year
355
MBA-H2040 Quantitative Techniques for Managers
Cost
given year
year and
A
B (Rs)
C (Rs)
D (in Rs)
E (Rs)
F (Rs)
1750
C = B
C/A
D + E
A
1
0
0
0
1750
1750
2
100
100
50
875
925
3
200
300
100
583
683
4
300
600
150
438
588
5
400
1000
200
350
550
6
500
1500
250
292
542
7
600
2100
300
250
550
8
700
2800
350
219
569
The value corresponding to any end-of-year (n) in Column F represents the average total cost of
using the equipment till the end of that particulars year.
In this problem, the average total cost decreases till the end of the year 6 and then it increases.
Hence the optimal replacement period is 6 years ie the economic life of the equipment is 6 years.
(e) When interest rate i = 12%
When the interest rate is more than 0% the steps to get the economic life are summarized in the
following table.
Calculation to determine Economic life
First Cost = Rs. 1750
Interest rate = 12%
Mai
Present
Summation
nten
worth
as of present
(A/P,
En
Present
ance
beginning
worth
of
12%,n)
Annual
d
simulator
cost
of years
maintenanc
=i (1+i)n
equipment
of
maintena
at
(P/F,12v,n)
1
of e
costs
(1+i)n-1
total
cost
ye
nce
cost
end
maintenanc
through the
through the
ar
and first
of
e costs
given year
giver year
(n)
cost
year
G
s
A
B
C
D
E
F
G
H
C =
B
1
E+
BxC
D
FxG
(iR)
(1+12/100) n
Rs. 1750
1
0
0.8929
0
0
1750
1.1200
1960
2
100
0.7972
79.72
79.72
1829.72
0.5917
1082.6
3
200
0.7118
142.36
222.08
1972.08
0.4163
820.9
4
300
0.6355
190.65
412.73
2162.73
0.3292
711.9
5
400
0.5674
226.96
639.69
2389.69
0.2774
662.9
6
500
0.5066
253.30
892.99
2642.99
0.2432
642.7
7
600
0.4524
271.44
1164.43
2914.430
0.2191
638.5
8
700
0.4039
282.73
1447.16
3197.16
0.2013
680.7
356
MBA-H2040 Quantitative Techniques for Managers
Identify the end of year for which the annual equivalent total cost is minimum in column. In this
problem the annual equivalent total cost is minimum at the end of year hence the economics life of the
equipment is 7 years.
5.3.5.
Simple probabilistic model for items which completely fail
Electronic items like bulbs, resistors, tube lights etc. generally fail all of a sudden, instead of gradual
failure. The sudden failure of the item results in complete breakdown of the system. The system may
contain a collection of such items or just an item like a single tube-light. Hence we use some
replacement policy for such items which would minimize the possibility of complete breakdown. The
following are the replacement policies which are applicable in these cases.
i) Individual replacement policy :
Under this policy, each item is replaced immediately after failure.
ii) Group replacement policy :
Under group replacement policy, a decision is made with regard the replacement at what equal internals,
all the item are to be replaced simultaneously with a provision to replace the items individually which
fail during the fixed group replacement period.
Among the two types of replacement polices, we have to decide which replacement policy we
have to follow. Whether individual replacement policy is better than group replacement policy. With
regard to economic point of view. To decide this, each of the replacement policy is calculated and the
most economic one is selected for implementation.
Exercise :
1. List and explain different types of maintenance
2. Discuss the reasons for maintenance.
3. Distinguish between breakdown maintenance and preventive maintenance.
4. Distinguish between individual and group replacement polices.
5. A firm is considering replacement of an equipment whose first cost is Rs.4000 and the scrap
value is negligible at the end of any year. Based on experience, it has been found that the
maintenance cost is zero during the first year and it is Rs.1000 for the second year. It increase by
Rs.300 every years thereafter.
a) When should the equipment be replace if i = 0%
b) When should the equipment be replace if i = 12%
Ans . a) 5 years
b) 5 years
6. A company is planning to replace an equipment whose first cost is Rs.1,00,000. The operating
and maintenance cost of the equipment during its first year of operation is Rs.10,000 and it
increases by Rs. 2,000 every year thereafter. The release value of the equipment at the end of the
357
MBA-H2040 Quantitative Techniques for Managers
first year of its operation is Rs.65,000 and it decreases by Rs.10,000 every year thereafter. Find
the economic life of the equipment by assuming the interest rate as 12%.
[Ans : Economic life = 13 years and the corresponding annual equivalent cost = Rs. 34,510]
7. The following table gives the operation cost, maintenance cost and salvage value at the end of
every year of machine whose purchase value is Rs. 12,000. Find the economic life of the
machine assuming.
a) The interest rate as 0%
b) The interest rate as 15%
Operation cost at Maintenance cost Salvage value at the end
End
of the end of year at the end of year of year (Rs)
year
(Rs)
(Rs)
1
2000
2500
8000
2
3000
3000
7000
3
4000
3500
6000
4
5000
4000
5000
5
6000
4500
4000
6
7000
5000
3000
7
8000
5500
2000
8
9000
6000
1000
Ans :
a) Economic life of the machine = 2 years
b) Economic life of the machine = 2 years
358
MBA-H2040 Quantitative Techniques for Managers
359
This post was last modified on 14 March 2022