Table Of Content

Chapter One ..........................................................................................................................1

Introduction To Operations Research ............................................................................ 1
1.1 Objective ...................................................................................................................1
1.2 Introduction .............................................................................................................1
1.3 History Of Operations Research ...........................................................................2
1.4 Nature And Significance Of Operations Research ...............................................3
1.5 Features Of Operations Research Approach .......................................................6
1.6 Model And Modeling In Operations Research ....................................................7
1.7 Features Of Or Solution .........................................................................................17
1.8 Basic Operations Research Models ........................................................................17
1.9 Operations Research Techniques ...........................................................................19
Chapter Two ..........................................................................................................................20
Linear Programming..................................................................................................... 20
2.1 Objective ...................................................................................................................20
2.2 Introduction .............................................................................................................21
2.3 Structure Of Linear Programming Model ............................................................21
2.4 Applications Area Of Lp ............................................................................................22
2.5 Formulation Of Lp...................................................................................................23
2.6 Graphical Solution Of Lpp .....................................................................................24
2.7 Some Complication And Their Resolution ............................................................30
2.8 Simplex Method .......................................................................................................42
2.9 Some Complications And Their Resolution .........................................................61
2.10 Duality In Lpp ..........................................................................................................69
2.11 Sensitivity Analysis .................................................................................................80
2.12 Summar.....................................................................................................................90
Chapter Three .......................................................................................................................92
Transportation And Assignmetn Problems .................................................................... 92
3.1 Objectives .................................................................................................................92
3.2 Transportation Models..............................................................................................92
3.3 Introduction...............................................................................................................93
3.3 Assignment Problems................................................................................................117
3.4 Solution Method For Assignment ..........................................................................121
3.5 Special Case In Assignment Problems ....................................................................126
Chapter Four.........................................................................................................................138
4.1. Characteristics Of Decision Theory ......................................................................... 138
4.2. The Payoff Table ................................................................................................... 139
4.3. Decision Making Under Certainty ........................................................................... 140
4.4. Decision Making Under Complete Uncertainty ....................................................... 140
4.5. Decision Making Under Risk (With Probabilities) .................................................... 146
4.6. Decision Trees........................................................................................................ 148
Chapter Five Network Models....................................................................................... 165
5.2.1. Basic Components Of Network Diagram ......................................................... 165
5.2.2. Rules For Drawing A Network ............................................................................. 167
5.2.3 Convention For Drawing Networks ......................................................................168
5.2.4. Common Errors In Drawing Networks.................................................................. 168
Chapter Six Game Theory.............................................................................................. 181
6.1 The Pay Off Matrix Of A Game..................................................................................183
6.2 Nash Equilibrium .........................................................................................................184
6.3 Mixed Strategies ................................................................................................... 186
 OPERATIONS RESEARCH
(Mgmt-441)

RIFT VALLEY UNVERSITY CENTER FOR

DISTANCE EDUCATION

Adama 2012
Operations Research

CHAPTER ONE
INTRODUCTION TO OPERATIONS RESEARCH

1.1. Objective
1.2. Introduction
1.3. The History of Operations Research
1.4. Nature and Significance of Operations Research
1.5. Features of Operations Research
1.6. Model and Modeling in Operations Research
1.7. Summary
1.8. Review Questions

1.1 Objective

Up on the completion of this unit, the learner would be able to:

 Define operations research

 Describe significance of OR
 Explain models and their importance
 Differentiate among different categories of models
 Elucidate methodology in OR
 Identify application areas of OR models
 Describe techniques in OR

1.2 Introduction

Decision-making in today’s social and business environment has become a complex task. The
uncertainty of the future and the nature of competition and social interaction greatly increase the
difficulty of managerial decision-making. Knowledge and technology are changing rapidly, the
new problems with little or no precedents these problems and provide leadership in the
advancing global age, decision-makers cannot afford to make decisions by simply applying their
personal experiences, guesswork or intuition, because the consequences of the wrong markets,

1
Operations Research

producing the wrong products, providing inappropriate services, etc., will have major, often
disastrous consequences for organizations.

Operations Research as one of the quantitative aid to decision-making, offers the decision maker
a method of evaluating every possible alternative (act or course of action) by using various
techniques to know the potential outcomes. This is not to say, however, that management
decision-making is simply about the application of operations research techniques.

In general, while solving a real-life problem, the decision-maker must examine in both from
quantitative as well as qualitative perspective. Information about the problem from both these
perspectives needs to be brought together and assessed in the context of the problem. Based on
some mixes of the two sources of information, a decision should be taken by the decision-maker.

 Dear learner, consider a problem of an investor considering investment in three

alternatives: Stock-market, real estate and Bank Deposit. Discuss some of the quantitative
and qualitative factors (information) to be considered to suggest an acceptable solution in
each case?
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
_____________________

The evaluation of each alternative is extremely difficult or time consuming for two reasons:
First, the amount and complexity of information that must be processed; second the number of
Alternative solutions could be so large that a decision maker simply cannot evaluate all of them
to select an appropriate one. For these reasons when there is lack of qualitative information,
decision makers increasingly turn to quantitative methods and use computers to arrive at their
optimal solution to problems involving large number of alternatives. The study of these methods
and how decision makers use them in the decision process is the essence of operations research
approach.

1.3 History of Operations Research

It is generally agreed that operations Research came is to existence as a discipline during World
War II when there was a critical need to manage scarce resources. The term “Operations
research” was coined as a result of research on military operations during this war. Since the war

2
Operations Research

involved strategic and tactical problems which were greatly complicated, to expect adequate
solutions from individual or specialists in a single discipline was unrealistic. Therefore, group of
individuals who collectively were considered specialists in mathematics, Economics, statistics
and probability theory, engineering, behavioral, and physical science were formed as special unit
within the armed forces to deal with strategic and tactical problems of various military
operations. The objective was the most effective utilization of most limited military resources by
the use of quantitative techniques.

After the war ended, scientists who had been active in the military OR groups made efforts to
apply the operations research approach to civilian problems, related to business, industry,
research and development, and even won Nobel prizes when they returned to their peacetime
disciplines.

There are three important factors behind the rapid development in the use of operations research
approach.

(i) The economic and industrial boom after World War II resulted in continuous
mechanization, automation, decentralization of operations and division of management
factors. This industrialization also resulted in complex managerial problems, and
therefore application of operations research to managerial decision-making become
popular.
(ii) Many operation researchers continued their research after war. Consequently, some
important advancement were made in various operations research techniques: linear
programming and its solution by a method known as simplex method, statistical quality
control, dynamic programming, queuing theory and inventory theory were well
developed during this time.
(iii) Analytic power was made available by high-speed computers. The use of computers
made it possible to apply many OR techniques for practical decision analysis.

1.4 Nature and Significance of Operations Research

The Operations research approach is particularly useful in balancing conflicting objectives (goals
or interests), where there are many alternative courses of action available to the decision-makers.

3
Operations Research

In a theoretical sense, the optimum decision must be one that is best for the organization as a
whole. It is often called global optimum. A decision that is best for one or more sections of the
organization is usually called suboptimum decision. The OR approach attempts to find global
optimum by analyzing inter-relationships among the system components involved in the
problem. In other words, operations research attempts to resolve the conflicts of interest among
various sections of the organization and seeks the optimal solution which may not be acceptable
to one department but is in the interest of the organization as a whole.

Dear learner, discuss with your colleagues about how conflicting interests arise in the
organization and how OR tries to balance these interests?
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________

1.4.1 Operation Research: Some Definitions

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"), "Decision Science" ("DS”) and “problem solving”. In recent
years there has been a move towards a standardization upon a single term for the field, namely
the term "OR".

Because of the wide scope of application of operations research, giving a precise definition is
difficult. However, a few definitions of OR are given below.
Operations research is the application of the methods of science to complex problems in the
direction and management of large systems of men, machines, materials and money in
industry, business, government and defense. The distinctive approach is to develop a
scientific model of the system incorporating measurements of factors such as chance and risk
with which to predict and compare the outcomes of alternative decisions, strategies, or
controls. The purpose is to help management in analyze its policy and actions scientifically.
Operations Research Society, UK

4
Operations Research

Operations research is concerned with scientifically defining how to best design and operate
man-machine systems usually requiring the allocation of scarce resources.
Operations Research Society, America
Apart from being lengthy, the definition given by ORSUK, has been criticized, because it
emphasizes complex problems and large systems leaving the impression that it is a highly
technical approach suitable only to large organizations. The definition of ORSA contains an
important reference to the allocation of scarce resources. The key words used in the above
definitions are scientific approach, scarce resources, system and model. The UK definition
contains no reference to optimization, while the American definition has no reference to the
word, best.

A few other definitions, which are commonly used and widely acceptable, are:
Operations research is a systematic application of quantitative methods, techniques and tools
to the analysis of problems involving the operation of systems.
Operations research is essentially a collection of mathematical techniques and tools which in
conjunction with systems approach, is applied to solve practical decision problems of an
economic or engineering nature.
Operations Research, in the most general sense, can be characterized as the application of
scientific methods, techniques and tools, to problems involving operations of a system so as
to provide those in control of the operations with optimum solutions to the problems.
Operation research seeks the determination of the optimum course of action of a decision
problem under the restriction of limited resources. It is quite often associated almost
exclusively with the use of mathematical techniques to model and analyze decision problems.
Operations research is the application of a scientific approach to solving management
problems in order to help managers make better decisions. As implied by this and other
definitions, operations research encompasses a number of mathematically oriented
techniques that have either been developed within the field of management science or been
adapted from other disciplines, such as natural sciences, mathematics, statistics, and
engineering.

 Dear learners, would you discuss on the above definitions and define OR in your words?
___________________________________________________________________________
___________________________________________________________________________
________________________________________________________________________
5
Operations Research

1.5 Features of Operations Research Approach

From the previous discussions and various definitions of OR, important features or
characteristics can be drawn. These features of OR approach to any decision and control
problems can be summarized as:

1.5.1 Inter-disciplinary approach

Interdisciplinary teamwork is essential because while attempting to solve a complex
management problem, one person may not have complete knowledge of all its aspects such as
economic, social, political, psychological, engineering, etc. This means we should not expect a
desirable solution to managerial problems from a single individual or discipline. Therefore, a
team of individuals specializing in mathematics, statistics, computer science, psychology, etc,
can be organized so that each aspect of the problem could be analyzed by a particular specialists
in that field.
But we shouldn’t forget that certain problem situations may be analyzed even by one individual.

 Dear learner, can you mention situations which can be and cannot be analyzed by a
single individual?
1.5.2. Methodological Approach
___________________________________________________________________________
___________________________________________________________________________
___________________________________________________________________________
Operation research is the application of scientific methods, techniques and tools to problems
involving the operations of systems so as to provide those in control of operations with optimum
solutions to the problems.
Note: A system is defined as an arrangement of components designed to achieve a particular
objective(s) according to plan. The components may be either physical or conceptual or both but
they share a unique relationship with each other and with the overall objective of the system.

1.5.2 Wholistic Approach or Systems Orientation

While arriving at a decision, an operation research team examines the relative importance of all
conflicting and multiple objectives and the validity of claims of various departments of the
organization from the perspective of the whole organization.

6
Operations Research

1.5.3 Objectivistic Approach

The OR approach seeks to obtain an optimal solution to the problem under analysis. For this, a
measure of desirability (or effectiveness) is defined, based on objective(s) of the organization. A
measure of desirability so defined is then used to compare alternative courses of action with
respect to their outcomes.
1.5.4 Decision Making – OR increases the effectiveness of management decisions. It is the
decision science which helps management to make better decisions. So the major premise
of OR is decision making, irrespective of the situation involved.
1.5.5 Use of Computers: OR often requires a computer to solve the complex mathematical
model or to perform a large number of computations that are involved.
1.5.6 Human factors: In deriving quantitative solution, we do not consider human factors
which doubtlessly play a great role in the problems. So study of the OR is incomplete
without a study of human factors.
Operations research, though young, is a recognized and established discipline in the field of
business administration. The application of management science techniques is widespread, and
they have been frequently credited with increasing efficiency and productivity of business firms.

Activity
1. Operations research is an aid for the executive in making his/her decisions by providing
the needed quantitative information, based on scientific method analysis. Discuss.

2. Discuss the significance and scope of OR in modern management and Ethiopian

context?

3. Quantitative techniques complement the experience and judgment of an executive in

decision making. They do not and cannot replace it. Discuss.

1.6 Model and Modeling in Operations Research

Both simple and complex systems can easily be studied by concentrating on some portion or key
features instead of concentrating on every detail of it. This approximation or abstraction,
maintaining only the essential elements of the system, which may be constructed in various
forms by establishing relationships among specified variables and parameters of the system, is
called a model. In general, models attempt to describe the essence of a situation or activity by

7
Operations Research

abstracting from reality so that the decision- maker can study the relationship among relevant
variables in isolation.
Models do not, and cannot, represent every aspect of reality because of the innumerable and
changing characteristics of the real life problems to be represented. Instead, they are limited
approximation of reality. For example, to study the flow of materials through a factory, a scaled
diagram on paper showing the factory floor, position of equipment, tools, and workers can be
constructed. It would not be necessary to give such details as the color of machines, the height
of the workers, or the temperature of the building. For a model to be effective, it must be
representative of those aspects of reality that are being investigated and have a major impact on
the decision situation.

A
model
Dearis learner,
constructed
can to
youanalyze
mentionand understand
some the given of
of the limitations system for the purpose of improving
models?
its performance. The reliability of the solution obtained from a model depends on the validity of
_________________________________________________________________________
the model in representing the system under study. A model, allows the opportunity to examine
_________________________________________________________________________
_________________________________________________________________________
the behavioral changes of a system without disturbing the on-going operations.
__
Note: The key to model building lies in abstracting only the relevant variables that affect the
criteria of the measures of performance of the given system and expressing the relationship in a
suitable form. But oversimplification of problem can lead to a poor decision. Model enrichment
is accomplished through the process of changing constants in to variables, adding variables,
relaxing linear and other assumptions, and including randomness.

1.8.1. Classification of OR Model

As we discussed earlier, OR model is an abstract representation of an existing problem situation.
It can be in the form of a graph or chart, but most frequently an OR model consist of a set of
mathematical relationships. These mathematical relationships are made up of numbers and
symbols.
There are many ways to classify models:
i) Classification based on structure
a) Physical Models
These models provide a physical appearance of the real object under study either reduced in size
or scaled up. Physical models are useful only in design problems because they are easy to
observe, build, and describe. Since these models can not manipulated and are not very useful for

8
Operations Research

prediction, problems such as portfolio analysis selection, media selection, production scheduling,
etc cannot be analyzed by physical models.
b) Symbolic models
These models use symbols (letters, numbers) and functions to represent variables and their
relationships to describe the properties of the system.
ii) Classification based on function or purpose
Models based on the purpose of their utility include:
a) Descriptive models
Descriptive models simply describe some aspects of a situation, based on observation, survey,
questionnaire results or other available data of a situation and do not recommend anything.
Example: Organizational chart, plant layout diagram, etc.
b) Predictive Models
These models indicate “If this occurs, then that follow”. They relate dependent and independent
variables and permit trying out, “what if” questions. In other words, these models are used to
predict outcomes due to a given set of alternatives for the problem. These models do not have an
objective function as a part of the model to evaluate decision alternatives.
For example, S = a + bA +cI is a model that describes how the sales (S) of a product changes in
advertising expenditures (A) and disposal personal income (I). Here, a, b, and c are parameters
whose values must be estimated.
c) Normative (Optimization) models
These models provide the “best” or “Optimal” solution to problems subject to certain limitations
on the use of resources. These models provide recommended courses of action. For example, in
mathematical programming, models are formulated for optimizing the given objective function,
subject to restrictions on resources in the context of the problem under consideration and non
negativity of variables. These models are also called prescriptive models, because they prescribe
what the decision maker ought to do.

iii) Classification Based on Time Reference

a) Static Models
Static models represent a system at some specified time and do not account for changes over
time. For example, an inventory model can be developed and solved to determine an economic

9
Operations Research

order quantity for the next period assuming that the demand in planning period would remain the
same as that for today.
b) Dynamic models
In dynamic models, time is considered as one of the variables and allows the impact of changes
due to change in time. Thus, sequences of interrelated decisions over a period of time are made
to select the optimal course of action to optimize the given objective. Dynamic programming is
an example of a dynamic model.
iv) Classification based on Degree of certainty
a) Deterministic Models
If all the parameters, constants and functional relationships are assumed to be known with
certainty when the decision is made, the model is said to be deterministic. Thus, in such a case,
the outcome associated with a particular course of action is known. That is, for a specific set of
input values, there is a uniquely, determined output which represents the solution of the model
under consideration of certainty. The results of the models assume single value. Linear
programming models are examples of deterministic models.
b) Probabilistic (Stochastic) models
Models in which at least one parameter or decision variable is a random variable are called
probabilistic (or stochastic). Since at least one decision variable is random, a dependant variable
which is the function of independent variable(s) will also be random. This means consequences
or pay off due to certain changes in the independent variable can not be produced with certainty.
However, it is possible to predict a pattern of values of both the variables by their probability
distribution. Insurance against risk of fire, accidents, sickness, etc are examples where the pattern
of events is studied in the form of a probability distribution.
v) Classification Based on Method of solution or Quantification
a) Heuristic Model
These models employ some sets of rules which, though perhaps not optimal, do facilitate
solutions of problems when applied in a consistent manner.
b) Analytical Models
These models have a specific mathematical structure and thus can be solved by known analytical
or mathematical techniques. Any optimization model (which requires maximization or
minimization of an objective function) is an analytical model.

10
Operations Research

c) Simulation Models
These models have a mathematical structure but are not solved by applying mathematical
techniques to get a solution. Instead, a simulation model is essentially a computer-assisted
experimentation on a mathematical structure of a real-life problem in order to describe and
evaluate its behavior under certain assumptions over a period of time.
Simulation models are more flexible than mathematical ones and therefore, can be used to
represent a complex system which otherwise cannot be represented mathematically. These
models do not provide general solution like those of mathematical Models.

 Dear learner, do you think that the above classification of models is mutually
exclusive? Support your response with evidence.
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
6. Advantage of Models
___
Models in general are used as an aid for analyzing complex problems. However, a model can
also serve other purposes as:
i) A model provides economy in representation of the realities of the system. That is,
models help decision makers to visualize a system so that he/she can understand the
system’s structure or operation in a better any. For example, it easier to represent a
factory lay out on paper than to construct it. It is cheaper to try out modifications of such
systems by rearrangement on paper.
ii) The problem can be viewed in its entirety, with all the components being considered
simultaneously.
iii) Models serve as aids to transmit ideas and visualization among people in the
organization. For example, process chart can help the management to communicate about
better work methods to workers.
iv) A model allows us to analyze and experiment in a complex situation to a degree that
would be impossible in the actual system and its environment. For example, the
experimental firing of satellite may be costly and require years of preparation.
v) Models simplify he investigation considerably and provide a powerful and flexible for
predicting the future state of the process or system.

11
 Operations Research

1.7.3. Methodology of Operations Research

For effective use of OR techniques, it is essential to follow some steps that are helpful for
decision-makers to make better solution. The flow diagram representing the methodology of OR
is shown as:

Abstraction System Model Building

Choosing a Logic
- Establish relationships among variables
Real World particular aspect and parameters of the system.
problem of reality which - Define objectives to be achieved and
needs attention limitations on resources

Solve the Model

Apply suitable OR technique to get solution in
terms of decision variables.

Modify the Model Testing the Model and its solution

- Put values of decision variables in the

model under consideration
Not acceptable - See whether solution in valid or not

Implementation and Control

- Interpret solution values

- Put the knowledge (result) gained Acceptable
form the solution to work through
organizational policies
- Monitor changes and exercise control.

Step 1. Observation and Defining a problem

The first step in OR process is the identification of a problem that exists is a system
(organization). The system must be continuously and closely observed so that problems can be
identified as soon as they occur or anticipated. Problems are not always the results of crisis; but
instead frequently involve an anticipatory or planning situation. Once it has been determined that
a problem exists, the problem must be clearly and concisely defined. Because improperly
defining a problem can easily result in no solution or an inappropriate solution. Since the
existence of a problem implies that the objectives of the firm are not being met in some way, the
goals (objectives of the organization) must also be clearly defined.

 Dear learner, can you identify individual(s) who are responsible to identify problems and
problems they face while they identify these problems?
___________________________________________________________________________12
_________
Operations Research

Step 2. Formulating a model

Model formulation involves an analysis of the system under study, determining objective of the
decision-maker, and alternative course of action, etc, so as to understand and describe, in precise
terms, the problem that an organization faces.

The major steps which have to be taken in to consideration for formulating the model are:
 Problem Components. The first component of the problem to be defined is the decision
maker who is not satisfied with the existing state of affairs. The interaction with the
decision maker will help the OR specialist in knowing his/her objectives. That is, either
he/she has already obtained some solution of the problem and wants to retain it, or he
wants to improve it to a higher degree. If the decision maker has conflicting multiple
objectives, he/she may be advised to rank the objectives in the order of preference;
overlapping objectives may be eliminated.
 Decision environment
It is desirable to know about the resources such as managers, employees equipments, etc which
are required to carry out the policies of the organization considering the social and ecological
environment in which the organization functions. Knowledge of such factors will help in
modifying the initial set of decision-maker’s objectives.
 Alternative courses of Action
The problem arises only when there are several courses of action available for a solution. An
exhaustive list of course of action can be prepared in process of going through the above steps of
formulating the problem. Courses of action which are not feasible with respect to objectives and
resources may be ruled out.

 Measure of effectiveness
A certain measure of effectiveness or performance is required in order to evaluate the merit of
the several courses of action. The performance or effectiveness can be measured in different
units such as birr (net profits), percentage (share of market desired), time dimension (service or
waiting time).

13
Operations Research

 Collecting Data and Constructing a Mathematical Model

After the problem is clearly defined and understood, the next step is to collect required data and
then formulate a mathematical model. Model construction consists of hypothesizing relationships
between variables subject to and not subject to control by decision-maker. Certain basic
components required in every decision problem model are:
o Controllable (decision) Variables - These are the issues or factors in the problem whose
values are to be determined (in the form of numerical values) by solving the model. The
possible values assigned to these variables are called decision alternatives (strategies or
courses of actions). Example, in LPP the number of units produced is a decision variable.
o Uncontrollable variable. These are the factors whose numerical value depends up on the
external environment prevailing around the organization. The values of these variables
are not under the control of the decision-maker and are also termed as state of nature.
 Objective function.
It is a representation of (i) the criterion that expresses the decision-maker’s manner of
evaluating the desirability of alternative values of decision variables, and (ii) how that
criterion is to be optimized (minimized or maximized. A stated objective helps to focus
attention on what the problem actually is.
 Constraints or Limitations
These are the restrictions on the values of the decision variables. These restrictions can arise due
to limited resources such as space, money, manpower, material, etc. The constraints may be in
the form of equations or inequalities.
 Functional relationships
In a decision problem, the decision variables in the objective function and in the constraints are
connected by a specific functional relationship. A general decision problem model might take the
form:
Optimize (Max or Min) Z = f(x)
Subject to the constraints:
gi(x) (≤, = ≥ ) bi; i = 1,2,…..m and x ≥ 0
Where, x = a vector of decision variables (x1, x2, x3, xn)
f(x) = Criterion or objective function to be optimized
gi(x) = the ith constraint
bi = fixed amount of the ith resource

14
Operations Research

A model is referred to as a linear model if all functional relationships among decision variables
X1, X2, Xn in f(x) and g(x) are of a linear form. But if one or more of the relationships are non –
linear, the model is said to be a non-linear model.
 Parameters- These are constants in the functional relationships. Parameters can be
deterministic or probabilistic in nature.

? Dear learner, take the case of furniture manufacturing unit in your nearby
village. Identify and define the unstructured problem of the unit and construct
a real OR model for the unit?
_______________________________________________________________
_______________________________________________________________
_______________________________________________________________
____
Step 3.Solving the Mathematical Model
This involves obtaining the numerical values of decision variables. Obtaining these values
depends on the specific form or type, of mathematical models. Solving the model requires the
use of various mathematical tools and numerical procedures. In general, there are two categories
of methods used for solving an OR model.
 Optimization model. These models yield the best value for the decision variables both for
unconstrained and constrained problems. In constrained problems, these values
simultaneously satisfy all of the constraints and provide an optimal or acceptable value
for the objective function or measure of effectiveness. The solution so obtained is called
the optimal solution to the Problem.
 Heuristic Model. These methods yield values of the variables that satisfy all the
constraints, but not necessarily provide optimal solution. However, these values provide
an acceptable value of the objective function.
Heuristic Methods are sometimes described as “rules of thumb” which work. These methods
are used when obtaining optimal solution is either very time consuming or the model is
complex.

15
Operations Research

(i) Failure to recognize the existence of a problem (iii) Failure to use all available
information
- Some people tend to personalize problems - The problem-solver fails to seek out
- Problems arise in context which people have had no information
experience.
ii)Failure to define the correct problem iv) Failure to recognize or question
- One situation may contain many intertwined assumptions.
problems. - it is assumed that there is a
- Obvious problems are often symptoms of much solution to every problem
deeper problems. - Rigid thinking blocks ones view
- The inability to identify accurately what is going on point.
can lead to inaccurate problem identification.
- Attitudes and beliefs can blind the problem solver
to the real causes on undesirable situations.
Difficulties in problem solving
Sometimes difficulties in problem solving arise due to lack of an appropriate methodology
for it and psychological perceptions on the part of the problem solver. The major difficulties
in problem solving:
Step 4. Validating (Testing) the solution
After solving the mathematical model, it is important to review the solution carefully to see
that values make sense and that the resulting decisions can be implemented. Some of the
reasons for validating the solution are:
(i) The mathematical model may not have enumerated all the limitations of the problem
under consideration.
(ii) Certain aspect of the problem may have been overlooked, omitted or simplified,
(iii) The data may have been incorrect estimated or recorded, perhaps when entered in to the
computer.
Step 5. Implementing the solution
The decision-maker has not only to identify good decision alternatives but also to select
alternatives that are capable of being implemented. It is important to ensure that any solution
implemented is continually reviewed and updated in the light of a changing environment.

16
Operations Research

Step 6. Modifying the Model

For a mathematical model to be useful, the degree to which it actually represents the system or
problem being modeled must be established. If during validation, the solution cannot be
implemented, one needs to (a) identify constraints that were omitted during the original problem
formulation or (b) find if some of the original constraints were incorrect and need to be modified.
In all such cases, one must return to the model formulation step and carefully make the
appropriate modifications to represent more accurately the given problem. A model must be
applicable for a reasonable time period and should be updated from time to time, taking in to
consideration the past, present, and future aspects of the problem.
Step 7. Establishing control over the solution
The dynamic environment and changes within the environment can have significant implications
regarding the continuing validity of models and their solutions. Thus, a control procedure has to
be established for detecting significant changes in decision variables of the problem so that
suitable adjustments can be made in the solution without having to build a model every time a
significant change occurs.
1.7 Features of OR solution
A solution that works but is quite expensive compared to the potential savings from its
application should not be considered successful. Also a solution that is well within the budget but
which does not accomplish the objective is not successful either. The following are features of
good solution:
 Technically appropriate. The solution should work technically; meet the constraints and
operate in the problem environment.
 Reliable. The solution must be useful for a reasonable period of time under the conditions
for which it was designed.
 Economically viable. The economic value should be more that what it costs to develop
and should be seen as wise investment in hiring OR talents.
 Behaviorally appropriate. The solution should be behaviorally appropriate and must
remain valid for reasonable period of time within the organization.
1.8 Basic Operations Research Models
There is no unique set of problems which can be solved by using OR models or techniques.
Some OR models or techniques include:

17
Operations Research

Allocation Models
Allocation models are used to allocate resources to activate in such a way that some measure of
effectiveness (objective function) is optimized. Mathematical programming is the broad term of
OR techniques used to solve allocation problems.
If the measure of effectiveness such as profit, cost, etc., is represented as a linear function of
several variables and if limitations of resources (constraints) can be expressed as a system of
linear inequalities or equalities, the allocation problem is classified as linear programming
problems.
But, if the objective unction of any or all constraints can not be expressed as a system of linear
equalities or inequalities, the allocation problem is classified as a non-linear programming
problem.
When the solution values or decision variables for the problem are restricted to being integer
values, the problem is classified as an integer programming. The problem having multiple,
conflicting and incommensurable objective function (goals) subject to linear constraints is called
goal programming. If decision variables in the linear programming problem depend on chance,
such a problem is called is called a stochastic programming problem.
If resources such as workers, machines or salesmen can be assigned to perform a certain number
of activities such as jobs or territories on one- to one basis so as to minimize total time, cost or
distance involved in performing a given activity, such problems are classified as assignment
problems. Conversely, if the resources can be used for more than one activity, the allocation
problem is classified as a transportation problem.
Inventory Model
Inventory Models deal with the problem of determination of how much to order at a point in time
and when to place an order. The main objective is to minimize the sum of three conflicting
inventory costs: the Cost of holding or carrying extra inventory, the cost of shortage or delay in
delivery of items when it is needed, a cost of ordering or set-up.
Competitive (Game Theory) Model
These models are used to characterize the behavior of two or more opponents (called players)
who compete for the achievement of conflicting goals.

18
 Operations Research

Network Models
These models are applied to the management (planning, controlling and scheduling) of large-
scale projects. PERT/CPM Techniques help in identifying potential trouble spots in a project
through the identification of the critical path. These techniques improve project coordination and
enable the efficient use of resources. Network methods are also used to determine time-cost trade
off, resource allocation and updating of activity time.
Decision analysis Model
These models deal with the selection of an optimal course faction given the possible pay offs and
their associated probabilities of occurrence. These models are broadly applied to problems
involving decision making under risk and uncertainty.

1.9. Operations Research Techniques

OR techniques can be loosely classified in to five categories

Management Science Techniques

Linear Mathematical Probabilistic Inventory Net work Other linear and non-
Programming Techniques Techniques Techniques linear Techniques

Linear programming Probability Certain Demand Net work flow Dynamic

models Decision Analysis Uncertain Demand CPM/PERT Programming
Graphic Analysis Game Theory Break-even Analysis
Simplex Method Markov Analysis Non-linear
Post optimality Queuing programming
Transportation and Simulation
assignment Forecasting
Integer linear
programming
Goal Linear
Programming

Source: Taylor, 1990, Introduction to Management Science, 3rd edition, Brown Publisher

19
Operations Research

CHAPTER TWO
LINEAR PROGRAMMING

2.1 Objective
2.2 Introduction
2.3 Structure of Linear programming Model
2.4 Application Area of LP
2.5 Formulation of LP
2.6 Graphical Solution of LPP
2.6.1 Maximization in LPP
2.6.2 Minimization in LPP
2.7 Some complication and their Resolution
2.8 The Simplex Method
2.8.1 Maximization Problem
2.8.2 Minimization Problem
2.9 Some complication and their Resolution
2.10 Duality in LPP
2.11 Sensitivity Analysis
2.12 Summary
2.13 Review Questions

2.1 Objective
Up on completion of this unit, the learner is expected to:

 Recognize problems that can be solved using LP models.

 Formulate an LP model in mathematical terms.
 Solve Linear Programming Problems (LPP) using both graphic and simplex approach.
 Explain special cases in both graphic and simplex techniques.

20
Operations Research

2.2 Introduction
Linear Programming is a mathematical process that as bee developed to help management in
decision making involving the efficient allocation of scares resources to achieve a certain
objective.
Diagrammatically,

Scares To be allocated to:

Resource

Resource
constraints
Objectives Constraints

Non-negativity
Constraints

Optimization

Maximization Minimization

2.3 Structure of Linear Programming Model

LP is a method for choosing the best alternative from a set of feasible alternatives
To apply LP, the following conditions must be satisfied:
a. Objective Function: Is the goal or objective of a management, stated as an intent to maximize
or to minimize some important quantity such as profits or costs.
b. Constraints: Are limitations or restrictions imposed by the problems. And constraints
include:

21
Operations Research

1. Resource constraints: Are restrictions that should be clearly identifiable and measurable in
quantitative terms, which arise from limitation of available resources.

Examples of limited resources:

 Plant capacity
 Raw materials availability
 Labor power
 Market demand, etc

2. Non-negativity constraints:
Are constraints that require the decision variables not to take on negative values
c. Linearity
The Objective Function and the constraints must be linear in nature in order to have a Linear
Programming Problems (LPP)

d. Feasible alternative
There should be a series of feasible alternative course of action available to the decision-making
determined by resource constraints. Thus, we have to choose the best alternative
Linear Programming Problems can be solved by using:
i. The Geometric method called” Graphical Method”
ii. The Algebraic method called” Simplex Method”

2.4 APPLICATIONS AREA OF LP

There is a wide range of problems that lend themselves to solution by linear programming
techniques. This discussion is only meant to give an indication of the LP techniques for
managerial decision making and the apparent diversity of situations to which linear programming
can be applied. Some of these include production management (product mix, blending problems,
production planning, Assembly line balancing…), Marketing management (media selection,
traveling sales man problem, physical distribution), Financial management (portfolio selection,
profit planning), agricultural application, military applications, personnel management (staffing
problem, Determination of equitable salary and etc.

22
Operations Research

Product Mix
Organizations often produce similar services that use the same resources. For example, labor,
material cost, etc. because of limited resources available during any time period a decision must
be made concerning how much of each product to produce or make available. Linear
programming answers what mix of output (or service) will maximize profit given the availability
of scarce resources.
Diet problem
It usually involves the mixing of raw materials or other ingredients to obtain an end product that
has certain characteristics. For example, what mix of inputs will achieve the desired results in the
output for the least cost? Other applications that fall into this category include mixing feed for
livestock, mixing pet foods, mixing building materials (concrete, mortar, paint), and so on.
Blending problems
They are very similar to diet problems. Strictly speaking, however, blending problems have
additional requirement, i.e. to achieve a mix that have specific consistency. For example, how
many quarts of the different juices each with different sugar content proportion must be mixed
together to achieve one gallon that has a sugar content of 17 percent?
Portfolio selection
These problems generally involve allocating a fixed dollar amount among a variety of
investments, such as bonds, sockets, real estates, etc. The goal usually is to maximize income or
total return. The problems take on an added dimension when certain other requirements are
specified (for example, no more than 40 percent of the portfolio can
2.5 FORMULATION OF LP
Decision variables are the variables whose values are unknown and are searched
The coefficients of the variables in the Objective Function are called the profit or cost
coefficients. They express the rate at which the value of the Objective Function increases or
decreases by including in the solution one unit of each of the decision variables.
The coefficients of the constraints’ variables are called the input- output coefficients that indicate
the rate at which the given resources are depleted or utilized.

23
Operations Research

Example:
Max.Z  50 X 180 X 2
St :
X 12 X 2 32
X 12 X 2 82
X1, X 2  0

2.6 GRAPHICAL SOLUTION OF LPP

To use the graphic method, the following steps are needed:
1. Identify the problem
i.e: The decision variables, the objective function and the constraints
2. Draw a graph including all the constraints and identify the feasible region
3. Obtain a point on the feasible region that optimizes the objective function-Optimal solution
4. Interprite the results
 Graphical LP is a two-dimensional model.

2.6.1 Maximization Problem

==>Maximize Z with inequalities of constraints in < form
Example: Consider two models of color TV sets; Model A and B, are produced by a company
to maximize profit. The profit realized is $300 from A and $250 from set B. The limitations are
a. availability of only 40hrs of labor each day in the production department.
b. a daily availability of only 45 hrs on machine time
c. ability to sale 12 set of model A.
How many sets of each model will be produced each day so that the total profit will be as large
as possible?
Resources used per unit
Constraints Model A Model B Maximum Available hrs.
(X1) (X2)

Labor hr. 2 1 40
Machine hr. 1 3 45
Marketing hr. 1 0 12
Profit $300 $250

24
Operations Research

Solution
1. Formulation of mathematical modeling of LPP
Max Z=300X1 +250X2
St:
2X1 +X2< 40
X1 +3X2< 45 LPP Model
X1 < 12
X1 , X2 > 0
2. Convert constraints inequalities into equalities
2X1 +X2 = 40
X1 +3X2= 45
X1 = 12
3. Draw the graph by intercepts
2X1 +X2 = 40 ==> (0, 40) and (20, 0)
X1 +3X2= 45==> (0, 15) and (45, 0)
X1 = 12==> (12, 0)
X1 , X2 = 0
2X1 +X2 = 40

X2
X1=0
40 X1=12

B
X1 +X2 = 45

15

Feasible C(12, 11)

Region X2=0
X1
D
A 12 20 45

4. Identify the feasible area of the solution which satisfies all constrains.
5. Identify the corner points in the feasible region
A (0, 0), B (0, 15), C (12, 11) and D (12, 0)
6. Identify the optimal point
7. Interprete the result
Corners Coordinates MaxZ=300 X1
+250X2
A (0, 0) $0
B (0, 15) $3750
C (12, 11) $6350
D (12, 0) $3600

25
Operations Research

Interpretation:
12 units of product A and 11 units of product B should be produced so that the total profit will be
$6350.

Exercise:
A manufacturer of light weight mountain tents makes two types of tents, REGULAR tent and
SUPER tent. Each REGULAR tent requires 1 labor-hour from the cutting department and
3labor-hours from the assembly department. Each SUPER tent requires 2 labor-hours from the
cutting department and 4 labor-hours from the assembly department .The maximum labor hours
available per week in the cutting department and the assembly department are 32 and 84
respectively. Moreover, the distributor, because of demand, will not take more than 12 SUPER
tents per week. The manufacturer sales each REGULAR tents for $160 and costs$110 per tent to
make. Where as SUPER tent ales for $210 per tent and costs $130 per tent to make.
Required:
A. Formulate the mathematical model of the problem
B. Using the graphic method, determine how many of each tent the company should manufacture
each tent the company should manufacture each week so as to maximize its profit?
C. What is this maximum profit assuming that all the tents manufactured in each week are sold in
that week?
Solution
_____________________________________________________________________
Labor hours per tent
Department REGULAR (X1) SUPER(X2) Maximum
labor-hours
Available per week
_____________________________________________________________________

Cutting department 1 2 32
Assembly department 3 4 84
Selling price per tent $160 $210
Cost per tent $110 $130
Profit per tent $50 $80

26
Operations Research

*The distributor will not take more than 12 SUPER tents per week. Thus, the manufacturer
should not produce more than 12 SUPER tents per week.
Let X1 =The No of REGULAR tents produced per week.
X2 =The No of SUPER tents produced per week.
X1 and X2 are called the decision variables
Max .Z  50 X 180 X 2
St :
X 12 X 2 ……….Cutting
32 department constraint
LPP Model ……….Assembly department constraint
3 X 14 X 2 84
……….Demand constraint
X 2  12
X 1 , X 2  0……….Non-negativity constraints

Corners Coordinates MaxZ=50 X1 +800X2

A (0, 0) $0
B (0, 12) $960
C (8, 12) $1360
D (20, 6) $1480
E (28, 0) $1400

Interpretation:
The manufacturer should produce and sale 20 REGULAR tents and 6 SUPERS tents to get a
maximum weekly profit of $1480.

2.6.2 Minimization Problem

==>Minimize Z with inequalities of constraints in > form
Example:
Suppose that a machine shop has two different types of machines; machine 1 and machine 2,
which can be used to make a single product .These machines vary in the amount of product
produced per hr., in the amount of labor used and in the cost of operation.

27
Operations Research

Assume that at least a certain amount of product must be produced and that we would like to
utilize at least the regular labor force. How much should we utilize each machine in order to
utilize total costs and still meets the requirement?

Solution
________________________________________________________________
Resource used
Machine 1 (X1) Machine (X2) Minimum required hours
_____________________________________________________________________

Product produced/hr 20 15 100

Labor/hr 2 3 15________
Operation Cost $25 $30

Min.Z  25 X 130 X 2
St :
20 X 115 X 2 100 LPP Model

2 X 13 X 2 15
X1, X 2  0
Constraint equation:
20X1 +15X2=100 ==> (0, 20/3) and (5, 0)
2X1+3X2=15 ==> (0, 5) and (7.5, 0)
X1 X2 > 0

X2
X1 =0
A (0, 20/3)

Feasible Region

B (2.5, 3.33)
_____________________________________________________________________
X2 =0
Corners Coordinates MinZ=25 X1 + 30X2
X1
A (0, 20/3) 200
5 C (7.5, 0)

28
Operations Research

B (2.5, 3.33) 162.5

C (7.5, 0) 187.5
_______________________________________________________________
X1 =2.5
X2=3.33 and
MinZ= 162.5
Exercise:
A company owns two flour mills (A and B) which have different production capacities for
HIGH, MEDIUM and LOW grade flour. This company has entered contract supply flour to a
firm every week with 12, 8, and 24 quintals of HIGH, MEDIUM and LOW grade respectively.
It costs the Co. $1000 and $800 per day to run mill A and mill B respectively. On a day, mill A
produces 6, 2, and 4 quintals of HIGH, MEDIUM and LOW grade flour respectively.
Mill B produces 2, 2 and 12 quintals of HIGH, MEDIUM and LOW grade flour respectively.
How many days per week should each mill be operated in order to meet the contract order most
economically standardize? Solve graphically.
Solution:
No of days per week of Minimum flour in
Mil A (X1) Mill B(X2) quintals
HIGH Capacity (in quintal) 6 2 12
MEDIUM Capacity (in quintal) 2 2 8
LOW Capacity (in quintal) 4 12 24
$1000 $800
Min.Z  100 X 1800 X 2
St :
6 X 12 X 2 12
2 X 12 X 2 8
4 X 1  12 X 2  24
X1, X 2  0
Constraint equation:

29
Operations Research

6 X 12 X 2 12
(0, 6), (2, 0)
2 X 12 X 2 8
(0, 4), (4, 0)
4 X 1  12 X 2  24
(0, 2), (6, 0)
X1, X 2  0
Corners MinZ=$1000 X1 + 800X2
(0, 6) $4800
(1, 3) $3400
(3, 1) $3800
(6, 0) $6000
X1 =1
X2=3 and
MinZ= $3400

X2
X1 =0
6 6X1+2 X2=12
2X1+2 X2=8
4 FR
4X1+12 X2=24
(1, 3)
(3, 1)
X2 =0
X1
2 4 6

Note:
-In maximization problems, our point of interest is looking the furthest point from the origin.
-In minimization problems, our point of interest is looking the point nearest to the origin.

2.7 SOME COMPLICATION AND THEIR RESOLUTION

1. Redundant Constraint
If a constraint when plotted on a graph doesn’t form part of the boundary making the feasible
region of the problem that constraint is said to be redundant.

Example:

30
Operations Research

A firm is engaged in producing two products A and B .Each unit of product A requires 2Kg of
raw material and 4 labor-hrs for processing. Whereas each unit of product B requires 3Kg of raw
materials and 3hrs of labor. Every unit of product A needs 4hrs to packaging and every unit of
product B needs 3.5hrs for packaging. Every week the firm has availability of 60Kg of raw
material, 96 labor-hours and 105 hrs I the packaging department.1 unit of product A sold yields
$40 profit and 1 unit of B sod yields $35 profit.

Required:
a. Formulate this problem as a LPP
b. Find the optimal solution

Solution
__________________________________________________________________
Products Resource available
Resources A B per week
_____________________________________________________________________

Raw materials (Kg) 2 3 60

Labor (hr) 4 3 96
Packaging (hr) 4 3.5 105
Profit per unit $40 $35

Let X1 =The No of units of product A produced per week

X2 =The No of units of product B produced per week
a. LPP Model
Max .Z  40 X 135 X 2
St :
2 X 13 X 2 60
4 X 13 X 2 96
4 X 1  3.5 X 2  105
X1, X 2  0

31
Operations Research

X2
(0, 32)

Labor: 4X1 +3X2 = 96

(0, 30)
Packaging: 4X1 +3.5X2 = 105
(0, 20) C (18,8)
Raw material: 2X1 +3X2 = 60
FR
X1
A (0, 0) D (24, 0) (26, 0) (30, 0)

 The packaging hr is redundant.

Corners Coordinates MinZ=40 X1 + 35X2
A (0, 0) 0
B (0, 20) 700
C (18, 8) 1000
D (24, 0) 960
X1 =18
X2=8 and
MinZ= 1000
Interpretation:
The company should produce and sale 18 units of product A and 8 units of product B per week
so as to get a maximum profit of 1000.
 By this production plan the entire raw material will be consumed.
 2X1 +3X2 <60
2(18) +3(8) =60
60=60==> N o idle or unused raw material
 4X1 +3X2 <96
4(18) +3(8) <96
96=96 ==>the entire labor hour will be consumed
 4X1 +3.5X2 <105
100<105==>There is to be idle or unused capacity of 5hrs in the packaging department.

32
Operations Research

Note: The packaging hour’s constraint does not form part of the boundary making the feasible
region. Thus, this constraint is of no consequence and is therefore, redundant. The
inclusion or exclusion of a redundant constraint does not affect the optimal solution of
the problem.

2. Multiple optimal Solutions

/Alternative optimal solutions/
-This is a situation where by a LPP has more than one optimal solution.
Multiple optimal Solutions will be found if two corners give optimal solution, then the line
segment joining these points will be the solution.
==>We have unlimited number of optimal solution with out increasing or decreasing the
objective function.

Example:
The information given below is for the products A and B.
_____________________________________________________________________
Machine hours per week Maximum available
Department Product A Product B per week
_____________________________________________________________________

Cutting 3 6 900
Assembly 1 1 200
Profit per unit $8 $16
_____________________________________________________________________
Assume that the company has a marketing constraint on selling products B and therefore it can
sale a maximum of 125units of this product.
Required:
a. Formulate the LPP of this problem
b. Find the optimal solution
Solution:
Let X1 =The No of units f product A produced per week
X2 =The No of units f product B produced per week

33
 Operations Research

a. The LPP Model of the problem is:

Max .Z  8 X 116 X 2
St :
3 X 16 X 2 900
X 1 X 2 200
X 2  125
X1, X 2  0

X2 X1=0
(0, 200)

(0,150)
B (0, 125)
D (100,100)
C (50, 125) X2=125 Marketing equation

Cutting: 3X1+6X2=900

FR X2=0
A (0, 0)
X1
E (200, 0) (300,0)

Corners Coordinates MaxZ=8 X1 + 16X2

A (0, 0) 0
B (0, 125) 2000
C (50, 125) 2400
D (100, 100) 2400
E (200, 100) 160
Interpretation:
Both C and D are optimal solutions. Any point on the line segment CD will also lead to the same
optimal solution.
==>Multiple optimal solutions provide more choices for management to reach their objectives.
3. Infeasible Solution
A solution is called feasible if it satisfies all the constraints and the constraints and non-
negativity condition. However, it is sometimes possible that the constraints may be inconsistent
so that there is no feasible solution to the problem. Such a situation is called infeasibility.

34
 Operations Research

Example:
MaxZ=20X1+30X2
St:
2X1+X2< 40
4X1+X2< 60
X1 > 30
X1, X2 > 0
Solution:

X2 X1=0
(0, 60) X1=30
4X1+X2= 60
(0, 40)

2X1+X2= 40
X2=0
X1
(15, 0) (20, 0) (30, 0)
Note:
-In the above graph, there is no common point in the shaded area.
-All constraints cannot be satisfied simultaneously and there is no feasible solution to the
problem.
4. Mix of constraints
Example:
ABC Gasoline Company has two refineries with different production capacities. Refinery A can
produce 4,000gallons per day of SUPER UNLEADD GASOLINE, 2000 gallons per day of
REGULAR UNLEADED GASOLINE and 1000 gallons per day of LEADED GASOLINE. On
the other hand, refinery B can produce 1000 gallons per day of SUPER UNLEADED, 3000
gallons per day of REGULAR UNLEADED and 4,000 gallons per day of LEADED.

35
Operations Research

The company has made a contract with an automobile manufacturer to provide 24000 gasoline’s
of SUPER UNLEADED, 42000 gallons of REGULAR UNLEADED and 36000 gallons of
LEADED .The automobile manufacturer wants delivery in not more than 14 days.
The cost of running refinery A is $1500 per day and refinery B is $2400 per day.
Required:
a. Formulate this problem as a LPP
b. Determine the number of days the gasoline company should operate each refinery in order to
meet the terms of the above contract most economical.(i.e. At a minimum running cost)
c. Which grade of gasoline would be over produced?
Solution:
_____________________________________________________________________
Production per day (in gallons) Contract with an automobile
manufacturer
Grade of gasoline A B
_____________________________________________________________________
SUPER UNLEADED 4000 1000 24,000
REGULAR UNLEADED 2000 3000 42,000
LEADED 1000 4000 36,000
Running cost per day $1,500 $2,400
_____________________________________________________________________
 The automobile manufacturer wants delivery in not more than 14 days.
Let X1 =The No of days refinery A should work.
X2 =The No of days refinery B should work.
a. LPP of the problem
MinZ=1500X1+2400X2
St:
4000X1+1000X2>24000
2000X1+3000X2>42000
1000X1+2000X2> 36000
X1 < 14
X2< 14
X1, X2 > 0
==>T o simplify the problem divide by 1000 the constraints

36
 Operations Research

MinZ=1500X1+2400X2
St:
4X1+1X2>24
2X1+3X2>42
X1+4X2 > 36
X1 < 14
X2< 14
X1, X2 > 0

24 Delivery time: X1=14

SUG: 4X1+X2 =24

A (2.5, 14) B (14, 14) Delivery time: X2=14

FR

D (12, 6) C (14, 5.5)

LG: X1+4X2=36

RUG: 2X1+3X2 =42

(6, 0) (14, 0) (21, 0) (36, 0)

Note: Point A, B, C, and D are solved by elimination-substitution method

________________________________________________________________
Corners Coordinates MaxZ=1500X1 + 2400X2
A (2.5, 4) $37350
B (14, 14) 54600
C (14, 5.5) 34200
D (12, 6) 32400
E (3, 12) 33300
_________________________________________________________________
Interpretation:
The oil company should operate refinery A for 12 days and refinery B for 6 days at a minimum
operating cost of $32,400.
c. Is there any over production

37
Operations Research

SUG: 4000X1+1000X2>24000
4000(12) +1000(6)>24000
54000 > 24000
Therefore, 30,000 gallons over production

RUG: 2000X1+3000X2>42000
2000(12) +3000(6)>42000
42000 > 42000
Therefore, there is no over production of RUG

LG: 1000X1+4000X2>36000
1000(12) +1000(6)>36000
36000 > 36000
Therefore, No over production of LG

5. Unbounded Solution
When the value of decision variables in LP is permitted to increase infinitely without violating
the feasibility condition, then the solution is said to be unbounded .Here, the objective function
value can also be increased infinitely. However, an unbounded feasible region may yield some
definite value of the objective function.
Example:
Use the graphical method to solve the following LPP.
1. MaxZ=3X1+4X2
St:
X1-X2<-1==> -X1+X2>1 since the quantity solution is positive
-X1+X2<0
X1, X2 > 0
X2 X1-X2 =-1

X1+X2 =0

1 Unbounded
Feasible Region

X1

38
Operations Research

Fig: Unbounded Solution

2. Max.Z=3X1+2X2
St:
X1-X2<1
X1+X2<3
X1, X2 > 0
X2

A(0,3) Unbounded
Feasible Region
X1-X2=1

B (2, 1)
X1+X2=3

X1

Note here that the two corners of the region are A(0,3) and .B(2,1).The value of MaxZ(A)=6 and
MaxZ(B)=8. But there exist number of points in the shaded region for which the value of the
objective function is more than 8.For example, the point (10, 12) lies in the region and the
function value at this point is 70 which is more than 8.

Remark:
An unbounded solution does not mean that there is no solution to the given LPP, but implies that
there exits an infinite number of solutions.

Exercise:
Use graphical method to solve the following LPP.
1. MaxZ=7/4X1+3/2X2 2. MaxZ=3X1+2X2
St: St:
8 X1+5X2 < 320 -2X1+3X2 < 9
4X1+5X2 < 20 X1-5X2 > -20
X1 > 15 X1, X2 > 0
X2> 10
X1, X2 > 0

39
Operations Research

Answer: No feasible solution Answer: Unbounded solution

3. Max.Z=3X1+2X2 4.Max.Z=X1+X2
St: St:
X1-X2 < 1 X1+X2 < 1
X1+X2> 3 -3X1+X2> 3
X1, X2> 2 X1, X2> 0
Answer: Unbounded solution Answer: Unbounded solution
5. Max.Z=6X1-4X2 6.Max.Z=X1+1/2X2
St: St:
2X1+4X2 < 4 3X1+3X2 < 12
4X1+8X2> 16 5X1 < 10
X1, X2 > 0 X 1 + X2 >8
Answer: Infeasible solution -X1 + X2 >4
X1, X2 > 0
Answer: Infeasible solution
Exercise
I. Solve the following LP problems using the graphical method.
1. Max.Z=15X1-10X2 2.Max.Z=2X1+X2
St: St:
4X1+6X2 < 360 X1+2X2 < 10
3X1+0X2< 180 X1 +X2 < 6
0X1+5X2< 280 X1 - X2 < 2
X1, X2 > 0 X1 -2X2 < 1
Answer: X1=60 , X2 =20 X1, X2 >0
and Max.Z=1,100 Answer: X1=4, X2 =2
and Max.Z=10

3. Max.Z=10X1+15X2 4.Min.Z=3X1+2X2
St: St:
2X1+X2 < 26 5X1+X2 > 10
2X1+4X2< 56 X1 +X2 > 6
-X1+X2< 5 X1 + 4 X2 > 12
X1, X2 > 0 X1, X2 >0
Answer: X1=4 , X2 =2 Answer:X1=1,X2=5
and Max.Z=230 and Min.Z=13

40
Operations Research

5. Min.Z=-X1+2X2 6.Min.Z=20X1+10X2
St: St:
-X1+3X2 < 26 X1+2X2 <40
X1 +X2 < 6 3X1 + 4 X2 > 30
X1-X2< 2 4X1+ 3X2> 60
X1, X2 > 0 X 1, X2 >0
Answer:X1=2 , X2 =0 Answer:X1=6,X2=12 and Min.Z=-2 and Min.Z=240

II.A manufacturer produces two different models; X and Y, of the same product .The raw
materials r1 and r2 are required for production. At least 18 Kg of r1 and 12 Kg of r2 must be used
daily. Almost at most 34 hours of labor are to be utilized .2Kg of r1 are needed for each model X
and 1Kg of r1 for each model Y. For each model of X and Y, 1Kg of r2 is required. It takes 3
hours to manufacture a model X and 2 hours to manufacture a model Y. The profit realized is
$50 per unit from model X and $30 per unit from model Y. How many units of each model
should be produced to maximize the profit?
Answer: 10 units of model X, 2 units of model Y and the maximum profit is $ 560.
III.A manufacturing firm produces two machine parts P1 and P2 using milling and grinding
machines .The different machining times required for each part, the machining times available
on different machines and the profit on each machine part are as given below:
____________________________________________________________________
Manufacturing time Maximum time
required (min) available per week
(min)
Machine P1 P2
_____________________________________________________________________

Lathe 10 5 25,000
Milling Machine 4 10 2000
Grinding Machine 1 1.5 450
Profit per unit ($) $50 $100
_____________________________________________________________________
Determine the number of pieces of P1 and P2 to be manufactured per week to maximize profit.
Answer:X1=375 , X2 =50 and Max.Z=23750

41
Operations Research

IV.A person requires 10, 12 and 12 units of chemicals A, B and C respectively for his garden. A
liquid product contains 5, 2 and 1 units of A, B and C respectively per jar. A dry product contains
1, 2 and 4 units of A, B and C per carton. If the liquid product sells for $3 per jar and the dry
product sells $2 per carton, how many of each should be purchased in order to minimize cost and
meet the requirement?
Answer: 1 Unit of liquid, 5 units of dry product and Min.Z=$13
2.8 SIMPLEX METHOD

The graphical method to solving LPPs provides fundamental concepts for fully understanding
the LP process. However, the graphical method can handle problems involving only two decision
variables (say X1 and X2).

In 19940’s George B.Dantzig developed an algebraic approach called the Simplex Method
which is an efficient approach to solve applied problems containing numerous constraints and
involving many variables that cannot be solved by the graphical method.

The simplex method is an ITERATIVE or “step by step” method or repetitive algebraic approach
that moves automatically from one basic feasible solution to another basic feasible solution
improving the situation each time until the optimal solution is reached at.

Note:
The simplex method starts with a corner that is in the solution space or feasible region and
moves to another corner I the solution space improving the value of the objective function each
time until optimal solution is reached at the optimal corner.

 Matrices
Different types of matrices:
1. Row matrix
A matrix which has exactly one row
Example:
(1 2 3 4 ) ==>1x 4 matrixes
2. Column matrix
A matrix which has exactly one column

42
Operations Research

Example:
5

6 ==> 3x1 Matrix

3. Square matrix
A matrix in which the No of rows (m) = the No of columns (n)
Example:

1 2
==> 2x2 square Matrix
3 4

4. Null or Zero Matrix

-A matrix each of whose elements is zero.
Example:

0 0

0 0 ==> 3x2 Zero Matrix

0 0
5. Diagonal Matrix
The elements aij (i=j) are called diagonal elements of a square matrix.

Example:

1 2 3

4 5 6
The diagonal elements are: a11=1, a22=5 and a33=9
7 8 9

A square matrix of every element other than diagonal elements is zero, is called a diagonal
matrix.

43
Operations Research

Example:

1 0 0 0 0
0 0
0 5 0
0 5
0 0 9 ; and 0 0

6. Identity Matrix
A diagonal matrix whose diagonal elements are all equal to 1(unity)
Example:
1 0

Is an identity matrix
0 1

 Scope of solution of LPP by simplex method

Following types of problems are solved by simplex method:

 Maximize Z with inequalities of constraints in “< “form.

 Minimize Z with inequalities of constraints in “>“form.

 Maximize Z with inequalities of constraints in “>“form.

 Maximize Z or Minimize Z with inequalities of constraints in “< “, “>“or “=”form.

2.8.1 MAXIMIZATION PROBLEMS

 Maximize Z with inequalities of constraints in “< “form

Example:
Solve the problem using the simplex approach
Max.Z=300x1 +250x2

44
Operations Research

Subject to:
2x1 + x2 < 40 (Labor )
x1+3x2 < 45 (Machine)
x1 < 12 (Marketing)
x1, x2 > 0
Solution
Step 1
Formulate LPP Model

Step 2
Standardize the problem
i.e Convert constraint inequality into equality form by introducing a variable called Slack
variable.
Slack Variables:
A slack variable(s) is added to the left hand side of a < constraint to covert the constraint
inequality in to equality. The value of the slack variable shows unused resource.
A slake variable emerges when the LPP is a maximization problem.
Slack variables represent unused resource or idle capacity. Thus, they don’t produce any product
and their contribution to profit is zero.
Slack variables are added to the objective function with zero coefficients.
Let that s1, s2, and s3 are unused labor, machine and marketing hrs respectively.
Max.Z=300x1 +250x2 + 0 s1 +0 s2+ 0 s3
St:
2 x1+x2 + s1 +0 s2+ 0 s3= 40 Standard form

x1+3x2 +0s1 + s2+ 0 s3= 45

x1 + 0s1 + 0s2+ s3= 12
x1 , x2 , s1 , s2, s3 > 0
Step 3
Obtain the initial simplex tableau
To represent the data, the simplex method uses a table called the simplex table or the simplex
matrix.

45
Operations Research

==> In constructing the initial simplex tableau, the search for of the optimal solution begins at
the origin. Indicating that nothing can be produced;
Thus, first assumption, No production implies that x1 =0 and x2=0

==>2 x1+x2 + s1 +0 s2+ 0 s3= 40 ==> x1+3x2 +0 s1 + s2+ 0 s3= 45

2(0) +0 + s1 +0 s2+ 0 s3= 40 0 +3(0) + 0s1 + s2+ 0 s3= 45
s1= 40 – Unused labor hrs. s2= 45 – Unused machine hrs.

==> x1+0s1 +0s2+ s3= 12

0 +0s1 +0 s2+ s3= 12
s3= 12 – Unused Marketing hrs.
Therefore, Max.Z=300x1 +250x2 + 0 s1 +0 s2+ 0 s3
=300(0) +250(0) + 0(40) +0(45) + 0(12)
=0
Note:
In general, whenever there are n variables and m constraints (excluding the non-negativity),
where m is less than n (m<n), n-m variables must be set equal to zero before the solution can be
solved algebraically.
a. Basic variables are variables with non-zero solution values.
Or: basic variables are variables that are in the basic solution. Basic variables have 0
values in the Cj-Zj row.
b. Non-basic variables are variables with zero solution values.
Or: non-basic variables are variables that are out of the solution.
==>n=5 variables (x1 , x2, s1, s2, and s3) and m=3 constraints (Labor, machine and marketing
constraints), excluding non-negativity.
[
Therefore, n-m=5-3=2 variables(x1 and x2) are set equal to zero in the 1st simplex tableau. These
are non-basic variables. 3 Variables (s1, s2, and s3) are basic variables (in the 1st simplex tableau)
because they have non-zero solution values.

46
Operations Research

Step 4
Construct the initial simplex tableau
Initial simplex tableau

Slack variables
variables column

Solution quantity
Basic or Solution

Real or decision

columns
variable column
Profit per unit
column

column
Profit per unit row
Cj 300 250 0 0 0

SV X1 X2 S1 S2 S3 Q

0 S1 2 1 1 0 0 40 R1
Constraint
equation rows
0 S2 1 3 0 1 0 45 R2

0 S3 1 0 0 0 1 12 Gross Profit row R3

Zj 0 0 0 0 0 0
Net Profit row
C j - Zj 300 250 0 0 0 /Indicator row/

Step 5:
Choose the “incoming” or “entering” variables
Note:
The entering variable is the variable that has the most positive value in the Cj - Zj row also called
as indicator row. Or the entering variable is the variable that has the highest contribution to
profit per unit.
a. X1 in our case is the entering variable
b. The column associated with the entering variable is called key or pivot column ( X1 column
in our case )

47
Operations Research

Step 6
Choose the “leaving “or “outgoing” variable
==> In this step, we determine the variable that will leave the solution for X1 (or entering
variable)
Note:
 The row with the minimum or lowest positive (non-negative) replacement ratio shows the
variable to leave the solution.

Replacement Ratio (RR) = Solution Quantity (Q)

Corresponding values in pivot column

Note: RR>0
 The variable leaving the solution is called leaving variable or outgoing variable.
 The row associated with the leaving variable is called key or pivot row (s3 column in our case)
 The element that lies at the intersection of the pivot column and pivot row is called pivot
element(No 1 in our case)
Step 7
Repeat step 3-5 till optimum basic feasible solution is obtained.
Or: repeat step 3-5 till no positive value occurs in the Cj - Zj row.
Note:
 Divide each element of the pivot row by the pivot element to find new values in the key or
pivot row.
 Perform row operations to make all other entries for the pivot column equal to zero.
2nd simplex tableau

Cj 300 250 0 0 0

SV X1 X2 S1 S2 S3 Q

0 S1 0 1 1 0 -2 16 R’1=R1-2R3

0 S2 0 3 0 1 -1 33 R’2=R2-R3
300 X1 1 0 0 0 1 12 R’3=R3
Zj 300 0 0 0 300 3600 1
C j - Zj 0 250 0 0 -300

48
Operations Research

3rd simplex tableau

Cj 300 250 0 0 0

SV X1 X2 S1 S2 S3 Q

0 S1 0 0 1 -1/3 -5/3 5
R’’1=R’1-R’2
250 X2 0 1 0 1/3 -1/3 11
R’’2=R2/3
300 X1 1 0 0 0 1 12
R’’3=R’3
300 250 0 250/3
Zj 6350
650/3
0 0 0 -250/3
C j - Zj
- 650/3
Since all the Cj -
Zj < 0 optimal solution is reached at.
Therefore, X1=12, X2=11, S1=5 and Max Z=6350
Exercise:
A Juice Company has available two kinds of food Juices: Orange Juice and Grape Juice. The
company produces two types of punches: Punch A and Punch B. One bottle of punch A requires
20 liters of Orange Juice and 5 liters of Grape Juice.1 Bottle of punch B requires 10 liters of
Orange Juice and 15 liters of Grape Juice.

From each of bottle of Punch A a profit of $4 is made and from each bottle of Punch B a profit of
$3 is made .Suppose that the company has 230 liters of Orange Juice and 120 liters of Grape
Juice available
Required:
a. Formulate this problem as a LPP
b. How many bottles of Punch A and Punch B the company should produce in order to maximize
profit? (Using the simplex method)
c. What is this maximum profit?

49
Operations Research

Solution:
Juice needed for one bottle of
Juice Punch A Punch B Juice Available
_____________________________________________________________________
Orange Juice (lt) 20 10 230
Grape Juice (lt) 5 15 120
Profit per tent $4 $3
______________________________________________________________________________
Let X1= the No of bottles of punch A produced.
X2= the No of bottles of punch B produced.
 LPP Model
Max Z=4X1 +3X2
St:
20X1 +10X2 < 230 Orange Constraint
5X1 +15X2 < 120 Grape Constraint
X1 , X2 >0 Non-negativity constraint
 Standard form
Max.Z=4x1 +3x2 + 0 s1 +0 s2+ 0 s3
St:
20 x1+3x2 + s1 +0 s2 = 230
Standard form
5x1+15x2 +0s1 + s2+ = 120
x1 , x2 , s1 , s2, >0
Where, s1 =Unused orange juice
s2 =Unused grape juice
Initial simplex tableau

Cj 4 3 0 0

SV X1 X2 S1 S2 Q

0 S1 20 10 1 0 230

0 S2 5 15 0 1 120
0 0 0
Zj 0
0
4 3 0
C j - Zj
0

50
Operations Research

2nd simplex tableau

Cj 4 3 0 0

SV X1 X2 S1 S2 Q

4 X1 1 1/2 1/20 0 11.5

0 S2 0 25/2 -1/4 1 62.5

4 2 1/5
Zj 46
0
0 1 -1/5
C j - Zj
0 Optimal simplex tableau

Cj 4 3 0 0

SV X1 X2 S1 S2 Q

4 X1 1 0 3/50 -1/25 9

0 X2 0 1 -1/50 2/25 5
4 3 0.12
Zj 51
0.08
0 0 - 0.12
C j - Zj
-0.08
X1 = 9 bottles of punch A
X2= 5 bottles of punch B
s1 =0
s2 =0
MaxZ=$51

Exercise:
Solve the following LPPs using the simplex method
1. Max.Z=3x1 +5x2 2. Max.Z=20x1 +10x2
Subject to: Subject to:
x2 < 6 5x1+4x2 < 250
3x1+2x2 < 18 2x1+5x2 < 150
x1, x2 > 0 x1, x2 > 0

51
Operations Research

Answer: Answer:
x1=2, x2 =6, s1 =0 , s2=0 and MaxZ=$36 x1=50, x2 =0, s1 =0 , s2=50 and
MaxZ=$1,000

2.8.2 MINIMIZATION PROBLEMS

 Minimize Z with inequalities of constraints in “> “form
There are two methods to solve minimization LP problems:
1. Direct method/Big M-method/
 Using artificial variables
2. Conversion method
 Minimization by maximizing the dual

 Surplus Variable (-s):

 A variable inserted in a greater than or equal to constraint to create equality. It
represents the amount of resource usage above the minimum required usage.
 Surplus variable is subtracted from a > constraint in the process of converting the
constraint to standard form.
 Neither the slack nor the surplus is negative value. They must be positive or zero.
Example:
1. 2x1+x2 < 40 ==>is a constraint inequality
x1= 12 and x2= 11==> 2x1+x2+s = 40 ==>2(12)+11+s = 40
==> s=5 unused resource
2. 5x1+3x2 < 45
x1= 12 and x2= 11==> 5x1+3x2+s = 45 ==>5(12)+3(11)+s = 45
==> s=0 unused resource (No idle resource)
3. 5x1+2x2 >20
x1= 4.5 and x2= 2==> 5x1+2x2- s = 20 ==>5(4.5)+2(2)-s = 20
==> s=6 unused resource
4. 2x1+x2 >40
x1= 0 and x2= 0(No production)==> 5x1+2x2- s = 20 ==>5(4.5)+2(2)-s = 20
==> s=-6(This is mathematically unaccepted)
Thus, in order to avoid the mathematical contradiction, we have to add artificial variable (A)

52
Operations Research

 Artificial variable (A):

Artificial variable is a variable that has no meaning in a physical sense but acts as a tool to create
an initial feasible LP solution.

Note:
Type of constraint To put into standard form
< ---------------------------------------------Add a slack variable
= ---------------------------------------------Add an artificial variable
> ---------------------- Subtract a surplus variable and add artificial variable

1. Big M-method
/Charnes Penalty Method/
The Big-M Method is a method which is used in removing artificial variables from the basis .In
this method; we assign coefficients to artificial variables, undesirable from the objective function
point of view. If objective function Z is to be minimized, then a very large positive price (called
penalty) is assigned to each artificial variable. Similarly, if Z is to be maximized, then a very
large negative price (also called penalty) is assigned to each of these variables.

Following are the characteristics of Big-M Method:

a. High penalty cost (or profit) is assumed as M
b. M is assigned to artificial variable A in the objective function Z.
c. Big-M method can be applied to minimization as well as maximization problems with the
following distinctions:
i. Minimization problems
 Assign +M as coefficient of artificial variable A in the objective function
Z
ii. Maximization problems:
 Here –M is assigned as coefficient of artificial variable A in the objective
function Z
d. Coefficient of S (slack/surplus) takes zero values in the objective function Z
e. For minimization problem, the incoming variable corresponds to the highest negative value
of Cj-Zj.

53
Operations Research

f. Solution is optimal when there is no negative value of Cj-Zj.(For minimization case)

Example:
1. Minimize Z=25x1 +30x2
Subject to:
20x1+15x2 > 100
2x1+ 3x2 > 15
x1, x2 >0

Solution
Step 1
Standardize the problem
Minimize Z=25x1 +30x2 +0s1+0s2 +MA1+MA2
Subject to:
20x1+15x2- s1+A1 = 100
2x1+ 3x2 –s2+A2 = 15
x1, x2 , s1, s2 ,A1 ,A2 > 0

Step 2
Initial simplex tableau
The initial basic feasible solution is obtained by setting x 1= x2= s1= s2=0
No production, x1= x2= s1=0==>20(0) +15(0) - 0+A1 = 100 ==> A1 = 100
x1= x2= s2=0==>0(0)+3(0) - 0+A2 =15==> A2 = 15
Initial simplex tableau

Cj
25 30 0 0 M M
X1 X2 S1 S2 A1
SV Q RR
A2
20 15 -1 0 1
M A1 100
0
2 3 0 -1 0 100/20=5
15/2=7.5
M A2 15
1
22M 18M -M -M M
Zj 115 M
M
25 -22M 30- 18M M M 0
C j - Zj
0

54
Operations Research

Note:
Once an artificial variable has left the basis, it has served its purpose and can therefore be
removed from the simplex tableau. An artificial variable is never considered for re-entry into the
basis.
2nd Simplex Tableau

Cj 25 30 0 0 M

X1 X2 S1 S2
SV Q
A2
R’1=R1/20
1 3/4 -1/20 0
25 X1 5
0 R’2=R2-2 R’

0 3/2 1/10 -1
M A2 5
1
25 75/4+3/2M -5/4+1/10M -M 125+5
Zj
M M
0 45/4-3/2M 5/4-1/10 M M 3rd Simplex
C j - Zj
0 Tableau
Cj 25 30 0 0

X1 X2 S1
SV Q
S2
1 0 -1/10
25 X1 5/2
1/2
0 1 1/15
30 X2 10/3
-2/3
25 30 -1/2
Zj 162.5
-15/2
0 0 1/2
C j - Zj
15/2
Cj - Zj > 0==>Optimal
solution is reached
X1=5/2
X2=10/3 and MinZ=162.5

55
Operations Research

Note:
As long as an “A” variable is available in the solution variable column, the solution is infeasible.
2. Use the penalty (Big-M) method to solve the following LPP
Min Z=5x1 +3x2
Subject to:
2x1+4x2 < 12
2x1+ 2x2 = 10
5x1+ 2x2 > 10
x1, x2 >0
Solution
Min Z=5x1 +3x2 +0s1+0s2 +MA1+MA2
Subject to: If no production
2x1+4x2+s1 = 12 ==>x1 =x2=0==>s1=0 (Solution Value in the initial simplex tableau)
2x1+2x2 +A1 =10 ==>x1 =x2=0==>A1 =15 (Solution Value in the initial simplex tableau)
5x1+2x2 –s2 +A1=10 ==>x1=x2=s2=0==>A2=10(Solution Value in the initial simplex
tableau) x1, x2 , s1, s2 ,A1 ,A2 > 0

Initial Simplex tableau

Cj 5 3 0 0 M M

RR
SV X1 X2 S1 S2 A1 A2 Q

6
0 S1 2 4 1 0 0 0 12

M A1 2 2 0 0 1 0 10 5

M A2 5 2 0 -1 0 1 10 2

7M 4M 0 M M
Zj 20 M
M
5 -7M 3- 4M 0 -M 0
C j - Zj
0

56
Operations Research

2nd simplex tableau

Cj 5 3 0 0 M

SV X1 X2 S1 S2 A1 Q

0 S1 0 16/5 1 2/5 0 8

M A1 0 6/5 0 2/5 1 6
5 X1 1 2/5 0 -1/5 0 2
5M 6/5M +2 0 2/5M -1 10+6
Zj
M M
0 -6/5M +1 0 -2/5M+1
C j - Zj
0

3rd simplex tableau

Cj 5 3 0 0 M

SV X1 X2 S1 S2 A1 Q RR

3 X2 0 1 5/16 1/8 0 2.5 20

M A1 0 0 -3/8 1/4 1 3 12

5 X1 0 0 -1/8 -1/4 0 1 -

5 3 -3/8M +5/6 M/4-7/8

Zj 12.5+3 M
M
0 0 3/8M -5/6 -M/4+7/8
C j - Zj
0

57
Operations Research

4th Simplex tableau

Cj 5 3 0 0

SV X1 X2 S1 S2 Q

3 X2 0 1 1/2 0 1

0 S2 0 0 -3/2 1 12
5 X1 0 0 -1/2 0 4
Zj 5 3 -1 0 23
C j - Zj 0 0 1 0

X1=4, X 2=1, S1=0, S2=12 and MinZ=23

3. Use the penalty (Big-M) method to solve the following LPP
Max Z=2x1 +x2+3x3
Subject to:
2x1+ x2 + x3 < 5
2x1+ 3x2 +4x3 = 12
x1, x2 , x3 >0
Solution

Initial Simplex tableau

Cj 2 1 3 0 -M

RR
SV X1 X2 X3 S1 A1 Q

1 1 2 1 0 5 2.5
0 S1

-M A1 2 3 4 0 1 12
3
-2M -3M -4M 0
Zj -12 M
-M
2M+2 3M+1 4M+3 0
C j - Zj
0

58
Operations Research

2nd simplex tablea

Cj 2 1 3 0

SV X1 X2 X3 S1 A1 Q
RR

5
3 X3 1/2 1/2 1 1 0 5

-M A1 2 3 4 0 1 12 2

3/2 -M +3/2 3 2M+3/2

Zj -2M+15/2
-M
1/22 M-1/2 0 -2M-3/2
C j - Zj
0

3rd simplex tableau

Cj 2 1 3 0

RR
SV X1 X2 X3 S1 Q

6
3 X3 1/2 0 1 3/2 1.5

1 X2 0 1 0 -2 2 Not defined

Zj 3/2 1 3 5/2 13/2

C j - Zj 1/2 0 0 -5/2

4th simplex tableau

Cj 2 1 3 0

SV X1 X2 X3 S1 Q

3 X1 1 0 2 3 3

1 X2 0 1 0 -2 2
2 1 4
Zj 8
4
0 0 -1
C j - Zj
-4 Cj - Zj < 0==>optimal

59
Operations Research

solution
X1=3, X2 =2, X3=0, S1=0 and Max Z=8
Exercise
Find the optimal solution using simplex method
1. Min Z=10x1 +5x2
Subject to:
2x1 + 5x2 > 150
3x1+ x2 > 120
x1, x2 >0
Ans: x1=450/13, x2 =210/13 and Min Z=$540

2. Min Z=4x1 +5x2

Subject to:
x1 + 2x2 > 80
3x1+ x2 > 75
x1, x2 >0
Ans: x1=14, x2 =33 and Min Z=$221

3. Min Z=7x1 +9x2

Subject to:
3x1 + 6x2 > 36
8x1+ 4x2 > 64
x1, x2 > 0 Ans: x1=20/3, x2 =8/3 and Min Z=212/3
Note:
To get an initial feasible solution

Types of constraint Presence of variables in the initial solution mix

1. < (Slack) Yes
2.> *(Surplus) No
*(Artificial) Yes
3. = (Artificial) Yes

60
Operations Research

2.9 Some Complications and their Resolution

2.9.1 Mixed constraints
Example
Max Z=6x1 +8x2
Subject to:
x2 < 4
x1+ x2 = 9
6x1+ 2x2 >24
x1, x2 >0
 Standard form
Max.Z=6x1 +8x2 + 0 s1 +0 s2+ 0 s3-M A2- M A3
St:
x2 + s 1 =4
x1 + x2 + A2 =9
Standard form
6x1+2x2 - s3 + A3 =24
All Variables >0
Initial simplex tableau

Cj 6 8 0 0 -M -M

SV X1 X2 S1 S3 A2 A3 Q

0 S1 0 1 1 0 0 0 4

-M A2 1 1 0 0 1 0 9

-M A3 6 2 0 -1 0 1 4
-7M -3M 0 +M -M
Zj 24
-M
7M +6 3M+8 0 -M 0
C j - Zj
0

Ans: At the 4th tableau: X1 =5 ,X2 =4 ,S3 =14 and MaxZ=62

61
Operations Research

Note:
For the initial basis, use artificial variables for constraints that have them. Otherwise, use a
constraint slack variable. Hence, surplus variables will not appear in an initial solution.

2.9.2 Two incoming variables

/ Or Tie for entering variables/

In order to break this tie, the selection for the key column (entering variable) can be made
arbitrary. However; the number of solution can be minimized by adopting the following rules:
1. If there is a tie between two decision variables, then the selection can be made arbitrary.
2. If there is a tie between a decision variable and a slack (or surplus) variable, then select the
decision variable to enter into basis first.
3. If there is a tie between slack or surplus variable, then selection can be made arbitrary.
Example:
If the equation is max Z:

Cj
SV X1 X2 S1 S3 Q

Zj
5 2 5
C j - Zj
0

In such a case,X1 is the entering variable

2.9.3 Infeasibility
A situation with no feasible solution may exist if the problem was formulated improperly.

Infeasibility comes about when there is no solution that satisfies all of the problem’s constraints.

62
Operations Research

In the simplex method, an infeasible solution is indicated by looking at the final tableau .In it, all
Cj - Zj row entries will be the proper sign to imply optimality, but an artificial variable (A) will
still be in the solution mix.
Example:
Minimization case

Cj 5 8 0 0 M

SV X1 X2 S1 S2 A2 Q

5 X1 1 1 -2 3 0 200

8 X2 0 1 1 2 0 100

M A2 0 0 0 -1 1 20
5 8 -2 31-M
Zj 1,800+200M
M
0 0 2 M-31
C j - Zj
0
Even though
all Cj - Zj are positive or 0(i.e the criterion for an optimal solution in a minimization case), no
feasible solution is possible because an artificial variable (A2) remains in the solution mix.

2.9.4 Unbounded Solutions

No finite solution may exist in problems that are not bounded .This means that a variable can be
infinitely large without violating a constraint.
In the simplex method, the condition of unboundedness will be discovered prior to reaching the
final tableau. We will note the problem when trying to decide which variable to remove from the
solution mix.
The procedure in unbounded solution is to divide each quantity column number by the
corresponding pivot column number. The row with the smallest positive ratio is replaced. But if
the entire ratios turn out to be negative or undefined, it indicates that the problem is unbounded.

63
Operations Research

Example:
Maximization case

Cj 6 9 0 0

SV X1 X2 S1 S2 Q
RR

9 X2 -1 1 2 0 30 30/-1=-30
Unacceptable RRs

0 S2 -2 0 -1 1 10 10/-2=-5

-9 9 18
Zj 270
0
C j - Zj 15 0 -18
0

Pivot Column

The solution in the above case is not optimal because not all Cj - Zj entries are 0 or negative, as
required in a maximization problem. The next variable to enter the solution should be X1.To
determine which variable will leave the solution, we examine the ratios of the quantity column
numbers to their corresponding numbers in the X1 or pivot column. Since both pivot column
numbers are negative, an unbounded solution is indicated.

No unbounded solutions, no outgoing variable will exist.

5. Degeneracy
/Tie for leaving basic variable (key row)/
If there is a tie for the smallest ratio, this is a signal tat degeneracy exists. Degeneracy can occur
right in the first (initial tableau).This normally happens when the number of constraints is less
than the number of variables in the objective function. Problem can be overcome by trial and
error method.

64
Operations Research

Cj 5 8 2 0 0 0

SV X1 X2 X3 S1 S2 S3 Q
RR
8 X2 1/4 1 1 -2 0 0 10
10/1/4=40
0 S2 4 0 1/3 -1 1 0 20
20/4=5 Tie for the smallest ratio
0 S3 2 0 2 2/5 0 1 10 indicates degeneracy.
10/2=5
2 8 8 16 0
Zj 80
0
C j - Zj 3 0 -6 -16 0
0
Degeneracy could lead to a situation known as cycling, in which the simplex algorithm
alternatives back and forth between the same non-optimal solutions, i.e, it puts a new variable in,
then takes it out in the next tableau, puts it back in ,and so on.
One simple way of dealing with the issue is to select either row (S2 or S3 in this case) arbitrary. If
we are unlucky and cycling does occur, we simply go back and select the other row.
Remark
When there is a tie between a slack and artificial variable to leave the basis, the preference shall
be given to artificial variable to leave the basis and there is no need to apply the procedure for
resolving such cases.
6. Multiple Optimal Solutions
Multiple optimal solutions exist when non-basic variable contains zero on its Cj - Zj row.
Example:
Maximization problem

Cj 3 2 0 0
SV X1 X2 S1 S2 Q
2 X2 3/2 1 1 0 6
0 S2 1 0 1/2 1 3
3 2 2
Zj 12
0
C j - Zj 0 0 -2
0
65
Operations Research

MaxZ=3X1+2X2
X1=0, X2=6, S2=3 and MaxZ=12 or: X1=3, X2=3/2 and MaxZ=12
The Cj - Zj value of the Non-basic variable (X1) is 0.Thus, there is alternative optimal solution.
Exercise:
1. Solve the following LPP by the simplex algorithm
Min Z=6x1 +8x2
Subject to:
x1+ 2x2 > 80
3x1+ x2 > 75
x1, x2 >0
What are the values of the basic variables at each iteration?
Which are the non-basic variables at each iteration?
Ans:X1=14, X2=33, and MinZ=221
2. At the 3rd iteration of a particular LP maximization problem, the following tableau is
established:

Cj

SV X1 X2 X3 S1 S2 S3 Q

5 X3 0 1 1 -2 0 0 5

6 X1 1 -3 0 0 0 1 12
0 S2 0 2 0 1 1 -1 10
6 -13 5 5 0
Zj 97
21
C j - Zj 0 16 0 -5 0
-21

What special condition exists as you improves the profit and move to the next iteration? Proceed
to solve the problem for optimal solution
Ans: Degeneracy; X1=27, X2=5, X3=0, and MaxZ=$177

66
Operations Research

3. Covert the following constraints and objective function into the standard form for use in the
simplex method
Min Z=4x1 +x2
Subject to:
3x1+ x2 = 3
4x1+ 3x2 > 6
x1+ 2x2 < 3
x1, x2 >0
Answer:
Min.Z=4x1 +x2 + 0 s1 +0 s2+ M A1+M A3
St:
3x1+ x2 + A1 =3
4x1+ 3x2 -s1 +A2 =6
x1+ 2x2 + s2 =3
All Variables >0
4. Solve the following LPP using simplex method
MaxZ=9x1 +7x2
Subject to:
2x1+ x2 < 40
x1+ 3x2 < 30
x1, x2 >0
Ans: X1=18, X2=4, and MaxZ=$190
5. Solve the following LPP to show that it has alterative optimal solutions.
a. MaxZ=6x1 +3x2 Ans: i. X1=4, X2=0, and MaxZ=24
Subject to: ii.X1=5/2, X2=3, and MaxZ=24
2x1+ x2 < 8
3 x1+ 3x2 < 18
x2 < 3
x1, x2 >0

67
Operations Research

b MinZ=2x1 +8x2 Ans: i. X1=32/6, X2=10/6, and MinZ=24

Subject to: ii.X1=12, X2=0, and MinZ=24
5x1+ x2 > 10
2x1+ 2x2 > 14
x1+ 4x2 > 14
x1, x2 >0
6. Solve the following LPP to show that it has unbounded solution.
a. MaxZ=-2x1 +3x2
Subject to:
x1 <5
2 x1-3x2 < 6
x1, x2 >0

b. MaxZ=3x1 +6x2
Subject to:
3x1+ 4x2 > 12

-2x1+ x2 < 4
x1, x2 >0
7. Solve the following LPP to show that it has no feasible solution.
a. MaxZ=-2x1 +3x2 Ans: X1=2, X2=0, A1=2 and MaxZ=4-2M
Subject to:
x1 - x2 > 4
x1+ x2 < 6
x1 <2
x1, x2 > 0
b. MaxZ=3x1 +3x2 Ans: X1=0, X2=2, A2=2 and MaxZ=4-4M
Subject to:
2x1+ x2 < 2
3x1 + 4x2 > 12
x1, x2 >0

68
Operations Research

2.10 Duality in LPP

The term ‘dual’ in a general sense implies two or double. Every linear programming problem can
have two forms. The original formulation of a problem is referred to as its Primal form. The
other form is referred to as its dual LP problem or in short dual form. In general, however, it is
immaterial which of the two problems called primal or dual, since the dual of the dual is primal.
The dual involves setting up and solving an LP problem that is almost a ‘mirror image’ of an LP
problem that has been formulated. Both in its formulation and solution, the dual is the flip flop
version of the primal.

In the context of LP, duality implies that each LP problem can be analyzed in two different ways,
but having equivalent solution. For example, consider the problem of production planning. By
using the primal LP problem, the production manager attempts to optimize resource allocation by
determining quantities for each product to be produced that will maximize profit. But through
dual LP problem approach, he attempts to achieve production plan that optimizes resource
allocation so that each product is produced at that quantity such that its marginal opportunity cost
equals its marginal return. Thus, the main focus of dual is to find for each resource its best
marginal value or shadow price. This value reflects the scarcity of resources, i.e., the maximum
additional prices to be paid to obtain one additional unit of the resources to maximize profit
under the resource constraints. If resource is not completely used, i.e., there is slack, then its
marginal value is zero.

The shadow price is also defined as the rate of change in the optimal objective function value
with the respect to the unit change in the availability of a resource. Precisely for any constraint,
we have,

Shadow Price = change in the optimal objective function value

unit change in the availability of a resource

Analysis of the dual can also enable a manager to evaluate the potential impact of a new product,
and it can be used to determine the marginal values of resources (i.e. constraints). Relative to a
product, a manager would want to know what impact adding a new product would have on the
solution quantities and the profit; relative to resources, a manager can refer to a dual solution to

69
Operations Research

determine how much profit one unit of each resource equivalent to. Whereas the primal gives
solution results in terms of the amount of profit gained from producing products, the dual
provides information on the value of the constrained resources in achieving that profit.

? Often the manager is less concerned about profit than about the use of resource
in dual. Do you agree? Dear learner, would you discuss with your class mates.
___________________________________________________________________
___________________________________________________________________
__________________________________________________________________

3.1.1. Formulating the Dual

There are two important forms of primal and dual problems, namely the symmetrical (or
canonical) and the standard form.

Symmetrical form
Suppose the primal LP problem is given in the form
Maximize Zx = c1x1+ c2x2 + ... +cnxn
Subject to
a11x1+a12x2+... + a1nxn  b1
a21x1+a22x2+... + a2nxn  b2

am1x1+am2x2+... + amnxn  bm
x1,x2... xn  0
and then the corresponding dual LP problem is defined as:
Minimize Zy = b1y1+ b2y2 + ... +bmym
Subject to
a11y1+a21y2+... + am1ym  c1
a12y1+a22y2+... + am2ym  c2

a1ny1+an2y2+... + amnym  cn
y1,y2... ym  0

70
Operations Research

The above pair of LP problems can be expressed in the general LP model form as

Primal Dual

n m
Max Zx =  cjxj
j 1
Min Zy =  biyi
i 1

Subject to constraints Subject to constraints

n m

 aijxj  bi ; i = 1, 2,.., m
j 1
 ajiyi  cj ; i= 1,2, ..., n
i 1

aij = aji
and xj  0; j = 1,2,…,n and yi  0; j = 1,2,…,m

The following rules which guide the formulation of the dual problem will give you the summary
of the general relationship between primal and dual LP problems.
1. If the primal’ objective is to minimize, the dual’s will be to maximize; and the vice
versa
2. The coefficient’s of the primal’s objective function become the RHS values for the
dual’s constraints.
3. The primal’s RHS values become the coefficients of the dual’s objective function.
4. The coefficients of the first “row” of the primal’s constraints become the coefficients
of the first “column” of the dual’s constraint …..
5. The ≤ constraints become ≥ and the vice versa.
Consider this Primal problem:
Minimize 40x1 + 44x2 + 48x3
Subject to: 1x1 + 2x2 + 3x3 > 20
4x1 + 4x2 + 4x3 > 30
x1 , x 2 , x3 > 0

71
Operations Research

The dual of this problem is:

Maximise Z = 20y1 + 30y2
Subject to:
1y1 + 4y2 < 40
2y1 + 4y2 < 44
3y1 + 4y2 < 48
y1, y2 > 0
The following table shows how the primal problem is transformed into its dual.
Primal Dual
Objective Minimize 40x1 + 44x2 + 48x3 Maximize 20y1 + 30y2
function and Subject to Subject to
right-hand 1  20 1  40
side values 2  30 2  44
3  48
Primal Dual
Constraint 1 1x1 + 2x2 + 3x3  1 1y1 + 4y2 
coefficients 2 4x1 + 4x2 + 4x3  2 2y1 + 4y2 
3 3y1 + 4y2 

We can see from the table that the original objective was to minimize, whereas the objective of
the dual is to maximize. In addition, the coefficients of the primal’s objective function become
the right-hand side values for the dual’s constraints, whereas the primal’s right-hand side values
become the coefficients of the dual’s objective function.
Note that the primal has three decision variables and two constraints; whereas the dual has two
decision variables and three constraints.
The constraint coefficients of the primal are constraint coefficients of the dual, except that the
coefficients of the first “row” of the primal become the coefficients of the first “column” of the
dual, and the coefficients of the second “row” of the primal become the coefficients of the
second “column” of the dual.

72
Operations Research

When the primal problem is a maximization problem with all < constraints, the dual is a
minimization problem with all > constraints.
3.1.2. Formulating the Dual when the Primal has Mixed Constraints
In order to transform a primal problem into its dual, it is easier if all constraints in a
maximization problem are of the < variety, and in a minimization problem, every constraint is of
the > variety.
To change the direction of a constraint, multiply both sides of the constraints by -1. For
example,
-1(2x1 + 3x2 > 18) is -2x1-3x2 < -18
If a constraint is an equality, it must be replaced with two constraints, one with a < sign and the
other with a > sign. For instance,
4x1 + 5x2 = 20 will be replaced by
4x1 + 5x2 < 20
4x1 + 5x2 > 20
Then one of these must be multiplied by -1, depending on whether the primal is maximization or
a minimization problem.

EXAMPLE:
Formulate the dual of this LP model.
Maximize z = 50x 1 + 80x 2

Subject to:
C1 3x 1 + 5x 2 ≤ 45

C2 4x 1 + 2x 2 ≥16

C3 6x 1 +6x 2 = 30

x 1 ,x 2 ≥ 0

73
Operations Research

SOLUTION
Since the problem is a max problem, put all the constraints in to the ≤ form. Subsequently, C 2
and C3 will be first adjusted in to ≤ constraints.
- C 2 will be multiplied by -1:

-1(4x 1 + 2x 2 ≥16) becomes -4x 1 - 2x 2 ≤ -16

- C3 is equality, and must be restated as two separate constraints. Thus, it becomes:

6x 1 +6x 2 ≤ 30 and 6x 1 +6x 2 ≥30. Then the second of these must be multiplied by -1.

-1(6x 1 +6x 2 ≥30) becomes -6x 1 -6x 2 ≤ -30

After making the above adjustments, rewrite the LP model again.

Maximize z = 50x 1 + 80x 2

Subject to:
C1 3x 1 + 5x 2 ≤ 45

C2 -4x 1 - 2x 2 ≤ -16

C3 6x 1 +6x 2 ≤ 30

C4 -6x 1 -6x 2 ≤ -30

x1 , x 2 ≥ 0

? Dear learner, would you guess the reason why the RHS quantity consists of
negative values and do you think that it contradicts with rules of LP?
____________________________________________________________________
____________________________________________________________________
____________________________________________________________________

74
Operations Research

The dual of the above problem will be:

Minimize 45y1 - 16y2 + 30y3 – 30y4
Subject to :
C 1 3y1 -4y2 + 6y3 – 6y4 ≥ 50

C 2 3y1 - 2y2 + 6y3 – 6y4 ≥ 80

y1, y2, y3, y4 ≥ 0
 Comparison of the Primal and Dual Simplex Solutions
Cross -referencing the values in the primal and dual final simplex tableaus is shown as follows.

PRIMAL HOW LEBELED/ WHERE CORRESPONDENCE IN

FOUND IN THE PRIMAL THE DUAL
-Decision variable x 1 ,x 2 x 3 ,….. s1,s2,s3,…….

- slack variable s1,s2,s3,……. y1,y2,y3,……..

- shadow prices Z row under slack column Quantity column in decision
variable rows.
- solution quantities Quantity column
C-Z row under slack and
decision variable columns
Example:
To show that the flip-flopping of values between the primal and the dual carries over to their
final simplex tableaus, let us look at the following tables. The first table contains the final tableau
for the dual and the second one contains the final tableau for the primal.

Final tableau of Dual solution to the Microcomputer problem

C 100 22 39 0 0 M M Quantity
Basis y1 y2 y3 s1 s2 a1 a2
Primal
y3 39 16/3 0 1 1/3 -2/3 -1/3 2/3 40/3
shadow
y2 22 -6 1 0 -1 1 1 -1 10
prices
Z 76 22 39 -9 -4 9 4 740
C-Z 24 0 0 9 4 9-M 4-M

Primal solution quantities

75
Operations Research

Final tableau of Primal solution to the Microcomputer problem:

C 60 50 0 0 0 Quantity
Basis x1 x2 s1 s2 s3
s1 0 0 0 1 6 -16/3 24
x1 60 1 0 0 1 -1/3 9
x2 50 0 1 0 -1 2/3 4
Z 60 50 0 10 40/3 740
C-Z 0 0 0 -10 -40/3

The primary concern with a simplex solution is often three fold:

1. Which variables are in solution?
2. How much of each variable is in the optimal solution?
3. What are the shadow prices for the constraints?
Let’s consider how we can obtain the answers to these questions from the dual solution. Notice
that the solution quantities of the dual are equal to the shadow prices of the primal (i.e. 40/3 and
10). Next, notice that the values of the solution quantities of the primal (i.e., 24, 9, and 4) can be
found in the bottom row of the dual. Now, in the primal solution, s1 equals 24. In the dual, the
24 appears in the y1 column. The implication is that a slack variable in the primal solution
becomes a real variable (i.e., a decision variable) in the dual. The reverse is also true: A real
variable in the primal solution becomes a slack variable in the dual. Therefore, in the primal
solution we have x1 = 9 and x2 = 4; in the dual, 9 appears under s1 in the bottom row, and 4
appears under s2. Thus, we can read the solution to the primal problem from the bottom row of
the dual: In the first three columns of the dual, which equate to slack variables of the primal, we
can see that the fist slack equals 24 and the other two are zero. Under the dual’s slack columns,
we can read the primal values of the decision variables. In essence, then, the variables of the
primal problem become the constraints of the dual problem, and vice versa.
 Economic Interpretation of the Dual
Suppose the microcomputer firm has been approached by a representative of a department store
chain that wants the firm to make computers that will be sold under the store’s brand name. The
microcomputer company has only a limited capacity for producing computers, and therefore,
must, decide whether to produce its own computers or produce computers for the department
store.

76
Operations Research

For convenience, the original problem is repeated here:

x1 = number of Model I
x2 = number of Model II’s
Maximize 60x1 + 50x2
Subject to: Assembly time: 4x1 + 10x2 < 100
Inspection time: 2x1 + 1x2 < 22
Storage space: 3x1 + 3x2 < 39
x1, x2 >= 0

The manager of the firm would reason in the following way. For each unit of Model I that the
firm sacrifices to produce computers for the department store, it will gain 4 hours of assembly
time, 2 hours of inspection time, and 3 cubic feet of storage space, which can be applied to the
store computers. However, it will also give up a unit profit of $60. Therefore, in order for the
firm to realistically consider the store’s offer, the amounts of scarce resources that will be given
up must produce a return to the firm that is at least equal to the foregone profit. Hence, the value
of 4 assembly hours + 2 inspection hours + 3 cubic feet of storage space > $60. By similar
reasoning, giving up one unit of Model II will require that the value received by giving up 10
assembly hours + 1 inspection hour + 3 cubic feet of storage must equal or exceed the Model II
profit of $50 per unit. These, then, become the constraints of the dual problem. Thus:

Value Received Resources Minimum Profit

Per Unit of Freed Up Required
Model I 4y1 + 2y2 + 3y3 $60
Model II 10y1 + 1y2 + 3y3 $50
Hence the constraints of the dual refer to the value of capacity (i.e., the scarce resources). The
formulation indicates that in order to switch from making units of Models I and II to making
computers for the department store, the value received from that switch must be at least equal to
the profit foregone on the microcomputer models. The variables y1, y2 and y3 are the marginal
values of scarce resources (assembly time, inspection time, and storage space). Solving the dual
will tell us the imputed values of the resources given our optimal solution.

77
Operations Research

Naturally, the department store would want to minimize the use of the scarce resources, because
the computer firm almost certainly would base its charges on the amount of resources required.
Consequently, the objective function for the dual problem focuses on minimizing the use of the
scarce resources. Thus:

Minimize: 100y1 + 22y2 + 39y3

Looking at the optimal solution to the dual of the microcomputer problem, we can see the
marginal values of y2 and y3 in the Quantity column, but not the value of y1. This is because the
optimal solution to the primal did not completely use up all of the assembly capacity.
Consequently, no amount of either x1 or x2 would need to be given up to obtain one free hour of
assembly time. Thus, the marginal value of one hour of assembly time is $0.

Finally, the optimal dual solution always yields the same value of the objective function as the
primal optimal. In this case, it is 740. The interpretation is that the imputed value of the
resources that are required for the optimal solution equals the amount of profit that the optimal
solution would produce.

 Adding another Variable

The concepts introduced in the preceding section also can be applied to evaluating the feasibility
of adding another decision variable to the problem. For example, suppose the microcomputer
firm is considering a third model of microcomputer that will yield a profit of Birr 70 per unit,
and this will need resource requirements of 8 hours of assembly time, 4 hours of inspection time,
and 5 cubic feet of storage space per unit. The dual constraint for this model would be:
8y1 + 4y2 + 5y3 > 70
In order to determine if this new variable would come into the optimal (primal) solution, we can
substitute the dual solution of y1 = 0, y2 = 10, and y3 = 40/3 (i.e., the shadow prices, or marginal
values of the resources) into this constraint to see if it would be satisfied. Thus, we find, 8(0) +
4(10) + 5(40/3) = 106.67. Because this amount is greater than the new dual constraint, the
original solution remains optimal. Hence, the new variable (x3) would not come into solution.

78
Operations Research

Conversely, if the dual constraint has not been satisfied, the new variable would have come into
solution. In this instance, the marginal value of scarce resources that would be required for the
new model exceeds the marginal contribution to profit that the new model would provide. That
is, Birr106.67 > Birr70.

Note: The value of this approach is that it is not necessary to rework the entire problem in order
to test the potential impact that adding a new decision variable would have on the optimal
solution.
 Managerial Significance of Duality
The importance of the dual LP problem in terms of the information which it provides about the
value of the resources. The economic analysis is concerned with deciding whether or not to
secure more resource and how much to pay for these additional resources. The significance of
the study dual is as follows;
 The dual variables provide the decision maker a basis for deciding how much to pay for
additional unit of resources.
 The maximum amount that should be paid for one additional unit of a resource is called
its shadow price (also called simplex multiplier).
 The total marginal value of the resources equals the optimal objective function value. The
dual variables equal the marginal value of resources (shadow prices).
Note: The value of ith dual variable is the rate at which the primal objective function value will
increase as the ith resource increases, assuming that all other data is unchanged.
Advantages of Duality
 It is advantageous to solve the dual of a primal having less number of constraints,
because the number of constraints usually equals the number of iterations required to
solve the problem.
 It avoids the necessity of adding surplus or artificial variables and solves the problem
quickly.
 The dual variables provide an important economic interpretation of the final solution of
an LP problem.
 It is quite useful when investigating change in the parameters of an LP problem ( called
Sensitivity analysis)

79
Operations Research

Activity
Convert the following in to dual and solve it using simplex approach
1. Min Z= 4x1 +x2
Subject to:
3 x1+ x2 = 3
4x1+ 3x2 > 6
x1+ 2x2 < 3
x1,x2 > 0
2. Min Z=2x1 +8x2
Subject to:
5x1+ x2 > 10
2x1+ 2x2 > 14
x1+ 4x2 > 14
x1, x2 > 0
3.2. Sensitivity Analysis

? Dear learner, discuss with your group members about why sensitivity analysis is
conducted?
_________________________________________________________________________
_________________________________________________________________________
___________________________________________________________________

2.11 Sensitivity Analysis

As it is indicated above, sensitivity analysis is the study of sensitivity of the optimal solution of
an LP problem due to discrete variations (changes) in its parameters. The degree of sensitivity of
the solution due to these changes can range from no change to all to a substantial change in the
optimal solution of the given LP problem.
Sensitivity analysis carries the LP analysis beyond the determination of the optimal solution and
begins with the final simplex tableau. Its purpose is to explore how changes in any of the
parameters of a problem, such as the coefficients of the constraints, coefficients of the objective
function, or the right hand side values; would affect the solution. For this, instead of resolving
the entire problem as anew problem with new parameters, we may consider the original optimal
solution as an initial solution for the purpose of knowing the ranges, both lower and upper,
within which a parameter may assume a value.

80
Operations Research

For instance, a manager may want to consider obtaining an additional amount of scarce resource
that might be available, in which case the manager would want to know the answers to questions
such as:
b. Will an increase in the right hand side of this constraint affect the objective function?
c. If so, how much of the resource can be used?
d. Given the answer to the preceding question, what will be the revised optimal solution?
Conversely, the decision maker might be facing a situation in which the amount of scarce
resource available is less than the original amount, in which case the issues would be:
a. Will the decreased level of the resource have an impact on the value of the objective
function?
b. If yes, how much of an impact will it have, and what will be the revised optimal solution?
The process of studying the sensitivity of the optimal solution of an LP problem is also called
post optimality analysis because it is done after an optimal solution, assuming a given set of
parameters, has been obtained for the model.

The discussion here covers only two types of changes: changes in the right hand side levels of
the constraints, and changes in the coefficients of the objective function as change in the
coefficient of constraints is beyond the coverage of this material.
To demonstrate sensitivity analysis, the microcomputer example will again be used.
Maximize 60x1 + 50x2
Subject to
Assembly time 4x1 + 10x2 ≤ 100
Inspection time 2x1 + 1x2 ≤ 22
Storage space 3x1 + 3x2 ≤ 39
X1, X2 > = 0
The most obvious way to ascertain the effect of a change in the parameter of the model is to
make the change in the original model, resolve the model, and compare the solution result with
the original solution. However, resolving a problem can be very time consuming and as it will

81
Operations Research

be demonstrated below, it is unnecessary. In most cases the effect of changes in the model can be
determined directly from the final simplex tableau.
 A Change in the RHS of a constraint
The first step in determining how a change in the RHS (Right Hand Side) of a constraint (e.g.,
the amount of scarce resource that is available for use) would influence the optimal solution is to
examine the shadow prices in the final simplex tableau. Shadow prices are the values in the Z
row in the slack columns.
The final tableau for the microcomputer is shown in the following table since sensitivity analysis
starts from final tableau.
60 50 0 0 0 Quantity
Cj x1 x2 s1 s2 s3
Basis
s1 0 0 0 1 6 -16/3 24
x1 60 1 0 0 1 -1/3 9
x2 50 0 1 0 -1 2/3 4
Z 60 50 0 10 40/3 740
C-Z 0 0 0 -10 -40/3

Negatives of
shadow prices

A shadow price is a marginal value; it indicates the impact that a one-unit change in the amount
of a constraint would have on the values of the objective function. As we can see from the table,
the shadow prices are $0 for s1 (Assembly time), $10 for s2 (Inspection time), and $40/3 for s3
(Storage space). These tell us that an increase in the amount of assembly time would have no
effect on profit; if inspection time is increased by one hour, it will increase the profit by $10, and
if storage space is increased by 1 cubic foot, profit would increase by $40/3. The reverse also
holds. If we decrease them by such respective amounts, the decrease in profit will take the same
figure.

82
Operations Research

What shadow prices do not tell us is the extents to which the level of a scarce resource can be
changed and still have the same impact per unit. The ability to use additional amounts of a
resource will disappear at some point because of the fixed amounts of the other constraints.
Therefore, we need to determine the range over which we can change the right hand side
quantities and still have the same shadow price. This is called the range of feasibility, or the right
hand side range.

The key to computing the range of feasibility for the constraint lies in each slack column of the
final simplex tableau. To compute the range for each constraint, the entries in the associated
slack column must be divided into the values in the Quantity column. For example, for the
storage column values, the resulting ratios are:

24 9 4
-16/3 = -4.5 -⅓ = -27 ⅔ = +6

Here, the smallest positive ratio indicates how much the constraint level can be decreased before
it reaches the lower limit of its range of feasibility, and the smallest negative ratio (i.e. the
negative ratio closest to 0) indicates how much the storage constraint can be increased
before it reaches its upper limit of feasibility. The general rule applies when computing the upper
and lower limits on the range of feasibility for maximization problem is:

Allowable decrease: The smallest positive ratio

Allowable increase: Negative ratio closest to zero.

Note: The reverse rule will be true for minimization problems.

Alternative Method
Assume q1 symbolically represent quantity in the original model, and  stands for change.
Max. Z = 60X1+50X2
Subject to 4X1+10X2 <100
2X1+X2 < 22
3X1+3X2 < 39

83
Operations Research

If we consider the same example as in the previous example, i.e., change in q 3 = 39. If q3 changes
by some amount, the new q’3 equals the original q3 + . The  changes form a separate column
identical to the s1 column in the final simplex tableau. Therefore, to determine the  change, we
need only observe the slack (si) column corresponding to the model constraint quantity
(qi) being changed.
Final simplex tableau
60 50 0 0 0 Quantity
Cj x1 x2 s1 s2 s3
Basis
s1 0 0 0 1 6 -16/3 24-16/3 
x1 60 1 0 0 1 -1/3 9-1/3 
x2 50 0 1 0 -1 2/3 4 + 2/3 
Z 60 50 0 10 40/3 740
C-Z 0 0 0 -10 -40/3
Recall that the requirement of simplex method is that the quantity values can not be negative. If
any q1 value becomes negative, the solution no longer be feasible. Thus the inequalities
24-16 /3 > 0
9- /3> 0
4 + 2 /3 > 0
are solved for :
24-16 /3 > 0 9- /3> 0 4 + 2 /3 > 0
72-16  > 0 27-  >0 12+2  >0
3 3 3
72-16 > 0 27-  > 0 12+2  > 0

72>16  27 >  2  > - 12

4.5 >   < 27  > -6
< 4.5
Since the new q3 = 39 + , then  = q3 -39. These values are substituted to the above inequalities
to result in:

84
Operations Research

 < 4.5  < 27  > -6

q3 -39 < 4.5 q3 -39 < 27 q3 -39> -6
q3 < 43.5 q3 < 66 q3 > 33
Summarizing these inequalities, we have:
33 < 43.5< q3 <66
The value 66 can be eliminated since q3 must be less than 43.5; thus
33 < q3 < 43.5
As long as q3 remains in this range, the present basic solution variables remain positive and
feasible. However, the quantity value of those basic variables may change.

? Dear learner, can you determine the range within which q &q 1 2 vary with out causing
the basic solution mix change?
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________

Once we have computed these limits, we proceed to assess the impact that a contemplated
change would have on the optimal solution.
Example
The manager in the microcomputer problem is contemplating a change in the level of the storage
constraint – an increase of 3 cubic feet. Determine the revised optimal solution for the change.
Solution
First, note that an increase of 3 cubic feet is within the range. Then, the effect of an increase of
three cubic feet is computed in the following way:
Basis s3 Current solution Change Revised solution

s1 -16/3 24 3(-16/3)= -16 8

x1 -⅓ 9 3(-⅓) = -1 8
x2 ⅔ 4 3(⅔) = +2 6
Z 40/3 740 3(40/3) = +40 780
For an increase in storage space of 8 cubic feet, we take the upper limit of 4.5 cubic feet for the
computation. The resulting revised figures, then, would be s1= 0, x1=7.5, x2=7, and Profit =800.

85
Operations Research

Note, however, that beyond the upper limit, s3 would come into solution, replacing S1, which
would no longer be slack. The amount of slack would be 8- 4.5=3.5 cubic feet. Consequently, the
revised solution would be s3= 3.5, x1=7.5, x2=7, and Profit = 800.

 A Change in an Objective Function Coefficient

It is useful to know how much the contribution of a given decision variable to the objective
function can change without changing the optimal solution. There are two cases: Changes for a
variable that is not currently in the solution mix, and Changes for a variable that is currently in
the solution mix.
If a variable is not currently in the solution in a maximization problem, its objective function
coefficient would have to increase by an amount that exceeds the C-Z value for that variable in
the final tableau in order for it to end up as a basic variable in the optimal solution. The range
over which a non-basic variable’s objective function coefficient can change without causing that
variable to enter the solution mix is called its range of insignificance.
Range of Insignificance: the range over which a non-basic variable’s objective function
coefficient can change without causing that variable to enter the solution mix.
Maximize 3x 1 + 2x 2 + 5x 3

Subject to:
C1 x 1 + 2x 2 + 2x 3 ≤ 18

C2 3x 1 + 2x 2 + 6x 3 ≤ 12

C3 2x 1 + 3x 2 + 4x 3 ≤ 12

x 1 ,x 2 x 3 , ≥ 0

C 3 2 5 0 0 0
Basis x1 x2 x3 s1 s2 s3 Quantity

s1 0 0 0 0 1 1/5 -4/5 4/5

x1 3 1 0 2 0 3/5 -2/5 12/5
x2 2 0 1 0 0 -2/5 3/5 12/5
Z 3 2 6 0 1 0 12
C-Z 0 0 -1 0 -1 0

86
Operations Research

In the above tableau we see that x 3 is a non- basic variable. To determine the range over which

the objective function coefficient of x 3 would change without changing the optimal solution,

recall how a variable in a max problem enters the solution mix. Because x 3 is not in solution, its
objective function coefficient would need to increase in order for it to come in to solution (i.e., to
make its c-z value positive). The amount of increase must be greater than the absolute

value of its c-z value, which is │-1│. Therefore, its objective function coefficient must increase
by more than 1. Hence, the range of insignificance for x 3 is 6 or less (i.e, its Z value). In general

the rule is: Range of insignificance for a non- basic variable is the Z value of the variable. The
allowable increase is the absolute value of its own C- Z value.

Range of Optimality: the range over which objective function coefficient of a variable that is in
solution can change without changing the optimal values of the decision variables. Note,
however, that such a change would change the optimal value of the objective function.

For variables that are in solution, the determination of the range of optimality (the range over
which the objective function coefficient of a variable that is in the solution can change without
changing the optimal values of the decision variables) requires a different approach. Divide the
values in row C-Z by the corresponding row values of the variable in question and follow the
following rule for both maximization and minimization problems.

Allowable increase: The smallest positive ratio of C-Z value and the
variable’s substitution rate.
Allowable decrease: the smallest negative ratio of C-Z value and the
variable’s substitution rate.

Note: If there is no positive ratio, it means that there is no upper limit on that variable’s
objective function coefficient.
Example
Determine the range of optimality for the decision variables in the microcomputer problem.
Solution

87
Operations Research

The final simplex tableau for that problem is repeated here for convenience.

Cj 60 50 0 0 0
Basis x1 x2 s1 s2 s3 Quantity
s1 0 0 0 1 6 -16/3 24
x1 60 1 0 0 1 -1/3 9
x2 50 0 1 0 -1 2/3 4
Z 60 50 0 10 40/3 740
C-Z 0 0 0 -10 -40/3

For x1 we find:
Column x1 x2 s1 s2 s3
C-Z value 0/1 = 0 0/0 = Undefined 0/0 = Undefined -10/1 = -10 -40/3 = +40
x1 row value -1/3

The smallest positive ration is +40. Therefore, the coefficient of X1 can be increased by Birr 40
without changing the optimal solution. The upper end of its range of optimality is this amount
added to its current (original) value. Thus, its upper end is (Birr60 +Birr40) = Birr100. Also, the
smallest negative ratio is -10; therefore, the x1 coefficient can be decreased by as much as Birr10
from its current value, making the lower end of the range equal to (Birr60 - Birr10 )= Birr 50.
For x2 we find:
Column x1 x2 s1 s2 s3
C-Z value 0/0 = Undefined 0/1= 0 0/0 = Undefined -10/-1 = +10 -40/3 = -20
x2 row value ⅔
The smallest positive ration is +10. This tells us that the x2 coefficient in the objective function
could be increased by Birr10 to (Birr50 + Birr10) = Birr 60. The smallest negative ratio is -20,
which tells us that the x2 coefficient could be decreased by Birr20 to (Birr50 - Birr20) = Birr30.
Hence, the range of optimality for the objective function coefficient of x 2 is Birr30 to Birr 60.

88
Operations Research

This can also be done in the same way as change in the constraints by using the concept of
change starting from the final tableau. That is, the range of cj that will maintain the optimal
solution can be determined directly from the optimal simplex tableau.
Cj 60+  50 0 0 0
Basis x1 x2 s1 s2 s3 Quantity
s1 0 0 0 1 6 -16/3 24
x1 60+  1 0 0 1 -1/3 9
x2 50 0 1 0 -1 2/3 4
Z 60+  50 0 +10 (40- )/3 740
C-Z 0 0 0 -10-  (-40+ )/3

Consider a  change for c1. This will change the c1 value from c1=60 to c1=60+ . This new value
is not only included on the top cj row, but also in the left-hand cj column. Since 60+  is in left-
hand column, it becomes a multiple of the column values when the new Zj row values and the
subsequent Cj – Zj row values.
The solution becomes optimal as long as the Cj –Zj row values remain negative. If Cj-Zj
becomes positive, the product mix will change, if it becomes zero, there will be an alternative
solution. Thus, for the solution to remain uniquely optimal,
-10-  < 0
and
(-40+ )/3 <0
Both of these inequalities must be solved for .
-10-  < 0 and (-40+ )/3 <0
-  < 10 (-40+ )<0
 > -10 < 40
Now recall that c1=60+ ; therefore,  = c1-60. Now substituting c1-60 for  in the above
inequalities, we get:
 > -10 < 40
C1-60 > -10 c1-60 < 40
C1> 50 C1<100

89
Operations Research

Therefore, the range of values for C1 over which the solution basis remain optimal (although) the
value of the objective function may change) is:
50<C1<100

? Dear learner, would you try to determine the range over which coefficient of Xin 2
the objective function change without changing the solution mix in the same way as
above?
______________________________________________________________________
______________________________________________________________________

2.12 SUMMAR
1. The standard form of LP problem should have the following characteristics:
I. All the constraints should be expressed as equations by slack or surplus and/or
artificial variables
II. The right hand side of each constraint should be made non-negative; if it is not,
this should be done by multiplying both sides of the resulting constraint by -1.

Example:
2X1+3X2-4X3+X3<-50, we multiply both sides by negative
III. Three types of additional variables, namely
a. Slack Variable(S)
b. Surplus variable (-S), and
c. Artificial variables (A)
Are added in the given LP problem to convert it into standard form for two reasons:
ii. to convert an inequality to have a standard form of an LP model, and
iii. to get an initial feasible solution represented by the columns of an identity matrix.

The summery of the extra variables needed to add in the given LP problem to convert it into
standard form is given below:

90
 Operations Research

Coefficient of extra variables Presence of variables

Types of in the objective function in the initial solution
Extra variables to be added
constraint MaxZ mix
MinZ
< Add only slack variable 0 Yes
0
Subtract surplus variable and 0 No
> 0
Add artificial variable -M Yes
+M
= Add artificial variable -M Yes
+M
Table
2. Test of optimality
i. If all Cj - Zj < 0, then the basic feasible solution is optimal (Maximization case)
ii. If all Cj - Zj > 0, then the basic feasible solution is optimal (Minimization case)
3. Variable to enter the basis
i. A variable that has the most positive value in the Cj - Zj row
(Maximization case)
ii. A variable that has the highest negative value in the Cj - Zj row (Minimization case)
4. Variable to leave the basis: The row with the non-negative and minimum replacement ratio
(For both maximization and minimization cases i.e: RR > 0

91
Operations Research

CHAPTER THREE
TRANSPORTATION AND ASSIGNMETN PROBLEMS

3.1 Objective
3.2 Introduction
3.3 Transportation Models
3.4 Introduction
3.4.1 Formulation of the model
3.4.2 Method of Finding initial feasible Solution
3.4.3 Test for Optimality
3.4.4 Developing an Improved Solution
3.4.5 Variation in Transportation Problem
3.4.6 Maximization in Transportation Problems
3.5 Assignment Problem
3.6 Solution Method for Assignment
3.7 Special Case in Assignment Problem
3.8 Summary
3.9 Review Questions
3.1 Objectives
Dear learners, in the previous chapters we have discussed about linear programming, which is
more general. In this chapter we will discuss about special types of linear programming models
called Transportation and Assignment Models
3.2. Transportation Models
Sub section objectives
Up on completion of this section, the learner will be able to:
 Explain transportation models
 Identify application areas of transportation models
 Formulate LP model for transportation problems
 Describe methods of finding initial feasible solution
 Check for optimality of transportation model using stepping stone and MODI techniques.
 Elucidate special cases in transportation models

92
Operations Research

3.3 Introduction
Dear learner, in the previous chapter we have been thoroughly discussed about LP models and
their applications. In this chapter we will try to see special types of LP models which are
Transportation and Assignment Models. Though it is possible to solve the two problems using
simplex method, the process would result in rather large simplex tableaus and numerous simplex
iterations. Because of the unique characteristics of each problem, however, alternative solutions
methods requiring considerably less mathematical computation than the simplex method have
been developed.

The purpose of using an LP model for transportation problems is to minimize transportation

costs, taking into account the origin supplies, the destination demands, and the transportation
costs. The transportation method is similar in certain respects to the simplex technique because
both involve an initial feasible solution that is evaluated to determine if it can be improved.
Moreover, both involve displaying initial and improved solutions in a series of tableaus or tables.
However, the transportation method requires considerably less computational effort. In addition,
it is not unusual to discover that the initial feasible solution in a transportation problem is the
optimum. Examples of transportation problems include shipments from warehouses to retail
stores, shipment from factories to warehouses, shipments between departments within a
company, and production scheduling.

? Discuss similarities and differences between transportation models and simplex

approach?
___________________________________________________________________
___________________________________________________________________
___________________________________________________________________

3.3.1 FORMULATING THE MODEL

A transportation problem typically involves a set of sending locations, which are referred to as
origins, and a set of receiving locations, which are referred to as destinations. In order to
develop a model of a transportation problem, it is necessary to have the following information:
a. Supply quantity (capacity) of each origin.
b. Demand quantity of each destination.
c. Unit transportation cost for each origin-destination route.

93
Operations Research

Assumptions
The transportation algorithm requires the assumption that:
 All goods be homogeneous, so that any origin is capable of supplying any destination,
and
 Transportation costs are a direct linear function of the quantity shipped over any route.
 Though it will be modified later in our discussion, we shall add one additional
requirement that the total quantity available is equal to the total demand.
Example
Let’s consider an example. Harley’s Sand and Gravel Pit has contracted to provide topsoil for
three residential housing developments. Topsoil can be supplied from three different “farms” as
follows:
Farm Weekly capacity (cubic yards)
A 100
B 200
C 200
Demand for the topsoil generated by the construction projects is:
Project Weekly demand(cubic yards)
1 50
2 150
3 300

The manager of the sand and gravel pit has estimated the cost per cubic yard to ship over each of
the possible routes:
Cost per cubic yard to
From Project #1 Project #2 Project #3
Farm A Birr4 Birr 2 Birr 8
Farm B 5 1 9
Farm C 7 6 3
This constitutes the information needed to solve the problem.

94
Operations Research

The next step is to arrange the information into a transportation table. This is shown in the
following table.
Transportation table for Harley’s sand and gravel
To:
From: Project Project Project Supply
4 2 8
From: #1 #2 #3
Farm A 100
5 1 9
Farm B 200
7 6 3
Farm C 200

Demand 50 150 300

 Note: It is possible to write this transportation problem in the form of LPM.
Nine ( 3x3) decision variables are there and the objective function is minimization.
LPM for this problem is:
Zmin: 4x11 + 2x12 +8x13 + 5x21 + x22 + 9x23 + 7x31 + 6x32 + 3x33
Subject to:
x11 + x12 +x13  100
x21 + x22 + x23  200 Capacity/ Source constraint
x31 + x32 + x33  200

x11 + x21 + x3 1  50
x12 + x22 +x32  150 Demand/Destination constraint

x13 + x23 + x33 300
This can be rewritten as:
Zmin: 4x11 + 2x12 +8x13 + 5x21 + x22 + 9x23 + 7x31 + 6x32 + 3x33
Subject to:
x11 + x12 + x13  100
x21 + x22 + x23  200
x31 + x32 + x33  200
x11 + x21 + x31  50
x12 + x22 + x32  150
x13 + x23 + x33  300
x11 , x12, x13 ,x21, x22 , x23 , x31, x32 , x33  0

95
Operations Research

3.3.2 METHOD OF FINDING INTIAL FEASIBLE SOLUTION

The starting point of the transportation method is a feasible solution. For an assignment to be
feasible, two conditions must be fulfilled:
 A feasible solution is one in which assignments are made in such a way that all
supply and demand requirements are satisfied.
 The number of nonzero (occupied) cells should equal one less than the sum of the
number of rows and the number of columns in a transportation table. In the case of
a table with 3 rows and 3 columns, the number of occupied cells should be 3 + 3 -1
= 5 in order to be able to use the transportation algorithm. Sometimes, fewer
occupied or completed cells appear in a solution. When that happens, the solution

is referred to as a degenerated solution; such a solution requires modification in order to be able

to determine if it is optimal. This topic will be covered later on.
A number of different approaches can be used to find an initial feasible solution. Three of these
are described here:
The northwest-corner method.
An intuitive approach/Least cost method
Vogel’s / Penalty Method

 The Northwest-Corner Method

The northwest corner method is a systematic approach for developing an initial feasible solution.
Its chief advantages are that it is simple to use and easy to understand. Its chief drawback is that
it does not take transportation costs into account. Consequently, such a solution may require
much additional effort to obtain the optimal solution.

The northwest corner method gets its name because the starting point for the allocation process is
the upper left-hand (Northwest) corner of the transportation table. For the Harley problem, this
would be the cell that represents the route from Farm A to Project #1. The following set of
principles guides the allocation:

96
Operations Research

i. Begin with the upper left-hand cell, and allocate as many units as possible to that
cell. This will be the smaller of the row supply and the column demand. Adjust
the row and column quantities to reflect the allocation.
ii. Remain in a row or column until its supply or demand is completely exhausted or
satisfied, allocating the maximum number of units to each cell in turn, until all
supply has been allocated (and all demand has been satisfied because we assume
total supply and demand are equal).

Initial Feasible Solution for Harley using Northwest-corner method :

To:
Project Project Project Supply
From: #1 #2 #3
4 2 8
Farm A 50 50 100
(first) (second)
5 1 9
Farm B 100 100 200
(third) (fourth)
7 6 3
Farm C 200 200
(last)

Demand 50 150 300 500

The total cost is found by multiplying the quantities in “completed” (i.e. nonempty) cells by the
cell’s unit cost and, then, summing those amounts. Thus:
Total cost = 50(4) + 50(2) + 100(1) + 100(9) + 200(3) = $1900
As noted earlier, the main drawback of the northwest-corner method is that it does not consider
cell (route) costs in making the allocation. Consequently, if this allocation is optimal, that can be
attributed to chance rather than the method used.

97
Operations Research

 The Intuitive Approach

This approach, also known as the minimum-cost method, uses lowest cell cost as the basis for
selecting routes. The procedure is as follows:
i. Identify the cell that has the lowest unit cost. If there is a tie, select one arbitrarily.
Allocate a quantity to this cell that is equal to the lower of the available supply for the
row and the demand for the column.
ii. Cross out the cells in the row or column that has been exhausted (or both, if both have
been exhausted), and adjust the remaining row or column total accordingly.
iii. Identify the cell with the lowest cost from the remaining cells. Allocate a quantity to
this cell that is equal to the lower of the available supply of the row and the demand
for the column.
iv. Repeat steps (ii) and (iii) until all supply and demand have been exhausted.
The initial feasible solution for the Harley problem completed using the above steps is shown
below.
Initial Feasible Solution for the Harley problem using the Intuitive approach
To:
Project Project Project Supply
From: #1 #2 #3
4 2 8
Farm A 50 50 100

5 1 9
Farm B 150 50 200

7 6 3
Farm C 200 200

Demand 50 150 300 500

98
Operations Research

We can easily verify that this is a feasible solution by checking to see that the row and column
totals of the assigned cell quantities equal the supply and demand totals for the rows and
columns. Now let us compute the total cost of this solution and compare it to that of the
northwest corner solution.
Total cost = 50(4) + 50(8) + 150(1) + 50(9) + 200(3) = $1800

Compared to the plan generated using the Northwest-corner method, this one has a total cost that
is $100 less. This is due to the fact that the previous one did not involve the use of cost
information in allocating units.

4.1.2.3. Vogel’s Approximation Method (VAM)

The third method for determining an initial solution, Vogel’s Approximation Method (also called
VAM), is based on the concept of penalty cost or regret. If a decision maker incorrectly chooses
from several alternative courses of action, a penalty may be suffered (and the decision maker
may regret the decision that was made). In transportation problem, the courses of action are the
alternative routes and a wrong decision is allocating to a cell that does not contain the lowest
cost.

This method is preferred over the other two methods because the initial feasible solution
obtained with VAM is either optimal or very close to the optimal. With VAM the basis of
allocation is unit cost penalty i.e. that column or row which has the highest unit cost penalty
(difference between the lowest and the next highest cost) is selected first for allocation and the
subsequent allocations in cells are also done keeping in view the highest unit cost penalty.
Steps in VAM
1. Construct the cost, requirement, and availability matrix i.e. cost matrix with column and row
information.
2. Compute a penalty for each row and column in the transportation table. The penalty is merely
the difference between the smallest cost and the next smallest cost element in that
particular row or column.
3. Identify the row and column with the largest penalty. In this identified row (column), choose
the cell which has the smallest cost and allocate the maximum possible quantity to this cell.
Delete the row (column) in which capacity (demand) is exhausted. When there is a tie for

99
Operations Research

penalty, select one arbitrarily. After allocation, cross that row or column and disregard it
from further consideration.
4. Repeat steps 1 to 3 for the reduced table until the entire capabilities are used to fill the
requirement at different warehouses.
5. From step 4 we will get initial feasible solution. Now for initial feasible solution find the
total cost.

The solution for our problem is as follows:

To: Penalty
From: Project Project Project Supply Hc NHC Hc- NHC
4 2 8
#1 #2 #3
Farm A 100
5 1 9 8 4 4
150
Farm B 200
1
7 6 3
Farm C 200 9 5 4 selected

Demand 50 150 300 7 6 1

HC 7 6 9 First allocation
NHC 5 2 8
Penalt 2 4 1
HC: Highest cost
NHC: Next Highest cost

100
 Operations Research

Table2 Penalty

To: Hc NHC Hc- NHC

From: Project Project Project Supply

4 2 8
#1 #2 #3
8 4 Farm A 100 4
5 1 9
150
Farm B 200 50
9 5 7 6 3 4
200
Farm C 2 200

7 3 Demand 50 150 300 100 4 selected

Penalty 2 - 1 Second allocation

Table 3

Third To:
allocation Penalty
From: Project Project Project Supply
4 2 8
#1 50 #2 #3
4 Farm A 3 100 50 selected
5 1 9
150
Farm B 200 50
7 6 3
4
200
Farm C 200

Demand 50 150 300 100

Penalty 1 - 1

101
Operations Research

Table 4
Since there is no penalty for the remaining cells, we allocate for these cells according to their
cost.

To:
From: Project Project Project #3 Supply
4 2 4 8
#1 50 #2 50
Farm A 100
5 1 50 9 50
5
150
Farm B 200 Fourth allocation
7 6 3
50
200
Farm C 200

Demand 50 150 300 100 50

Fifth allocation
Therefore the final allocation will be:

To:
From: Project Project Project #3 Supply
4 2 8
#1 50 #2 50
Farm A 100
4
5 1 50 9
150
Farm B 200
7 6 3
200
Farm C 200

5
Demand 50 150 300

102
Operations Research

Total cost:
1x150 + 3x200+4x50 + 8x50 + 9x50
= Birr 1800

? Dear learners, could you make comparison among costs of the three
approaches? Why do you think cost of North West corner method has the
highest cost?
_______________________________________________________________
_______________________________________________________________
_______________________________________________________________

3.3.3 TEST for Optimality

The test for optimality for a feasible solution involves a cost evaluation of empty cells (i.e.,
routes to which no units have been allocated) to see if an improved solution is possible. We shall
consider two methods for cell evaluation:
The Stepping-stone method
The MODI method
The stepping-stone method involves a good deal of more effort than the MODI method.
However, it provides an intuitive understanding of the evaluation process. Moreover, when a
solution is not optimal, the distribution plan must be revised by reallocating units into and out of
various cells, and only the stepping-stone method can be used for the reallocation.

 The Stepping-stone method

The Stepping-stone method involves tracing a series of closed paths in the transportation table,
using one such path for each empty cell. The path represents a shift of one unit into an empty
cell, and it enables the manager or analyst to answer a “what-if” question: What impact on total
cost would there be if one unit were shifted into an unused route? The result is a cost change per
unit shifted into a cell. If the shift would result in a cost savings, the stepping-stone path also can
be used to determine the maximum number of units that can be shifted into the empty cell, as
well as modifications to other completed cells needed to compensate for the shift into the
previously unused cell.
Reconsider the initial feasible solution we found using the northwest-corner method. Only the
unoccupied cells need to be evaluated because the question at this point is not how many units to
allocate to a particular route but only if converting a cell from zero units to nonzero (a positive

103
Operations Research

integer) would decrease or increase total costs. The unoccupied cells are A-3, B-1, C-1, and C-2.
They must be evaluated one at a time, but in no particular order.
Rules for tracing Stepping-stone paths:
i. All unoccupied cells must be evaluated. Evaluate cells one at a time.
ii. Except for the cell being evaluated, only add or subtract in occupied cells. (It is
permissible to skip over unoccupied cells to find an occupied cell from which the path
can continue.)
iii. A path will consist of only horizontal and vertical moves, starting and ending with the
empty cell that is being evaluated.
iv. Alter + and – signs, beginning with a + sign in the cell being evaluated.

 Note that it is not necessary to actually alter the quantities in the various cells to reflect the
one-unit change, the + and – signs suffice.

The general implication of the plus and minus signs is that cells with + signs mean one unit
would be added, cells with a – sign indicate one unit would be subtracted. The net impact of such
a one-unit shift can be determined by adding the cell costs with signs attached and noting the
resulting value. Thus, for cell B-1, we have a net change of +2 (i.e., 5+2-4-1) which means that
for each unit shifted into cell B-1, the total cost would increase by $2.
Computed in similar way, the evaluations of cells C-1, A-3, and C-2 result in +10, -2, and +11
respectively. The negative value for cell A-3 indicates an improved solution is possible: For each
unit we can shift into that cell, the total cost will decrease by $2. The following table shows how
empty cell C-1 can be evaluated using the Stepping stone method.

104
Operations Research

Evaluation path for cell C-1

To:
Project Project Project Supply
From: #1 #2 #3
4 2 8
Farm A 50 50 100
–- +
5 1 9
Farm B 100 100 200
– +
7 6 3
Farm C + – 200 200

Demand 50 150 300 500

 The MODI method

The MODI (Modified Distribution) method of evaluating a transportation solution for optimality
involves the use of index numbers that are established for the rows and columns. These are based
on the unit costs of the occupied cells. The index numbers can be used to obtain the cell
evaluations for empty cells without the use of stepping-stone paths.

There is one index number for each column and one for each row. These can be conveniently
displayed along the left and upper edges of a matrix. The index numbers are determined in such a
way that for any occupied cell, the sum of the row index and the column index equal the cell’s
unit transportation cost:

Row index + Column index = Cell cost

ri + kj = cij

105
Operations Research

The index numbers are determined sequentially in a manner dictated by the position of occupied
cells. The process always begins by assigning a value of zero as the index number of row 1.
The method will be illustrated by developing index numbers for the initial feasible solution for
the Harley problem generated by the northwest-corner method. We begin assigning a value of
zero for row 1. Once a row index has been established, it will enable us to compute column index
numbers for all occupied cells in that row. Similarly, once a column index number has been
determined, index numbers for all rows corresponding to occupied cells in that column can be
determined. The complete set of row and column index numbers is shown in the following table.
Cell evaluations for Northwest Corner solution for the Harley Problem using the MODI method
k1 k2 k3
+4 +2 +10
To:
Project Project Project Supply
From: #1 #2 #3
4 2 8
r1 0 Farm A 50 50 100
-2

5 1 9
r2 -1 Farm B +2 100 100 200

7 6 3
+10 +11
r3 -7 Farm C 200 200

Demand 50 150 300 500

The cell evaluations (improvement potentials) for each of the unoccupied cells are determined
using the relationship:

Cell evaluation = Cell cost – Row index – Column index

eij = cij – ri – kj

106
Operations Research

For example, the cell evaluations for A-3 is 8 – 0 – 10 = -2. Similarly, the evaluation for B-1 is
+2, for C-1, +10, and for C-2, it is +11. Note that they agree with the values we computed earlier
using the stepping-stone method.

When cell evaluations are positive or zero, an optimal solution has been found. If one or more is
negative, the cell with the largest negative should be brought into solution because that route has
the largest potential for improvement per unit. In this case, we found that cell A-3 had an
evaluation of –2, which represented an improvement potential of $2 per unit. Hence, an
improved solution is possible.

3.3.4 DEVELOPING AN IMPROVED SOLUTION

Developing an improved solution to a transportation problem requires focusing on the

unoccupied cell that has the largest negative cell evaluation. Improving the solution involves
reallocating quantities in the transportation table. More specifically, we want to take advantage
of the improvement potential of cell A-3 by transferring as many units as possible into that cell.
The stepping-stone path for that cell is necessary for determining how many units can be
reallocated while retaining the balance of supply and demand for the table. The stepping-stone
path also reveals which cells must have quantity changes and both the magnitude and direction
of changes. The + signs in the path indicate units to be added, the – signs indicate units to be
subtracted. The limit on subtraction is the smallest quantity in a negative position along the cell
path. With each iteration (new solution), it is necessary to evaluate the empty cells to see if
further improvement is possible. This requires use of either the MODI or the Stepping-stone
method. Both will yield the same values.

After reallocating the units using the stepping stone method, the empty cells will be A-2, B-1, C-
1 and C-2. Suppose we use the MODI method for evaluation. After assigning all the row and
column indices, we find the cell evaluations to be as follows:
Cell A-2: 2 – 0 – 0 = +2
Cell B-1: 5–1–4= 0
Cell C-1: 7-(-5)- 4 = +8
Cell C-2: 6-(-5)- 0 = +11

107
Operations Research

Because none of these numbers is negative, this is an optimal solution. You may recall that this
was the same solution obtained using the intuitive method for the initial feasible solution. At that
point, it was determined that the total cost for the distribution plan was $1800.

Optimal Solution for the Harley problem

+4 0 +8
To:
Project Project Project Supply
From: #1 #2 #3
4 2 8
+2
0 Farm A 50 50 100

0
5 1 9
+1 Farm B +8 150
+11 50 200

7 6 3
-5 Farm C 200 200

Demand 50 150 300 500

? Dear learner, which of the two approaches for improvement is lengthy and
difficulty to use?
___________________________________________________________________
___________________________________________________________________
___________________________________________________________________

108
Operations Research

3.3.5 Variation in Transportation

There are a number of special issues in relation to the transportation model. They are:
i. Determining if there are alternate optimal solutions.
ii. Defining and handling degeneracy (too few occupied cells to permit evaluation of a
solution).
iii. Avoiding unacceptable or prohibited route assignments.
iv. Dealing with problems in which supply and demand are not equal.
v. Solving maximization problems.

 Alternate Optimal Solutions

Sometimes, transportation problems have multiple optimal solutions. In such instances, it can be
useful for a manager to be aware of alternate solutions, because this gives the manager an option
of bringing non-quantitative considerations into the decision.
In the case of the transportation problem, the existence of an alternate solution is signaled by an
empty cell’s evaluation equal to zero. In fact, you may have noted that cell B-1 had an evaluation
equal to zero in the final solution of the Harley problem. We can find out what that alternate
solution is by reallocating the maximum number of units possible around the stepping-stone path
for that cell. After the cell evaluation of, and reallocating units to cell B-1, we find the following
table to be an alternate optimal solution.
Alternate optimal solution for the Harley problem
To:
Project Project Project Supply
From: #1 #2 #3
4 2 8
Farm A 100 100
5 1 9
Farm B 50 150 200

7 6 3
Farm C 200 200

Demand 50 150 300 500

109
Operations Research

 Degeneracy
A solution is degenerate if the number of occupied cells is less than the number of rows plus the
number of columns minus one. i.e., there are too few occupied cells to enable all the empty cells
to be evaluated. In the case of the stepping-stone method, this means that there will be at least
one empty cell for which an evaluation path cannot be constructed. For the MODI method, it
means that it will be impossible to determine all of the row and column index numbers.
Obviously, some modification has to be made to determine if such a degenerate solution is
optimal. The modification is to treat some of the empty cells as occupied cells. This is
accomplished by placing a delta () in one of the empty cells. The delta represents an extremely
small quantity (e.g., 0.001 units); it is so small that supply and demand for the row and column
involved will be unaffected even without modifying other quantities in the row or column, and so
small that total cost will not change.

The purpose of the delta is to enable evaluation of the remaining empty cells. The choice of
location for the delta can be somewhat tricky: some empty cells may be unsuitable if they do not
enable evaluations of the remaining empty cells. Moreover, the delta cannot be placed in a cell
which later turns out to be in a negative position of a cell path involved in reallocation because
delta will be the “smallest quantity in a negative position” and shifting that minute quantity
around the cell path will leave the solution virtually unchanged. Consequently, a certain amount
of trial and error may be necessary before a satisfactory location can be identified for delta.

The technique can be demonstrated for the degenerate alternate solution of the Harley problem.
Suppose that after some experimentation, cell A-1 has been selected for the location of delta. The
resulting index numbers generated using MODI and the improvement potential for empty cells
based on delta in cell A-1 are shown in the following table. This confirms that the solution is
optimal Harley Alternate Solution Modified for Degeneracy.

110
Operations Research

+4 0 +8
To:
Project Project Project Supply
From: #1 #2 #3
4 2 8
+2
0 Farm A Δ 100 100
5 1 9
0
+1 Farm B 50 150 200

7 6 3
-5 Farm C +8 +11 200 200

Demand 50 150 300 500

 Unacceptable Routes
In some cases, an origin-destination combination may be unacceptable. This may be due to
weather factors, equipment breakdowns, labor problems, or skill requirements that either
prohibit, or make undesirable, certain combinations (routes).

Suppose that in the Harley problem route A-3 was suddenly unavailable because of recent
flooding. In order to prevent that route from appearing in the final solution (as it originally did),
the manager could assign a unit cost to that cell that was large enough to make that route
uneconomical and, hence, prohibit its occurrence. One rule of thumb would be to assign a cost
that is 10 times the largest cost in the table (or a very big +M). Then, this revised problem could
be solved using either of the methods we have discussed earlier. Note that the prohibited route
may appear in a non-optimal solution, but it will be eliminated by the time the optimal solution is
reached.

111
Operations Research

Dear learner, would you find the optimal solution if route A-3 is not available? How do you
compare the cost with the original cost?

 Unequal Supply and Demand

Up to this point, examples have involved cases in which supply and demand were equal. As you
might guess, there are situations in which the two are not equal. When such a situation is
encountered, it is necessary to modify the original problem so that supply and demand are equal.
This is accomplished by adding either a dummy column or a dummy row; a dummy row

is added if supply is less than demand and a dummy column is added if demand is less than
supply. The dummy is assigned unit costs of zero for each cell, and it is given a supply (if a row)
or a demand (if a column) equal to the difference between supply and demand. Quantities in
dummy routes in the optimal solution are not shipped. Rather, they serve to indicate which
supplier will hold the excess supply, and how much, or which destination will not receive its
total demand, and how much it will be short.

Let’s consider an example. Suppose that Farm C in the Harley problem has experienced an
equipment breakdown, and it will be able to supply only 120 cubic yards of topsoil for a period
of time. Therefore, total supply will be 80 units less than total demand. This will require adding
a dummy origin with a supply of 80 units. The final solution is shown in the following table.
We interpret the solution indicating that Project #3 will be short by 80 units per week until the
equipment is repaired. Note, though, that this analysis has considered only transportation costs,
and that other factors, such are shortage costs or schedules of the projects, may dictate some
other course of action.

If the intuitive approach is used to obtain the initial feasible solution when a dummy is involved,
make assignments to the dummy last. Hence, begin by assigning units to the cell with the lowest
nonzero cost, then the next lowest nonzero cost, and so on. For the Harley problem this would
mean that units would be assigned first to cell B-2 because its cost of $1 is the lowest nonzero
cell cost.

112
Operations Research

Solution using the Dummy Origin

To:
Project Project Project Supply
From: #1 #2 #3
4 2 8
Farm A 50 50 100

5 1 9
Farm B 150 50 200

7 6 3
Farm C 120 120

0 0 0
Dummy 80 80

Demand 50 150 300 500

? Dear learner, would you complete the above case by evaluating the
assignments? What would assignments in the dummy row or column in the
optimal solution indicate?
_________________________________________________________________
_________________________________________________________________
_________________________________________________________________
______________________________

3.3.6 Maximization Transportation Problems

Some transportation type problems concern profits or revenues rather than costs. In such cases,
the objective is to maximize rather than to minimize. Such problems can be handled by adding
one additional step at the start: Identify the cell with the largest profit and subtract all the other

113
Operations Research

cell profits from that value. Then replace the cell profits with the resulting values. These values
reflect the opportunity costs that would be incurred by using routes with unit profits that are less
than the largest unit profit. Replace the original unit profits with these opportunity cost solution.
This will be identical to maximizing the total profit. For example, suppose in the Harley
problem, the cell values had been unit profits instead of unit costs. Cell B-3 had the largest dollar
value: $9. Hence, each cell’s dollar amount would be subtracted from 9. For cell A-1, the
resulting opportunity cost would have been 9-4 = 5 and so on. Cell B-3 would have an
opportunity cost of 0 making it the most desirable route.

The remainder of the steps for developing an initial feasible solution, evaluation of empty cells,
and reallocation are identical to those used for cost minimization. When the optimal distribution
plan has been identified, use the original cell values (i.e., profits) to compute the total profit for
that plan.

? Dear learner, would you complete the above case by evaluating the assignments?
Which one would you consider the original price or opportunity cost of each cell to
calculate the total profit in the optimal solution?
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________

ACTIVITY
1. A transportation problem involves the following costs, supply, and demand.
To
From 1 2 3 4 Supply
1 $5 750 300 450 12
2 65 800 400 600 17
00
040
3 700 500 550 11
0
Demand 10 10 10 10
b. Find the initial solution using the northwest corner method, the minimum cell cost
method, and Vogel's Approximation Method. Compute total cost for each.
c. Using the VAM initial solution, find the optimal solution using the modified
distribution method (MODI).

114
Operations Research

4 Oranges are grown, picked, and then stored in warehouses in Tampa, Miami, and Fresno.
These warehouses supply oranges to markets in New York, Philadelphia, Chicago, and
Boston. The following table shows the shipping costs per truckload ($100s), supply, and
demand.
To
From New York Philadelphia Chicago Boston Supply
Tampa 9 14 12 17 200
Miami 11 10 6 10 200
Fresno 12 8 15 7 200
Demand 130 170 100 150
Because of an agreement between distributors, shipments are prohibited from Miami to
Chicago.
a.Set up the transportation tableau for this problem and determine the initial solution
using the minimum cell cost method.
b.Solve using MODI.
c.Are there multiple optimum solutions? Explain. If so, identify them.
d. Formulate this problem as a linear programming model.

5 Steel mills in three cities produce the following amounts of steel:

Location Weekly Demand (tones)
Bethlehem 130
Birmingham 210
Gary 320
620
These mills supply steel to four cities where manufacturing plants have the following
demand.
Location Weekly Demand (tons)
1.Detroit 130
2.St. Louis 70
3.Chicago 180
4.Norfolk 240

115
Operations Research

Shipping costs per ton of steel are as follows

To
From 1 2 3 4
A $14 9 16 18
B 11 8 7 16
C 16 12 10 22

Because of a truckers' strike, shipments are presently prohibited from Birmingham to

Chicago.
a. Set up a transportation tableau for this problem and determine the initial solution.
Identify the method used to find the initial solution.
b. Solve this problem using MODI.
c. Are there multiple optimum solutions? Explain. If so, identify them.
d. Formulate this problem as a general linear programming model.

4. Below is a linear programming problem.

Minimize Z = 17x11 + l0x12 + 15x13 + llX21 + 14x22 + l0x23 + 9x31 + 13x32 + llx33 + 19x41 + 8x42
+ 12x43
Subject to
x11 + x 12 + x 13 = 120
x 21 + x 22 + x 23 = 70
x 31 + x 32 + x 33 = 180
x 41 + x 42 + x 43 = 30
x l1 + x 21 + x 31 + x 41 = 200
x 12 + x 22 + x 32 + x 42 = 120
x 13 + x 23 + x 33 + x 43 = 80
a. Set up the transportation tableau for this problem and determine the initial solution
using VAM.
b.Solve using the stepping-stone method.
c.Calculate the optimal allocation and total cost

116
Operations Research

3.4 ASSIGNMENT PROBLEMS

In the previous section, we discussed about the transportation problem. Now we consider another
type of special linear programming problem, called the assignment problem.

Section objectives:
Up on completion of this section, the learner will be able to:
 Define assignment model
 Formulate LP model for assignment problems
 Identify areas of application
 Find optimal allocation or solution to the model
 Deal with special cases of the model

? Dear learner, can you make any difference between assignment and transportation
problems?
___________________________________________________________________________
___________________________________________________________________________
__________________________________________________________________

There are many situations where the assignment of people or machines and so on, may be called
for. Assignment of workers to machines, clerks to various counters, salesmen to different sales
areas, service crews to different districts, are typical examples of these. The assignment is a
problem because people posses varying abilities for performing different jobs and, therefore, the
costs of performing the jobs by different people are different. Obviously, if all persons could do a
job in the same time or at the same cost then it would not matter who of them is assigned the job.
Thus, in assignment problem, the question is how should the assignment be made in order that
the total cost involved is minimized (or the total value is maximized when pay-offs are given in
terms of, say, profits).
A typical assignment problem follows:
Example
A production supervisor is considering how he should assign the four jobs that are
performed, to four of the workers working under him. He want to assign the jobs to the
workers such that the aggregate time to perform the job in the least. Based on the previous

117
Operations Research

experience, he has the information on the time taken by the four workers in performing
these jobs, as given in below
Job
Worker A B C D
1 45 40 51 67
2 57 42 63 55
3 49 52 48 64
4 41 45 60 55
Example
A computer centre has got four Expert Programmers. Centre needs four application
programmers to be developed. The head of dept., estimate the computer required by the
respective experts to develop the application programs as follows. Make appropriate
selection of experts.

Program A B C D
Expert
1 120 100 80 90
2 80 90 110 70
3 110 140 120 100
4 90 90 80 90

3.4.1 THE ASSIGNMENT PROBLEM (A.P) definition:

1. The Assignment Problem (A.P) is a special case of Transportation Problem (T.P) under

the condition that the number of origins is equal to the number of destinations,
i.e. m=n
Hence assignment is made on the basis of 1:1
Number of jobs equal to number of machines or persons.
Each man or machine is loaded with one and only one job.

118
Operations Research

Each man or machine is independently capable of handling any of the jobs being
presented.
Loading criteria must be clearly specified such as “ minimizing operating time or “
maximizing profit”, or “ minimizing production cost” or minimizing production cycle
time e.t.c
A.P
The Assignment Problem (A.P.) is a special case of Transportation Problem (T.P.) in
which the number of sources and destinations are the same, and the objective is to assign
the given job (task) to most appropriate machine (person) so as to optimize the objective
like minimizing cost.
Cost Vector (Cij)
Cost vector (Cij) is the cost of assigning ith job (Task) to jth machine (person),
Assignment Vector (Xij) is defined as follows
Xij = 0; If ith job (Task) is not assigned by jth machine (person) or
Xij = 1; If ith job (Task) is assigned to jth machine (person).
Cost Matrix (Cij)
Assignment problem is stated in the form of (n*n) matrix. This is called cost
matrix. This is illustrated as given below.
Cij = cost of assigning ith job to jth machine
(Symbols j = Job (Task) M = Machine (person)
General Assignment Problem Table
M 1 2 3 … k … n
J
1 C11 C12 C13 … C1k … C1n
2 C21 C22 C23 … C2k … C2n
3 C31 C32 C33 … C3k … C3k
… … … … … … … …
k Ck1 Ck2 Ck3 … Ckk … Ckn
… … … … … … … …
n Cn1 Cn2 Cn3 … Cnk … Cnn

119
Operations Research

Effective Matrix
A cost Matrix in A.P. is called an “Effectiveness Matrix” when there is at least
one zero in each row and column. Following is an example of Effectiveness
Matrix.
1 2 3 4
1 0 1 3 4
2 5 0 6 0
3 7 8 0 9
4 0 4 3 8
2. Mathematical Modeling of an A.P.
Let there be ‘n’ jobs in a manufacturing plant. Let there be ‘n’ machines to process the
product. A job i (i = 1, 2… n) when processed in a machine j (j = 1, 2… n), it is assumed
to incur a cost of Cij.
The assignment is made in such a way that one job is associated with one machine (see
assumption). Hence we have the following:
n n
 i  1Xij  1 =  j  1Xij  1

Where: - xij is the assignment- vector discussed earlier

Cij is the cost vector associated with the assignment vector x ij.
Total assignment cost (Z) is, therefore, given by the following:

 
n n
Z= i 1 j 1
Cij * xij

Thus the mathematical formulation of assignment Problem is given as follows:

Minimize

 
n n
Z= i 1 j 1
Cij * xij

Subject to
n

x i 1
ij 1 i = 1, 2, 3… n

x
j 1
ij 1 j = 1, 2, 3…, n

120
Operations Research

This special structure of A.P. makes solution much easy compared to the conventional T.P.
Remark:
1. It may be noted that assignment problem is a variation of transportation problem with two
characteristics (i) the cost matrix is square matrix, and (ii) the optimum solution for the problem
would always be such that there would be only one assignment in a given row or column of the
cost matrix.

In assignment problem if a constant is added or subtracted from every element of any row or
column in the given cost matrix then an assignment that minimizes the total cost in one matrix
also minimizes the total cost in the other matrix.

3.5 Solution Method for Assignment

 HUNGARIAN ASSIGNEMNT METHOD (HAM)

A method, designed specially to handle the assignment problems in an efficient way, called the
Hungarian Assignment Method, is available, which is based on the concept of opportunity cost.
It is shown in figure 6.1 and discussed here. For a typical balanced assignment problem
involving a certain number of persons and an equal number of jobs, and with an objective
function of the minimization type, the method is applied as listed in the following steps:

Step 1. Locate the smallest cost element in each row of the cost table. Now subtract this smallest
from each element in that row. As a result, there shall be at least one zero in each row
of this new table, called the Reduced Cost Table (Row Reduction).

Step 2. In the reduced cost table obtained, consider each column and locate the smallest element
in it. Subtract the smallest value from every other entry in the column. As a
consequence of this action, there would be at least one zero in each of the rows and
columns of the second reduced cost table (Column Reduction).

Step 3. Draw the minimum number of horizontal and vertical lines (not diagonal ones) that are
required to cover the entire ‘zero’ elements. If the number of lines drawn is equal to n
(the number of rows/columns) the solution is optimal, and proceeds to step 6. If the
number of lines drawn is smallest than n, go to step 4.

121
Operations Research

Step 4. Select the smallest uncovered (by the lines) cost element. Subtract this element from all
uncovered elements including it self and add this element to each value located at the
intersection of any lines. The cost elements through which only one line passes
remain unaltered.
Step 5. Repeat steps 3 and 4 until an optimal solution is obtained.
Step 6. Given the optimal solution, make the job assignments as indicated by the ‘zero’ elements.
This done as follows:
a) Locate a row which only ‘zero’ element. Assign the job corresponding to this element to
its corresponding person. Cross out the zeros, if any, in the column corresponding to the
element, which is indicative of the fact that the particular job and person are no more
available.
b) Repeat (a) for each of such rows which contain only one zero. Similarly, perform the
same operation in respect of each column containing only one ‘zero’ element, crossing
out the zero(s), if any, in the row in which the element lies.
c) If there is no row or column with only a single ’zero’ element left, then select a
row/column arbitrarily and choose one of the jobs (or persons) and make the assignment.
Now cross the remaining zeros in the column and row in respect of which the assignment
is made.
d) Repeat steps (a) through (c) until all assignments are made.
e) Determine the total cost with reference to the original cost table.
Example
Solve the assignment problem given in Illustrative Example 1 for optimal solution using HAM.
The information is reproduced in the following table
Time Taken (in minutes) by 4 workers
Job
Worker A B C D
1 45 40 51 67
2 57 42 63 55
3 49 52 48 64
4 41 45 60 55

The solution to this problem is given here in a step-wise manner.

122
Operations Research

Step 1: The minimum value of each row is subtracted from all elements in the row. It is shown in
the reduced cost table, also called opportunity cost table, given as follows.
Reduced Cost Table 1
Job
Worker A B C D
1 5 0 11 27
2 15 0 21 13
3 1 4 0 16
4 0 4 19 14

Step 2: For each column of this table, the minimum value is subtracted from all the other values.
Obviously, the columns that contain a zero would remain unaffected by this operation. Here only
the fourth column values would change. The table below shows this.
Reduced Cost Table 2
Job
Worker A B C D
1 5 0 11 14
2 15 0 21 0
3 1 4 0 3
4 0 4 19 1
Step 3: Draw the minimum number of lines covering all zeros. As a general rule, we should first
cover those rows/columns which contain larger number of zeros. The above table is reproduced
in the next table and the lines are drawn.
Reduced Cost Table3
Job
Worker A B C D
1 5 0 11 14
2 15 0 21 0
3 1 4 0 3
4 0 4 19 1

123
Operations Research

Step 4: Since the number of lines drawn is equal to 4(=n), the optimal solution is obtained. The
assignments are made after scanning the rows and columns for unit zeros. Assignments made are
shown with squares, as shown in the following table.
Assignment of Jobs
Job
Worker A B C D
1 5 0 11 14
2 15 0 X 21 0
3 1 4 0 3
4 0 4 19 1

Assignments are made in the following order. Rows 1, 3, and 4 contain only one zero each. So
assign 1-B, 3-C and 4-A. Since worker 1 has been assigned job B, only worker 2 and job D are
left for assignment. The final pattern of assignments is 1-B, 2-D, 3-C, and 4-A, involving a total
time of 40+55+48+41=184 minutes. This is the optimal solution to the problem-the same as
obtained by enumeration and transportation methods.
Example
Using the following cost matrix, determine (a) optimal assignment, and (b) the cost of
assignments.
Reduced Cost Table 1
Job
Machinist 1 2 3 4 5
A 10 3 3 2 8
B 9 7 8 2 7
C 7 5 6 2 4
D 3 5 8 2 4
E 9 10 9 6 10

124
Operations Research

Iteration 1: Obtain row reductions.

Reduced Cost Table 1
Job
Machinist 1 2 3 4 5
A 8 1 1 0 6
B 7 5 6 0 5
C 5 3 4 0 2
D 1 3 6 0 2
E 3 4 3 0 4
Iteration 2: Obtain column reductions and draw the minimum number of lines to cover all
zeros.

Reduced Cost Table2

Job
Machinist 1 2 3 4 5
A 7 0 0 0 4
B 6 4 5 0 3
C 4 2 3 0 0
D 0 2 5 0 0
E 2 3 2 0 2

Since the number of lines covering all zeros is less than the number of columns/rows, we modify
the Table 6.13. The least of the uncovered cell values is 2. This value would be subtracted from
each of the uncovered values and added to each value lying at the intersection of lines
(corresponding to cells A-4, D-4, A-5 and D-5). Accordingly, the new table would appear as
shown as follows.

125
Operations Research

Iteration 3
Job
Machinist 1 2 3 4 5
A 7 0 0 X 2 6
B 4 2 3 0 3
C 2 0 X 1 0 X 0
D 0 2 5 2 2
E 0 X 1 0 0 X 2
The optimal assignments can be made as the least number of lines covering all zeros in Table
6.14 equals 5. Considering rows and columns, the assignments can be made in the following
order:
i. Select the second row. Assign machinist B to job 4. Cross out zeros at cells C-4 and E-4.
ii. Consider row 4, Assign machinist D to job 1. Cancel the zero at cell E-1.
iii. Since there is a single zero in the row, put machinist E to job 3 and cross out the zero at
A-3.
iv. There being only a single left in each of the first and third rows, we assign job 2 to
machinist A and job 5 to C.
The total cost associated with the optimal machinist-job assignment pattern A-2, B-4, C-5, D-1
and E-3 is 3+2+4+3+9 = 21
3.5 Special Case in Assignment Problems
When we solve assignment problems, there are cases which are treated differently from the usual
way.
Unbalanced Assignment Problems
The Hungarian Method of solving an assignment problem requires that the number of
columns should be equal to the number of rows. They are equal, the problem is balanced
problem, and when not, it is called an unbalanced problem. Thus, where there are 5 workers
and 4 machines, or when there are 4 workers and 6 machines, for instance, we have
unbalanced situations in which one-to-one match is not possible. In case the machines are in
excess, the excess machine(s) would remain idle and so is the case when men are in excess-
the number of excess people would get an assignment.

126
Operations Research

In such situations, dummy column(s)/row(s), whichever is smaller in number, are inserted

with zeros as the cost elements. For example, when the given cost matrix is of the dimension
4*5, a dummy row would be included. In each column in respect of this row, a ‘zero’ would
be placed. After this operation of introducing dummy columns/rows, the problem is solved in
the usual manner.

Example:
A company has 4 machines to do 3 jobs. Each job can be assigned to one and only one
machine. The cost of each job on each machine is given below. Determine the job
assignments which will minimize the total cost.
Machine
W X Y Z
A 18 24 28 32
Job
B 8 13 17 18
C 10 15 19 22
Constrained/Prohibited/ Assignment Problems
It happens sometimes that a worker cannot perform a certain job or is not to be assigned a
particular job. To cope with this situation, the cost of performing that job by such person is
taken to be extremely large (which is written as M). Then the solution to the assignment
problem proceeds in the manner discussed earlier. The effect of assigning prohibitive cost to
such person-job combinations is that they do not figure in the final solution.
Example:
You are given the information about the cost of performing different jobs by different
persons. The job-person marking X indicates that the individual involved cannot perform the
particular job. Using this information, state (i) the optimal assignment of jobs, and (ii) the
cost of such assignment,
Job
J1 J2 J3 J4 J5
P1 27 18 X 20 21
31 24 21 12 17
person

P2
P3 20 17 20 X 16
Job

P4 22 28 20 16 27

127
Operations Research

Solution: - Balancing the problem not assigning a high cost to the pairings P1-J3 and P3-J4,
we have the cost given in the table below.
Job
J1 J2 J3 J4 J5
P1 27 18 M 20 21
P2 31 24 21 12 17
P3 20 17 20 M 16
P4 22 28 20 16 27
person

P5 0 0 0 0 0
Job

dummy

Now we can derive the reduced cost table as shown in table shown below. Note that the cells
with prohibited assignments continue to be shown with the cost element M, since M is
defined to be extremely large so that subtraction or addition of value does not practically
affect it. To test optimality, lines are drawn to cover all zeros. Since the number of lines
covering all zeros is less than n, we select the lowest uncovered cell, which equals 4. With
this value, we can obtain the revised reduced cost table, shown in the final table.
Reduced Matrix
Job
J1 J2 J3 J4 J5
P1 9 0 M 2 3
P2 19 12 9 0 5
P3 4 1 44 M 0
P4 6 12 4 0 11
person

P5 0 0 0 0 0
Job

dummy

128
Operations Research

Reduced Matrix
Job
J1 J2 J3 J4 J5
P1 9 0 M 2 3
P2 19 12 9 0 5
P3 4 1 4 M 0
P4 6 12 4 0 11
person

P5 Job 0 0 0 0 0
dummy

Assignment of Job
Job
J1 J2 J3 J4 J5
P1 9 0 M 2 3
P2 15 8 9 0 1
P3 4 1 4 M 0
P4 2 8 0 0 X 7
person

P5 0 0X 0X 0 0 X
Job

dummy
The number of lines covering zeros is equal to 5(=n), hence the optimal assignment can be
made. The assignment is: P1-J2, P2-J4, P3-J5, P4-J3, while job J1 would remain unassigned.
This assignment pattern would cost 18+12+16+20=66 in the aggregate.
Unique Vs Multiple Optimal Solutions
In the process of making assignments, it was stated earlier that we select a row/column with
only a single zero to make an assignment. However, a situation may wherein the various
rows and columns, where assignment are yet to be done, have all multiple zeros. In such
cases, we get multiple optimal solutions to the given problem. In any of the problems
discussed so far, we have not experienced such a situation. Hence, each one of them has had
a unique optimal solution. When a problem has a unique optimal solution, it means that no
other solution to the problem exist which yields the same objective function value (cost, time,

129
Operations Research

profit e. t. c) as the one obtained from the optimal solution derived. On the other hand in a
problem with multiple optimal solutions, there exists more than one solution which all is
optimal and equally attractive. Consider the following example.
Example:
Solve the following assignment problem and obtain the minimum cost at which all the jobs
can be performed.
Job (cost in ’00 Br.)
Worker 1 2 3 4 5
A 25 18 32 20 21
B 34 25 21 12 17
C 20 17 20 32 16
D 20 28 20 16 27
Solution: This problem is unbalanced since number of jobs is 5 while the number of workers
is 4. We first balance it by introducing a dummy worker E, as shown in table below.

Balancing the Assignment Problem

Job (cost in ’00 Br.)
Worker 1 2 3 4 5
A 25 18 32 20 21
B 34 25 21 12 17
C 20 17 20 32 16
D 20 28 20 16 27
E 0 0 0 0 0

Step 1: Obtain reduced cost values by subtracting the minimum value in each row from
every cell in the row. This is given in Table below.

130
Operations Research

Reduced Cost 1
Job (cost in ’00 Br.)
Worker 1 2 3 4 5
A 7 0 14 2 3
B 22 13 9 0 5
C 4 1 4 16 0
D 4 12 4 0 11
E 0 0 0 0 0

Since there is at least one zero in each row and column, we test it for optimality.
Accordingly, lines are drawn. All zeros are covered by 4 lines, which is less than 5 (the order
of the given matrix). Hence, we proceed to improve the solution. The least uncovered value
is 4. Subtracting from every uncovered value and adding it to every value lying at the
intersection of lines, we get the revised values as shown below.

Reduce cost 2 and Assignment

Job (cost in ’00 Br.)
Worker 1 2 3 4 5
A 7 0 14 6 3
B 18 9 5 0 1
C 4 1 4 20 0
D 0 8 0 0 X 7
E 0 X 0 X 0 4 0 X

The solution given in the reduced cost 2 table is optimal since the number of lines covering
zeros matches with the order of the matrix. We can, therefore, proceed to make assignments.
To begin with, since each of the columns has multiple zeros, we cannot start making
assignments considering columns and have, therefore, to look through rows. The first row
has a single zero. Thus, we make assignment A-2 and cross out zero at E-2.

131
Operations Research

Further, the second and the third rows have one zero each. We make assignments B-4 and C-
5, and cross out zeros at D-4 and E-5. Now, both the rows left two zeros each and so have
both the columns. This indicates existence of multiple optimal solutions. To obtain the
solutions, we select zeros arbitrarily and proceed as discussed below.

i. Select the zero at D-1, make assignment and cross out zeros at D-3 and E-1 (as both,
worker D and job 1, are not available any more). Next, assign worker E to job 3,
corresponding to the only zero left. Evidently, selecting the zero at E-3 initially would have
the effect of making same assignments.
ii. Select the zero at D-3, make assignment and cross at D-1 and E-3. Next, make assignment
at the only zero left at E-1. Obviously, selecting the zero at E-1 making assignment in the
first place would lead to the same assignments.

To conclude, the problem has two optimal solutions as given below.

Solution 1 (’00 Br.) Solution 2 (’00 Br.) cost

Worker Job cost Worker Job cost
A 2 18 A 2 18
B 4 12 B 4 12
C 5 16 C 5 16
D 1 20 D 3 20
Job left 3 Job left 1
Total 66 Total 66
Maximization Case in Assignment Problem
In some situations, the assignment problem may call for maximization of profit, revenue,
e.t.c. as the objective. For example, we may be faced with the problem of assignment of
salesmen in different regions in which they can display different qualities in making sales
(reflected in amounts of sales executed by them). Obviously, assignment would be made in
such a way that the total expected revenue is maximized.

132
Operations Research

For dealing with a maximization problem, we first change it into an equivalent minimization
problem. This is achieved by subtraction each of the elements of the given pay-off matrix
from a constant (value) k. Thus, we may simply put a negative sign before each of the pay-
off values (which is equivalent to subtracting each value from zero). Usually, the largest
value of all values in the given matrix is located and then each one of the values is subtract
from it (the largest value is taken so as to avoid the appearance of negative signs). Then the
problem is solved the same way as a minimization problem is.

Example:
A company plans to assign 5 salesmen to 5 districts in which it operates. Estimates of sales
revenue in thousands of birr for each salesman in different districts are given in the following
table. In your opinion, what should be the placement of the salesmen if the objective is to
maximize the expected sales revenue?

District
Salesman D1 D2 D3 D4 D5
S1 40 46 48 36 48
S2 48 32 36 29 44
S3 49 35 41 38 45
S4 30 46 49 44 44
S5 37 41 48 43 47
Solution: Since it is a maximization problem, we would first subtract each of the entries in
the table from the largest one, which equals 49 here. The resultant data are given in Table
below.
Opportunity Loss Matrix
District
Salesman D1 D2 D3 D4 D5
S1 9 3 1 13 1
S2 1 17 13 20 5
S3 0 14 8 11 4
S4 19 3 0 5 5
S5 12 8 1 6 2

133
Operations Research

Now, we will proceed as usual.

Step 1: Subtract value in each row from every value in the row. The resulting values are
given in table below.
Reduced cost Table 1
District
Salesman D1 D2 D3 D4 D5
S1 8 2 0 12 0
S2 0 16 12 19 4
S3 0 14 8 11 4
S4 19 3 0 5 5
S5 11 7 0 5 1

Step 2: Subtract minimum value in each column in reduced cost table 1 from each value in
the column. Test for optimality by drawing lines to cover zeros. These are shown in table
below (in the Reduced cost Table 2)
Reduced cost Table 2
District
Salesman D1 D2 D3 D4 D5
S1 8 0 0 7 0
S2 0 14 12 14 4
S3 0 12 8 6 4
S4 19 1 0 0 5
S5 11 5 0 0 1

Since the number of lines covering all zeros is fewer than n, we select uncovered cell value,
which equals 4. With this, we can modify the table as given in the Reduced Cost Table 3.
Steps 4, 5, 6: Find improved solution. Test for optimality and make assignments.
Reduced cost Table 3
District
Salesman D1 D2 D3 D4 D5
S1 12 0 0 X 7 0 X

S2 0 10 8 10 0 X
S3 0 X 8 4 2 0
S4 23 1 0 0 X 5
S5 15 5 0 X 0 1

134
Operations Research

More than one optimal assignment is possible in this case because of the existence of
multiple zeros in different rows and columns. The possible assignments are:
S1-D2, S2-D5, S3-D1, S4-D3, S5-D4; or
S1-D2, S2-D1, S3-D5, S4-D3, S5-D5; or
S1-D2, S2-D1, S3-D1, S4-D4, S5-D3; or
S1-D2, S2-D1, S3-D5, S4-D4, S5-D
Each of these assignment patterns would lead to expected aggregated sales equal to 231 thousand
birr. START

Write the problem in

tabular form

Is it a Add dummy
No
balanced row(s)/column(s)
problem?

Yes

Convert it into a maximization

problem, either by changeling
Is it a the signs of the elements of the
Yes
maximum signs of the elements of the
problem? table; or by subtracting all the
values from the largest value.
No
Obtain reduced cost tables.
i.Subtract from all entries in each
row the least value in the row.
ii.From this table, subtract from all
entries column the least value in
the column.

Improve the solution. For this:

i. Select the minimum of the uncovered
Can all zeros (by lines) cell values.
be covered Yes ii. Subtract this value from all uncovered
by less than n cell values.
lines? iii. Add this value to the cells lying on the
intersection of any pair of lines.
iv. Leave the cell values covered by only
No one undisturbed.
Make assignments on
one-to-one match basis
considering zeros in
rows/columns

STOP

135
Operations Research

Fig. Schematic Presentation of Hungarian Assignment Method.

Summary
 The purpose of using an LP model for transportation problems is to minimize
transportation costs, taking into account the origin supplies, the destination
demands, and the transportation costs
 A transportation problem typically involves a set of sending locations, which are
referred to as origins, and a set of receiving locations, which are referred to as
destinations.
 For allocation to be feasible, two conditions must be fulfilled:
 A feasible solution is one in which assignments are made in such a way that all
supply and demand requirements are satisfied.
 The number of nonzero (occupied) cells should equal one less than the sum of
the number of rows and the number of columns in a transportation table
 A number of different approaches can be used to find an initial feasible solution.
Which include:
The northwest-corner method.
An intuitive approach/Least cost method
Vogel’s / Penalty Method
 The test for optimality for a feasible solution involves a cost evaluation of empty
cells (i.e., routes to which no units have been allocated) to see if an improved
solution is possible. Two methods for cell evaluation:
The Stepping-stone method
The MODI method
 The Assignment Problem (A.P.) is a special case of Transportation Problem (T.P.)
in which the number of sources and destinations are the same, and the objective is to
assign the given job (task) to most appropriate machine (person) so as to optimize
the objective like minimizing cost.
Review Questions
1. An electronics firm produces electronic components, which it supplies to various
electrical manufacturers. Quality control records indicate that different employees
produce different numbers of defective items. The average number of defects produced
by each employee for each of six components is given in the following table.
Component
Employee A B C D E F
2
1 30 24 16 26 30
2
12
2 22 28 14 30 20
32
2
3 18 16 25 14 12
2
3
4 14 22 18 23 21
0
2
5 25 18 14 16 16
8
6 32 14 10 14 18 2
0
0 136
 Operations Research

Determine the optimal assignment that will minimize the total average
number of defects per month.
2. The Bunker manufacturing firm has five employees and six machines, and wants to
assign the employees to the machines so as to minimize cost. A cost table showing the
cost incurred by each employee on each machine is given below.
Machine
Employee A B C 0 E F
1 $12 7 20 14 8 10
2 10 14 13 20 9 11
3 5 3 6 9 7 10
4 9 11 7 16 9 10

5 10 6 14 8 10 12
Because of union rules regarding departmental transfers, employee 3 cannot be assigned
to machine E and employee 4 cannot be assigned to machine B. Solve this problem,
indicate the optimal assignment, and compute total minimum cost.

3. A university department head has five instructors to be assigned to four different courses.
All of the instructors have taught the courses in the past and have been evaluated by the
students. The rating for each instructor for each course is given in the following table (a
perfect score is 100).
Course
Instructor A B C D
1 80 75 90 85
2 95 90 90 97
3 85 95 88 91
4 93 91 80 84
5 91 92 93 88
The department head wants to know the optimal assignment of instructors to courses that
will maximize the overall average evaluation. The instructor who is not assigned to teach a
course will be assigned to grade exams. Solve this problem using the assignment method.

137
Operations Research

CHAPTER FOUR
DECISION THEORY
Unit objective:
After completing this unit, the learner should be able to:
 Describe the basic characteristics of decision theory problems
 Differentiate between decision analysis under certainty and uncertainty.
 Describe the different approaches (criteria) to decision making under complete
uncertainty.
1. Maximax
2. Maximin
3. Minimax regret
4. Hurwicz
 Use decision tree as decision making tools.

? Dear learner, from your previous courses, can you define decision making? What
steps are involved in it? Can you mention conditions in which decision at different
levels are made?
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________

Dear learner, in the previous units dealing with LP, models were formulated and solved in order
to aid the manager in making decision. The solutions to the models were represented by values
for the decision variables. However, these LP models are formulated under the assumption that
certainty existed. In actual practice, however, many decision making situations occur under
conditions of uncertainty. For example, the demand for a product may be not 100 units next
week, but 50 or 200 units, depending on the market (which is uncertain).

4.1. Characteristics of Decision Theory

Decision theory problems are characterized by the following:

1. List of alternatives: are a set of mutually exclusive and collectively exhaustive decisions
that are available to the decision maker (some times, not always, one of these alternatives
will be to “do nothing”.)

138
Operations Research

2. States of nature: - the set of possible future conditions, or events, beyond the control of
the decision maker, that will be the primary determinants of the eventual consequence of
the decision. The states of nature, like the list of alternatives, must be mutually exclusive
and collectively exhaustive.
3. Payoffs: - the payoffs might be profits, revenues, costs, or other measures of value.
Usually the measures are financial. Usually payoffs are estimated values. The more
accurate these estimates, the more useful they will be for decision making purposes and
the more likely, it is that the decision maker will choose an appropriate alternative. The
number of payoffs depends on the number of alternative/state of nature combination.
4. Degree of certainty: - the approach often used by a decision maker depends on the
degree of certainty that exists. There can be different degrees of certainty. One extreme is
complete certainty and the other is complete uncertainty. The later exists when the
likelihood of the various states of nature are unknown. Between these two extremes is
risk (probabilities are unknown for the states of nature). Knowledge of the likelihood of
each of the states of nature can play an important role in selecting a course of action.
5. Decision criteria: - the decision maker’s attitudes toward the decision as well as the
degree of certainty that surrounds a decision. Example; maximize the expected payoffs.
4.2. THE PAYOFF TABLE
A payoff table is a device a decision maker can use to summarize and organize information
relevant to a particular decision. It includes a list of alternatives, the possible future states of
nature, and the payoffs associated with each of the alternative/state of nature combinations. If
probabilities for the states of nature are available, these can also be listed. The general format
of the table is illustrated below:
States of nature
S1 S2 S3
A1 V11 V12 V13
Alternatives A2 V21 V22 V23
A3 V31 V32 V33
where:
Ai = the ith alternative
Sj = the jth states of nature

139
Operations Research

Vij = the value or payoff that will be realized if alternative i is chosen and event j
occurs.
Decision situations can be categorized in to three classes: Situation of certainty, Situations
where probabilities can not be assigned to future occurrences and Situations where
probabilities can be assigned to future occurrences. In this chapter we will discuss each of
these classes of decision situations separately.
4.3. DECISION MAKING UNDER CERTAINTY
The simplest of all circumstances occurs when decision making takes place in an
environment of complete certainty. When a decision is made under conditions of complete
certainty, the attention of the decision maker is focused on the column in the payoff table that
corresponds to the state of nature that will occur. The decision maker then selects the
alternative that would yield the best payoff, given that state of nature.
EXAMPLE
The following payoff table provides data about profits of the various states of
nature/alternative combination.
S1 S2 S3
A1 4 16 12
A2 5 6 10
A3 -1 4 15
If we know that S2 will occur, the decision maker then can focus on the first raw of the
payoff table. Because alternative A1 has the largest profit (16), it would be selected.
4.4. DECISION MAKING UNDER COMPLETE UNCERTAINTY (With out
probabilities)
Under complete uncertainty, the decision maker either is unable to estimate the probabilities
for the occurrence of the different state of nature, or else he or she lacks confidence in
available estimates of probabilities, and for that reason, probabilities are not included in the
analysis.
A decision making situation includes several components- the decision themselves and the
actual event that may occur future, known as state of nature. At the time the decision is made,
the decision maker is uncertain which state of nature will occur in the future, and has no
control over them.

140
Operations Research

Decisions made under these circumstances are at the opposite end of the spectrum from the
certainty case just mentioned. Once the decision has been organized in to a payoff table,
several criteria are available making the actual decision. There are several approaches
(criteria) to decision making under complete uncertainty. Some of these discussed in this
section include: maximax, maximin,minimax regret, Hurwicz, and equal likelihood.
4.4.1. MAXIMAX
With the maiximax criterion, the decision maker selects the decision that will result in the
maximum of the maximum payoffs. (In fact this is how this criterion derives its name-
maximum of maximum). The maximax is very optimistic. The decision maker assumes that
the most favorable state of nature for each decision alternative will occur. For example, the
investor would optimistically assume that good economic conditions will prevail in the
future. The best payoff for each alternative is identified, and the alternative with the
maximum of these is the designated decision.
For the previous problem:
S1 S2 S3 Row Maximum
A1 4 16 12 16*maximum
A2 5 6 10 10
A3 -1 4 15 15

Decision: A1 will be chosen.

Note: If the pay off table consists of costs instead of profits, the opposite selection would
be indicated: The minimum of minimum costs. For the subsequent decision criteria we
encounter, the same logic in the case of costs can be used.

? Dear learner, how would the decision in the above example change if the
values in the table stand for costs instead of profit?
________________________________________________
________________________________________________
________________________________________________
___

141
Operations Research

4.4.2. Maximin Criteria

This approach is the opposite of the previous one, i.e. it is pessimistic. This strategy is a
conservative one; it consists of identifying the worst (minimum) payoff for each alternative,
and, then, selecting the alternative that has the best (maximum) of the worst payoffs. In
effect, the decision maker is setting a floor on the potential payoff by selecting maximum of
the minimum; the actual payoff can not be less than this amount. It involves selecting best of
the worst.
For the previous problem:
S1 S2 S3 Row minimum
A1 4 16 12 4
A2 5 6 10 5*maximum
A3 -1 4 15 -1
Decision: A2 will be chosen.
Note: If it were cost, the conservative approach would be to select the maximum cost for
each decision and select the minimum of these costs.

? Dear learner, how would the decision in the above example change if the
values in the table stand for costs instead of profit?
________________________________________________
________________________________________________
______________________________________________

4.4.3. MINIMAX REGRET

Both the maximax and maximin strategies can be criticized because they focus only on a
single, extreme payoff and exclude the other payoffs. Thus, the maximax strategy ignores the
possibility that an alternative with a slightly smaller payoff might offer a better overall
choice. For example, consider this payoff table:
S1 S2 S3 Row Max.
A1 -5 16 -10 16*max
A2 15 15 15 15
A3 15 15 15 15

142
Operations Research

A similar example could be constructed to demonstrate comparable weaknesses of the

maximin criterion, which is also due to the failure to consider all payoffs.
An approach that does take all payoffs in to consideration is Minimax regret. In order to use
this approach, it is necessary to develop an opportunity loss table. The opportunity loss
reflects the difference between each payoff and the best possible payoff in a column (i.e.,
given a state of nature). Hence, opportunity loss amounts are found by identifying the best
payoff in a column and, then, subtracting each of the other values in the column from that
payoff. Therefore, this decision avoids the greatest regret by selecting the decision alternative
that minimizes the maximum regret.
EXAMPLE:
S1 S2 S3
A1 4 16 12
A2 5 6 10
A3 -1 4 15
opportunity loss table:
S1 S2 S3
A1 5-4=1 16-16=0 15-12=3
A2 5-5=0 16-6=10 15-10=5
A3 5-(-1)=6 16-4=12 15-15=0
The values in an opportunity loss table can be viewed as potential “regrets” that might be
suffered as the result of choosing various alternatives. A decision maker could select an
alternative in such a way as to minimize the maximum possible regret. This requires
identifying the maximum opportunity loss in each row and, then, choosing the alternative
that would yield the best (minimum) of those regrets.
S1 S2 S3 Max. Loss
A1 5-4=1 16-16=0 15-12=3 3*minimum
A2 5-5=0 16-6=10 15-10=5 10
A3 5-(-1)=6 16-4=12 15-15=0 12

Decision: A1 will be chosen.

143
Operations Research

? Dear learner, how would the decision in the above example change if the
values in the table stand for costs instead of profit?
_______________________________________________
_______________________________________________
_______________________________________________
_____
Although this approach makes use of more information than either Maximin or Maximax, it
still ignores some information, and, therefore, can lead to a poor decision.
EXAMPLE:
Opportunity loss table
S1 S2 S3 S4 Max. Loss
A1 0 0 0 24 24
A2 15 15 15 0 15*minimum
A3 15 15 15 0 15*minimum

4.4.4. PRINCIPLE OF INSUFFICIENT REASON/ Equal likelihood/ Laplace

The Minimax regret criterion’s weakness is the inability to factor row differences. Hence,
sometimes the minimax regret strategy will lead to a poor decision because it ignores certain
information.
The principle of insufficient reason offers a method that incorporates more of the information. It
treats the states of nature as if each were equally likely, and it focuses on the average payoff for
each row, selecting the alternative that has the highest row average.

EXAMPLE
S1 S2 S3 S4 S5 Row Average
A1 28 28 28 28 4 23.2*maximum
A2 5 5 5 5 28 9.6
A3 5 5 5 5 28 9.6
Decision: A1 is selected
The basis for the criterion of insufficient reason is that under complete uncertainty, the decision
maker should not focus on either high or low payoffs, but should treat all payoffs (actually, all
states of nature), as if they were equally likely. Averaging row payoffs accomplishes this.

144
Operations Research

? Dear learner, how would the decision in the above example change if the values in the
table stand for costs instead of profit?
__________________________________________________________________________
__________________________________________________________________________
__________________________________________________________________________
_
4.4.5. The Hurwitz Criterion
The Hurwitz criterion strikes a compromise between the maximax and maximin criterion. The
principle underlying this decision criterion is that the decision maker is neither totally optimistic,
nor totally pessimistic. With Hurwitz criterion, the decision payoffs are weighted by a coefficient
of optimism, a measure of a decision maker’s optimism. The coefficient of optimism, which is
defined as, is between zero and one (0< <1). If  = 1, then the decision maker is said to be
completely optimistic, if = 0, then the decision maker is completely pessimistic. Given this
definition, if  is coefficient of optimism, 1- is coefficient of pessimism.
The Hurwitz criterion requires that for each alternative, the maximum payoff is multiplied by 
and the minimum payoff be multiplied by 1-.
Example: If  = 0.4 for the above example,
A1 = (0.4x16) + (0.6x4)
= 8.8
A2 = (0.4x10) + (0.6x5)
=7
A3 = (0.4x15) – (0.6x1)
= 5.4
Decision: A1 is selected
A limitation of Hurwicz criterion is the fact that  must be determined by the decision maker.
Regardless of how the decision maker determines , it is still a completely a subjective measure
of the decision maker’s degree of optimism. Therefore, Hurwicz criterion is a completely
subjective decision making criterion.

? Dear learner, can you mention conditions under which Hurwicz criterion criteria can be
considered as maximin or minimax criterion
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
__ 145
Operations Research

4.5. DECISION MAKING UNDER RISK (WITH PROBABILITIES)

Dear learner, the decision making criteria’s just presented were based on the assumption that no
information regarding the likelihood of the states of the nature was available. Thus, no
probabilities of occurrence were assigned to the states of nature, except in the case of the equal
likely hood criterion.
It is often possible for the decision maker to know enough about the future state of nature to
assign probabilities to their occurrences. The term risk is often used in conjunction with partial
uncertainty, presence of probabilities for the occurrence of various states of nature. The
probabilities may be subjective estimates from managers or from experts in a particular field, or
they may reflect historical frequencies. If they are reasonably correct, they provide the decision
maker with additional information that can dramatically improve the decision making process.
Given that probabilities can be assigned, several decision criteria are available to aid the decision
maker. Some of these are discussed below.
4.5.1. EXPECTED MONETARY VALUE (EMV)
The EMV approach provides the decision maker with a value which represents an average payoff
for each alternative. The best alternative is, then, the one that has the highest EMV. The average
or expected payoff of each alternative is a weighted average:
k
EMVi = Σ Pj.Vij
i=1
Where:
EMVi = the EMV for the ith alternative
Pi = the probability of the ith state of nature
Vij = the estimated payoff for alternative i under state of nature j.
Note: the sum of the probabilities for all states of nature must be 1.

EXAMPLE:

Probability 0.20 0.50 0.30

S1 S2 S3 Expected payoff

146
Operations Research

A1 4 16 12 12.40*maximum
A2 5 6 10 7
A3 -1 4 15 6.30
Decision: A1 will be chosen.

Note that it does not necessarily follow that the decision maker will receive a payoff equal
to the expected monetary value of a chosen alternative. Similarly, the expected payoffs for
either of the other alternatives do not equal any payoffs in those rows. What, then, is the
interpretation of the expected payoff? Simply a long-run average amount; the approximate
average amount one could reasonably anticipate for a large number of identical situations.

4.5.2. Expected Opportunity Loss (EOL)

The table of opportunity loss is used rather than a table of payoffs. Hence, the opportunity
losses for each alternative are weighted by the probabilities of their respective state of nature
to compute a long run average opportunity loss, and the alternative with the smallest
expected loss is selected as the best choice.
EOL (A1) = 0.20(1) + 0.50(0) + 0.30(3) = 1.10 *minimum
EOL (A2) = 0.20(0) + 0.50(10) + 0.30(5) = 6.50
EOL (A3) = 0.20(6) + 0.50(12) + 0.30(0) = 7.20

Note: The EOL approach resulted in the same alternative as the EMV approach
(Maximizing the payoffs is equivalent to minimizing the opportunity losses).

4.5.3. Expected Value of Perfect Information (EVPI)

It can some times be useful for a decision maker to determine the potential benefit of knowing
for certain which state of nature is going to prevail. The EVPI is the measure of the difference
between the certain payoffs that could be realized under a condition involving risk.
If the decision maker knows that S1 will occur, A2 would be chosen with a payoff of $5.
Similarly for S2 $16 (for A1) and for S3, $15 (with A3) would be chosen.
Hence, the expected payoff under certainty (EPC) would be:
EPC = 0.20(5) + 0.50(16) + 0.30(15) = 13.50

147
Operations Research

The difference between this figure and the expected payoff under risk (i.e., the EMV) is the
expected value of perfect information. Thus:
EVPI = EPC – EMV
= 13.50 – 12.40 = 1.10
Note: The EVPI is exactly equal to the EOL. The EOL indicates the expected opportunity
loss due to imperfect information, which is another way of saying the expected payoff that could
be achieved by having perfect information.

Note: The expected value approach is particularly useful for decision making when a number
of similar decisions must be made; it is a long-run approach. For one-shot decisions, especially
major ones, other methods (perhaps, maximax or maximin) may be preferable. In addition, non
monetary factors, although not included in a payoff table, may be of considerable importance.
Unfortunately, there is no convenient way to include them in an expected value analysis.

? Dear learner, can you make differences between decision making situations under
??uncertainty and risk?
___________________________________________________________________________
___________________________________________________________________________
___________________________________________________________________________
___
4.6. DECISION TREES
Decision trees some times are used by decision makers to obtain a visual portrayal of decision
alternatives and their possible consequences. The term gets its name from the tree-like
appearance of the diagram.
Decision tree format:

Decision tree, like probably tree is composed of squares, circles, and lines:
 The squares indicate decision points
 Circles represent chance events( circles and squares are called nodes)
o The lines (branches) emanating from squares represent alternatives.

148
Operations Research

o The lines from circles represent states of nature

 The tree is read from right to left.
It should be noted that although decision trees represent an alternative approach to payoff tables,
they are not commonly used for problems that involve a single decision. Rather, their greatest
benefit lies in portraying sequential decisions (i.e., a series of chronological decisions). In the
case of a single decision, constructing a decision tree can be cumbersome and time consuming.
Example
Pay off table for Real Estate investment
State of Nature
Good economic Poor economic
Decision conditions conditions
(Purchase) 0.6 0.4

Apartment building 50,000 30,000

Office building 100,000 -40,000
Warehouse 30,000 10,000

The decision tree for the above example will be:

Good economic $50,000

conditions (0.6

Apartment 2 $30,000
building Poor economic
conditions (0.4)
Purchase
Good economic $100,000
conditions (0.60
Office building
1 3
Poor economic -$40,000
conditions (0.4)
Warehouse Good economic
conditions (0.60 $30,000
4

Poor economic $10,000

conditions
(0.0.4)
The circles ( ) and squares (  ) in the above figure are referred to as nodes. The
squares are decision nodes, and the branches emanating from a decision node reflect the
alternative decisions possible at that point. For example, in the above figure, node 1

149
Operations Research

signifies a decision to purchase an apartment building. Office building, or ware house. The
circles are probability nodes, and the branches emanating from them indicate the state of
nature that can occur: good economic conditions or poor economic conditions. The
decision tree represents the sequence of events in a decision situation. First, one of the
three decision choices is selected at node 1. Depending on the branch selected, the decision
maker arrives at probability node 2, 3, or 4, where one of the states of nature will prevail,
resulting in one of six possible payoffs.

Determining the best decision using a decision tree involves computing the expected value
at each probability node. This is accomplished by starting with the final outcomes
(payoffs) and working backward through the decision tree toward node 1. First, the
expected value of the payoffs is computed at each probability node.
EV(node 2) = .60($ 50,000) + .40($ 30,000) = $42,000
EV(node 3) = .60($100,000) + .40($-40,000) = $44,000
EV(node 4) = .60($ 30,000) + .40($ 10,000) = $22,000
These values are now shown as the expected payoffs from each of the three branches
emanating from node 1 in figure below. Each of these three expected values at nodes 2, 3,
and 4 is the outcome of a possible decision that can occur at node 1. Moving toward node
1, we select the branch that comes from the probability node with the highest expected
payoff. In figure below, the branch corresponding to the highest payoff, $44,000 is from
node 1 to node 3. This branch represents the decision to purchase the office building. The
decision to purchase the office building, with an expected payoff of $44,000, is the same
result we achieved earlier using the expected value criterion. In fact, when only one
decision is to be made (i.e., there is not a series of decisions), the decision tree will always
yield the same decision and expected payoff as the expected value criterion. As a result, in
these decision situations the decision tree is not very useful. However, when a sequence or
series of decisions is required, the decision tree can be very useful.

150
Operations Research

$42,000 Good economic

Conditions (0.6)

2
$50,000

Apartment $44,000

building
3
1 Poor economic
Conditions(0.4) $30,000

$22,000 Good economic

Office building Conditions (0.6)
4
$100,000
Purchase
Poor economic
Conditions -$40,000

Good economic
Warehouse Conditions (0.6)
$30,000

Poor economic
Conditions(0.4) $10,000
4.6.1. Sequential Decision Trees
As noted above, when a decision situation requires only a single decision, an expected value
payoff table will yield the same result as a decision tree. However, a payoff table is usually
limited to a single decision situation like our real estate investment example. If a decision
situation requires a series of decisions, then a payoff table cannot be created and a decision
tree becomes the best method for decision analysis.

151
Operations Research

In order to demonstrate the use of a decision tree for a sequence of decisions, we will alter
our real estate investment example to encompass a ten-year period during which several
decisions must be made. In this new example, the first decision facing the investor is whether
to purchase an apartment building or land. If the investor purchases the apartment building,
two states of nature are possible. Either the population of the town will grow (with a
probability of 0.60), or the population will not grow (with a probability of 0.40). Either state
of nature will result in a payoff. On the other hand, if the investor chooses to purchase land,
three years in the future another decision will have to be made regarding the development of
the land. The decision tree for this example, shown in figure below, contains all the
pertinent data, including decisions, states of nature, probabilities, and payoffs.

At decision node 1 in figure below, the decision choices are to purchase an apartment
building and to purchase land. Notice that the cost of each venture ($800,000 and $200,000,
respectively) is shown in parentheses. If the apartment building is purchased,
two states of nature are possible at probability node 2. The town may exhibit population
growth, with a probability of .60, or there may be no population growth or a decline, with a
probability of .40. If the population grows, the investor will achieve a payoff of $2,000,000
over a ten-year period. (Note that this whole decision situation encompasses a ten-year time
span.) However, if no population growth occurs, a payoff of only $225,000 will result.

If the decision is to purchase land, two states of nature are possible at probability node 3.
These two states of nature and their probabilities are identical to those at node 2; however,
the payoffs are different. If population growth occurs for a three-year period, no payoff will occur,
but the investor will make another decision at node 4 regarding development of the land. At that
point either apartment will be built at a cost of $800,000 or the land will be sold with a payoff of
$450,000. Notice that the decision situation at node 4 can occur only if population growth occurs
first. If no population growth occurs at node 3, there is no payoff and another decision situation
becomes necessary at node 5: the land can be developed commercially at a cost of $600,000 or
the land can be sold for $210,000. (Notice that the sale of the land results in less profit if there is
no population growth than if there is population growth.)
If the decision at decision node 4 is to build apartments, two states of nature are possible. The

152
 Operations Research

population may grow, with a conditional probability of .80, or there may be no population
growth, with a conditional probability of .20. The probability of population growth is higher (and
the probability of no growth is lower) than before because there has already been population
growth for the first three years, as shown by the branch from node 3 to node 4. The payoffs for
these two states of nature at the end of the ten-year period are $3,000,000 and $700,000,
respectively, as shown in figure below.
If the investor decides to develop the land commercially at node 0. 5, then two states of nature
can occur. Population growth can occur, with a probability of .30 and an eventual payoff of
$2,300,000, or no population growth can occur, with a probability of .70 and a payoff of
$1,000,000. The probability of population growth is low (i.e., .30) because there has already
been no population growth, as shown by the branch from node 3 to node 5
Sequential Decision Tree
$ 2,000,000

2 $ 225,000
$ 3,000,000
.60
6
6
.20
.40 ,00
1 0,0
.80 $ 700,000
4 00
$ 450,000
Sell land

.30
3 ,00
7 0,0 $ 2,000,000
00
.70
,00
0,0
5 00 $ 1,000,000
Sell land
$ 210,000

.60

153
Operations Research

.40
This decision situation encompasses several sequential decisions that can be analyzed using the
decision tree approach outlined in our earlier (simpler) example. As before, we start at the end of
the decision tree and work backward toward a decision at node 1.
First we must compute the expected values at nodes 6 and 7.
EV (node 6) = .80($3,000,000) + .20($ 700,000) = $2,540,000
EV (node 7) = .30($2,300,000) + .70($1,000,000) = $1,390,000
Both of these expected values (as well as all other nodal values) are shown in boxes in Figure 4.4.
At decision nodes 4 and 5, we must make a decision. As with a normal payoff table, we make
the decision that results in the greatest expected value. At node 4 we have a choice between two
values: $1,740,000, the value derived by subtracting the cost of building an apartment building
($800,000) from the expected payoff of $2,540,000, or $450,000, the expected value of selling
the land computed with a probability of 1.0. The decision is to build the apartment building, and
the value at node 4 is $1,740,000.
This same process is repeated at node 5. The decisions at node 5 result in payoffs of $790,000
(i.e., $1,390,000 - 600,000 = $790,000) and $210,000. Since the value $790,000 is higher, the

decision is to develop the land commercially.

Sequential Decision Tree with Nodal Expected Values
$ 2,000,000

$1,290,000

2 $ 225,000 $ 7540,000
.60
6
Purchase
6 $ 3,000,000
.20
Apartment .40 .80
$1,740,000 ,00
1Building 0,0
$1,160,000 00
4
$ 450,000
Sell land

$1,390,000 .20
.60 .30
3 .40 ,00
7 0,0 $ 700,000
00
$1,360,000 $790,000 .70
,00
0,0
5 00 $ 2,300,000
$ 1,000,000
Sell land
$ 210,000

154
 Operations Research

Next we must compute the expected values at nodes 2 and 3.

EV (node 2) = .60($2,000,000) + .40($225,000) = $1,290,000

EV (node 3) = .60($1,740,000) + .40($790,000) = $1,360,000

(Note that the expected value for node 3 is computed from the decision values previously
determined at nodes 4 and 5.)
Now we must make the final decision for node 1. As before, we select the decision with
the greatest expected value after the cost of each decision is subtracted.

Apartment building: $1,290,000 - 800,000 = $ 490,000

Land: $1,360,000 - 200,000 = $1,160,000

Since the highest net expected value is $1,160,000, the decision is to purchase land and the
payoff of the decision is $1,160,000.
This example demonstrates the usefulness of decision trees for decision analysis. The
decision tree allows the decision maker to see the logic of decision making, because it
provides a picture of the decision process. Decision trees can be used for decision problems
more complex than the example above without too much difficulty.

4.6.2. DECISION ANALYSIS WITH ADDITIONAL INFORMATION

Earlier in this unit we discussed the concept of the expected value of perfect information. We
noted that if perfect information could be obtained regarding which states of nature would
occur in the future the decision maker could obviously make better decisions. Although
perfect information about the future is rare, it is often possible to gain some amount of
additional (imperfect) information that will improve decisions.
In this section we will present a process for using additional information m the decision-
making process by applying Bayesian analysis, a probabilistic technique. This process will
be demonstrated using the real estate investment example employed throughout this
chapter. To briefly review this example, a real estate investor is considering three
alternative investments, which will occur under one of two possible economic conditions

155
Operations Research

Payoff Table for the Real Estate Investment Example

States of Nature
Decision Good Economic Conditions Poor Economic Conditions
(purchase) .60 .40

Apartment building $ 50,000 $ 30,000

Office building 100,000 - 40,000
Warehouse 30,000 10,000
(states of nature) shown in table above.

Recall that using the expected value criterion, we found the best decision to be the purchase of
the office building, with an expected value of $44,000. We also computed the expected value of
perfect information to be $28,000. Therefore, the investor would be willing to pay up to $28,000
for information about the states of nature, depending on how close to perfect the information
was.
Now suppose that the investor has decided to hire a professional economic analyst who will
provide additional information about future economic conditions. The analyst is constantly
researching the economy, and the results of this research are what the investor will be
purchasing.
The economic analyst will provide the investor with a report predicting one of two outcomes.
The report will be either positive, indicating that good economic condition are most likely to
prevail in the future, or negative, indicating that poor economic conditions will probably occur.
Based on the analyst's past record in forecasting future economic conditions, the investor has
determined conditional probabilities of the different report outcomes given the occurrence of each
state of nature in the future. We will use the following notations to express these conditional
probabilities:
g = good economic conditions
p = poor economic conditions
P = positive economic report

156
Operations Research

N = negative economic report

The conditional probability of each report outcome given the occurrence of each state of nature
is shown below.
P(P/g) = .80
P(N/g) = .20
P(P/p) = .10
P(N/p) = .90
For example, if future economic conditions are in fact good (g), the probability that a positive
report (P) will have been given by the analyst, P(P/g), is .80. The other three conditional
probabilities can be interpreted similarly. Notice that these probabilities indicate that the analyst
is a relatively accurate forecaster of future economic conditions.
The investor now has quite a bit of probabilistic information available - not only the conditional
probabilities of the report, but also the prior probabilities that each state of nature will occur.
These prior probabilities that good or poor economic conditions will occur in the future are
P(g) = .60
P(p) = .40
Given the conditional probabilities, the prior probabilities can be revised to form posterior
probabilities by means of Bayes's rule. If we know the conditional probability that a positive
report was presented given that good economic conditions prevail, P(P / g), the posterior
probability of good economic conditions given a positive report, P(g/P), can be determined using
Bayes's rule, as follows.
P(g /P) = P(P / g)P(g)
P(P/g)P(g) + P(P/p)P(p)
(.80) (.60)
(.80) (.60) + (.10) (.40)
= .923
The prior probability that good economic conditions will occur in the future is .60. However,
by obtaining the additional information of a positive report from the analyst, the investor can
revise the prior probability of good conditions to a .923 probability that good economic
conditions will occur. The remaining posterior (revised) probabilities are

157
Operations Research

P(g/N) = .250
P(p /P) = .077
P(p/N) = .750
4.6.3. Decision Trees with Posterior Probabilities
The original decision tree analysis of the real estate investment example is shown in Figures 4.1
and 4.2. Using these decision trees, we determined that the appropriate decision was the purchase
of an office building, with an expected value of $44OOO. However, if the investor hires an
economic analyst, the decision regarding which piece of real estate to invest in will not be made
until after the analyst presents the report. This creates an additional stage in the decision-making
process, which is shown in the decision tree in Figure 4.5.
The decision tree shown in Figure 4.5 differs in two respects from the decision trees in Figures
4.1 and 4.2. The first difference is that there are two new branches at the beginning of the
decision tree that represent the two report outcomes. Notice, however, that given either report
outcome, the decision alternatives, the possible states of nature, and the payoffs are the same as
those in the first two figures.

The second difference is that the probabilities of each state of nature are no longer the prior
probabilities given in Figure 4.1; instead they are the revised posterior probabilities computed in
the previous section using Bayes's rule. If the economic analyst issues a positive report, then the
upper branch in figure below (from node 1 to node 2) will be taken.
If an apartment building is purchased (the branch from node 2 to node 4), the probability of
good economic conditions is .923, whereas the probability of poor conditions is .077. These are
the revised posterior probabilities of the economic conditions given a positive report. However,
before we can perform an expected value analysis using this decision tree, one more piece of
probabilistic information must be determined-the initial branch probabilities of a positive and a
negative economic report.

The probability of a positive report, P(P), and of a negative report, P(N), can be determined
according to the following logic. Recall from Chapter 10 that the probability that two dependent
events, A and B, will both occur is
P(AB) = P(A/B) P(B)

158
Operations Research

Decision Tree with Posterior Probabilities

$50,000
Pig p1=.923
4
Apartment
building
2
5 P/P/p) = 077 $30,000
Office building PigP)= 923
$100,00
Positive 6

report P(P/P)=077 -$ 40,000

1
Warehouse P(g/p)= 923
$ 30,000
7

8 P(p/p)= 077 $ 10,00

3
P(g/N)= 250
9 $ 50,00
Apartment
Building
Negative $30,000
report Office building P(p/N) = 750
P(g/N) +250 $100,000
-$40,000
P(g/N)= . 750
Warehouse P(g/N)= .250 $ 30,000

$ 10,000
P(p/N)= 750

159
Operations Research

If event A is a positive report and event B is good economic conditions, then according to the
above formula,
P(Pg) = p(p / g)P(g)
We can also determine the probability of a positive report and poor economic conditions the
same way.
P(Pp) = P(P/p)P(p)
Next consider the two probabilities P(Pg) and P(Pp). These are, respectively, the probability of a
positive report and good economic conditions and the probability of a positive report and poor
economic conditions. These two sets of occurrences are mutually exclusive, since both good and
poor economic conditions cannot occur simultaneously in the immediate future. Conditions will
be either good or poor, but not both. To determine the probability of a positive report, we add the
mutually exclusive probabilities of a positive report with good economic conditions and a
positive report with poor economic conditions, as follows.
P(P) = P(Pg) + P(Pp)
Now, if we substitute into this formula the relationships for P(Pg) and P(Pp) determined
earlier, we have
P(P) = P(P/g)P(g) + P(P/p)P(p)
You might notice that the right-hand side of this equation is the denominator of the Bayesian
formula we used to compute P(g/P) in the previous section. Using the conditional and prior
probabilities that have already been established, we can determine that the probability of a
positive report is
P(P) = P(P/g)P(g) + P(P/p)P(p) = (.80) (.60) + (.10) (.40)
= .52
Similarly, the probability of a negative report is
P(N) = P(N/g)P(g) + P(N/p)P(p)
= (.20) (.60) + (.90) (.40)
= .48
Now we have all the information needed to perform a decision tree analysis. The decision tree
analysis for our example is shown in Figure 11.6. To see how the decision tree analysis is
conducted, consider the result at node 4 first. The value $48,460 is the expected value of the
purchase of an apartment building given both states of nature. This expected value is computed
as follows.

160
Operations Research

EV (Apartment building) = $50,000(.923) + 30,000(.077) = $48,460

The expected values at nodes 5, 6, 7, 8, and 9 are computed similarly.
The investor will actually make the decision as to which investment to make at nodes 2 and 3. It
is assumed that the investor will make the best decision in each case. Thus, the decision at node
2 will be to purchase an office building, with an expected value of $89,212; the decision at node
3 will be to purchase an apartment building, with an expected value of $35,000. These two
results at nodes 2 and 3 are referred to as decision strategies. They represent a plan of decisions to
be made given either a positive or a negative report from the economic analyst.
The final step in the decision tree analysis is to compute the expected value of the decision
strategy given that an economic analysis is performed. This expected value, shown as $63,190 at
node 1 in Figure 11.6, is computed as follows.
EV (strategy) = $89,212(.52) + 35,000(.48) = $63,190
Decision Tree Analysi
$50,000

$30,000

$100,000

-$40,000

$30,000

$10,000

$50,000

$30,000

$100,000

-$40,000

$30,000

$10,000
$15,000

P (p/N) =.750

161
Operations Research

This amount, $63,190, is the expected value of the investor's decision strategy given that a report
forecasting future economic condition is generated by the economic analyst.
5.6.4. Computing Posterior Probabilities with Tables
One of the difficulties that can occur with this type of decision analysis with additional
information is that as the size of the problem increases (i.e., as we add more decision
alternatives and states of nature) the application of Bayes's rule to compute the posterior
probabilities becomes more complex. In such cases, the posterior probabilities can be computed
using tables. This tabular approach will be demonstrated with our real estate investment
example. The table for computing posterior probabilities for a positive report and P(P) is initially
set up as shown in table below.
The posterior probabilities for either state of nature (good or poor economic conditions) given a
negative report are computed similarly.
(1) (2) (3) (4) (4)
States Prior Conditional Prior Posterior
Of Nature Probabilities Probabilities Probability x Probabilities:
Conditional (4)   (4)
Probability:
(2) x (3)
Good P(g) = .6 P(P/g) = 0.80 P(Pg) = .48 .48
P(g/P) =  .923
Conditions .52
Poor P(P) = .4 P(p/P) = 0.10 P ( Pp )  .04 .04
P(p/P)= .  .077
Conditions   P ( P )  .52 52

No matter how large the decision analysis, the steps of this tabular approach can be
followed the same way as in this relatively small problem. This approach is more
systematic than the direct application of Bayes's rule, making it easier to compute the
posterior probabilities for larger problems.

4.6.5. The Expected Value of Sample Information

Recall that we computed the expected value of our real estate investment example to be
$44,000 when we did not have any additional information. After obtaining the additional
information provided by the economic analyst, we computed an expected value of

162
Operations Research

$63,190 using the decision tree in Figure 4.6. The difference between these two expected
values is called the expected value of sample information (EVSI), and it is computed as follows.
EVSI = EV with information - EV with out information
For our example, the expected value of sample information is
EVSI = $63,190 - 44,000
= $19,190
This means that the real estate investor would be willing to pay the economic analyst up
to $19,190 for an economic report that forecasted future economic conditions.
After we computed the expected value of the investment without additional information,
we computed the expected value of perfect information, which equaled $28,000. However, the
expected value of the sample information was only $19,190. This is a logical result, since
it is rare than absolutely perfect information can be determined. Since the additional
information that is obtained is less than perfect, it will be worth less to the decision maker.
We can determine how close to perfect our sample information is by computing the
efficiency of sample information as follows.
EVSI
Efficiency = EVPI

= $ 19,190
28,000
=.68
Thus, the analyst's economic report is viewed by the investor to be 68% as efficient as
perfect information. In general, a high efficiency rating indicates that the information is
very good, or close to being perfect information, and a low rating indicates that the
additional information is not very good. For our example, the efficiency of .68 is relatively
high; thus it is doubtful that the investor would seek additional information from an
alternative source. (However, this is usually dependent on how much money the decision
maker has available to purchase information.) If the efficiency had been lower, however,
the investor might seek additional information elsewhere.

163
Operations Research

Summary

 Decision theory problems are characterized by, list of alternatives, states of nature,
payoffs, degree of certainty, decision criteria.

 Decision situations can be categorized in to three classes: Situation of certainty,

Situations where probabilities can not be assigned to future occurrences and Situations
where probabilities can be assigned to future occurrences.

 There are several approaches (criteria) to decision making under complete uncertainty.
Some of these discussed in this section include: maximax, maximin,minimax regret,
Hurwitz, and equal likelihood.

 Decision making under risk (with probabilities) involves several decision criteria
including Expected Monetary Value (EMV), Expected Opportunity Loss (EOL),
Expected Value of Perfect Information (EVPI).

 Decision trees represent an alternative approach to payoff tables; which are used for
problems that involve a series of chronological decisions by portraying sequential
decisions graphically.

 The circles ( ) and squares (  ) in the above figure are referred to as nodes. The
squares are decision nodes, and the branches emanating from a decision node reflect
the alternative decisions possible at that point.

164
Operations Research

CHAPTER FIVE
NETWORK MODELS

? Dear learner, could you discuss about importance of network diagrams with your
colleagues?
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
_______________
Drawing a logic diagram is a skill requiring practice and ingenuity, and for major projects may
require two or three attempts before a satisfactory network or diagram is completed correctly.
Computer packages are often used to carry out this process, cutting drastically the time taken.
5.2.1. Basic Components of Network Diagram
i) Network
It is the graphic representation of logically and sequentially connected arrows and nodes
representing activities and events of a project. Networks are also called arrow diagram.
ii) Activity
An activity represents some action and is a time consuming effort necessary to complete a
particular part of the overall project. Thus each and every activity has a point where it ends.
It is represented in the network by an arrow.
A

Here A is called an activity.

The straight lines connecting the circles represent a task that takes time or resources; these lines
are called activities. The arrow heads show the direction of the activity.
The beginning and end points of an activity are called events or nodes. Event is a point in the line
and does not consume any resource. A numbered circle represents it. The head event has always
a number higher than the tail event.

Activity
Tail Head

165
Operations Research

Merge and burst events

It is not necessary an event to be the ending event of the only one activity but can be the ending
event of two or more activities. Such event is defined as a Merge event.

Merge event

If the event happened to be the beginning event of two or more activities, it is defined as a Burst
event.

Burst event

iii) Preceding, succeeding and concurrent activities

Activities, which must be accomplished before a given event can occur, are termed as preceding
activities. Activities, which cannot be accomplished until any event has occurred, are termed as
succeeding activities. Activities that can be accomplished concurrently are known as concurrent
activities.
This classification is relative, which means that one activity can be proceeding to a certain event,
and the same activity can be succeeding to some other event or it may be a concurrent activity
with one or more activities.

? Dear learner, can you define the following terms?

a) Preceding activity____________________________________________________
b) Succeeding activities_________________________________________________
c) Concurrent activities__________________________________________________

iv) Dummy Activity

Certain activities, which neither consume time nor resources but are used simply to represent a
connection or a link between the events, are known as dummies. It is shown in the network by a
dotted (broken) line.
The purpose of introducing dummy activity is:
 To maintain uniqueness in the numbering system as every activity may have
distinct set of events by which the activity can be identified.

166
Operations Research

 To maintain a proper logic in the network.

D
A
B

Dummy

C
v) Numbering the Events
After the network is drawn in a logical sequence, every event is assigned a number. The number
sequence must be in such a way that it should reflect the flow of the network. In event
numbering, the following rules should be observed:
i. Event numbers should be unique.
ii) Event numbering should be carried out a sequential basis from
left to right.
iii) The initial event, which has all outgoing arrows with no incoming arrow, is
numbered 0 or 1.
iv) The head of an arrow should always bear a number higher than the one
assigned at the tail of the arrow.
vi) Gaps should be left in the sequence of event numbering to accommodate subsequent
inclusion of activities, if necessary.
5.2.2. Rules for drawing a network
 A complete network should have only one point of entry -the start event, and one point of
exit -the finish event.
 Every activity must have one preceding event -the tail, and one following event - each
activity has one head.
 Several activities may use the same tail event, and the same head event, but no two
activities may share the same head and tail events.
 Time flows from left to right.
 An activity must be completed in order to reach the end-event.
 Dummy activities should only be introduced if absolutely necessary.

167
Operations Research

5.2.3 Convention for Drawing Networks

In addition to the rules described above, certain conventions are followed for the sake of clarity
and uniformity. There are two slightly different conventions for constructing the net work
diagrams. Under one convention, the arrows are used to designate activities, where as under the
other convention, the nodes are used to designate activities. These conventions are referred to as
activity- on-arrow (A-O-A) and activity – on- node (A-O-N) respectively. In order to avoid
confusion, the discussion here focuses primarily on activity- on- arrow convention. When we use
this convention:
 Networks proceed from left to right -the start event is at the left hand side of the diagram
and the end event at the right hand side
 Networks are not drawn to scale.
 Arrows representing activities should have their head to the right of their tail unless it is
impossible to draw the network in that way.
 Events or nodes should be numbered so that an activity always moves from a lower
numbered event to a higher one. This convention is relatively easy to accomplish in a
simple network but in a complex network it may be necessary to number in tens to allow
for extra activities to be added without the need for a complete renumbering of the whole
diagram
 Lines that cross should be avoided if possible
 The start event may be represented as a line instead of a circle, particularly when several
activities may begin at the start point.
5.2.4. Common Errors in Drawing Networks
There are three types of errors, which are most common in network construction. These are:
a) Formation of a loop: If an activity were represented as going back in time, a closed loop
would occur. In a network diagram looping error is also known as cycling error. Cycling
(looping) in a network can result through a simple error or while developing the activity plans,
one tries to show the repetition of an activity before beginning the next activity.
A closed loop would produce an endless cycle in computer programmers with a built- in routine
for detection or identification of the cycle. Thus one property of a correctly constructed network
diagram is that it is non-cyclic.

168
Operations Research

B
A

b) Dangling: No activity should end without being joined to the end event. If it is not so,
a dummy activity is introduced in order to maintain the continuity of the system.
Such end-events other than the end of the project as a whole are called dangling
events. All activities must contribute to the progression of the network or be
discarded as irrelevant.

B
A E
B
D
B C
B F
Danglin
g
d) Redundancy: If a dummy activity is the only activity emanating from an event, it
can be eliminated.

Dummy D F

A
C
E
B
? Dear learner, can you mention and discuss on the common mistakes made in drawing
net work diagrams.
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
____________

5.2 TYPES OF NETWORK MODELS

Section objective:
Up on completion of this section, the learner will be able to:
 Identify the three models of network

169
Operations Research

 Explain their importance’s

 Distinguish among the three models of network
5.3.1. The Shortest Route Problem
The objective of this network model is to obtain the shortest path in which one can minimize
distance, time or costs involved form the origin to destination.
If we have “n” location in the network, the total steps required to solve the shortest route
problem will be n-1. The determination of shortest route involves labeling procedure in which
each node is assigned with two numbers where by the 1st label represent distance from the
original (source) node and the 2nd number refers the node that immediately precede the labeled
node. The labeling procedure begins with the original node where the label will be (O, S) to
indicate the distance is zero and it is starting point.
In the shortest route algorithm nodes are labeled either permanently to indicate the final labeling
or temporary labeling to indicate that the labeling might be revised. Labels remain temporary
until it can be ascertained that no shorter route to a node exists.
The Shortest Route Algorithm
Step 1: Start at node 1 and find the distance from node 1 to other nodes that can be directly
reached form node 1. Temporarily label each of these nodes with their distance from node 1
Followed by a comma and a number 1. Then select the node that has smallest distance from node
1, and make its label permanent. This can be done by shading the node.
Step 2: Find the distance from the new permanent node to each non permanent nodes that can
reached directly form this node. Temporarily label each of these nodes with the cumulative
distance from node 1 if a node has no label. Or, change the earlier assigned temporary label of a
node if its cumulative distance from node 1 through the new permanent node is less than the
previous temporary label. Then permanently label (shade) the node that has the smallest
cumulative distance form node 1.
Step 3: Repeat the preceding steps until all nodes have permanent labels.
Step 4: Identify the shortest route to each node from node 1 by working backward through the
tree according to the nodes label specified on the node.
Example:
Find the shortest route of the following network starting form node one. Travel between nodes
can be in either direction.

170
Operations Research

2 13

5
5 4
6
7
8 3

1 1
7 16 6

4
1st step
Shade node 1 and temporarily label node 2, 3 and 4, which can be directly reached from node 1.
The label should show the distance from node 1.
(5.1)

2 13
5 4
(O,
( S) (8, 1)
8 7
3 5 6
1

7
1
(7, 1)
16 6
2nd Step 4

Select the node, which has the smallest distance from node 1, and label it permanent and shade
the node (i.e. node 2), then temporarily label nodes that can be directly reached from node 2.
Therefore node 5 is labeled as (18, 2.)

171
 Operations Research

(5, 1)

2
13
5 4 (18,2)
(O, S)
(8,1)
8
3 5
7
7 6
7, 1

1 16
6
3rd Step
Identify the smallest distance from node 1 i.e. from node 3, 4, and 5, which have values of 8, 1,
7,1 and 18,2 respectively. Permanently label and shade node 4, which has the smallest distance
of 7. Identify non-permanent labels; directly reached from this node and temporarily label it by
the distance from node 1.
(5,1)
13
2
5 (18, 2)
(O, S) 4
(8, 1)
7 5
8 3
3
1
6
7 3

(7,1) (23, 4)
1
3
16
4 6

172
Operations Research

4th Step
The node with the smallest temporary label is node 3. Thus, its label becomes permanent and the
node is shaded. Node 3 can be reached directly from node 1 through node 4 with the same
distance of 8 (i.e. tie exists). Next, find each non-Permanent label that can be reached directly
from node 3. Node 5 is the only node with a value of (18, 2). However, the cumulative distance
of node 5 from node 1 through node 3 is 15 (i.e. 8+7=15). Because the route through node 3 is
shorter, we update the temporary label of node 5 to reflect this shorter route.

(5,1)

2
13
5 (15, 3)
(O, S)
(18,2)
(8,1) tie

8 7
1 5
3

7 6
1
(7, 1)

(23,4)
4 16
6

5th Step
Identify the node, which have smallest label from the un-shaded nodes or temporary labels. Node
5 with a distance of 15,3 is selected and becomes permanent label.

173
 Operations Research

(5,1)

2
13
5 4 (15,3)
18,2
(8,1)
(O,S) tie
8 3 7 5
1 6
7 (7,1) 1 (23,4)

16
4 6

6th Step
The only node which is not shaded and can be reached directly from the new permanent node is
node 6. Using this route, its cumulative distance from node 1 would be 15+6=21. Because this is
less than its current table update label to 21, 5 and make permanent & shade it.

13 (15, 3)
18,2
5
4
(O,S)
8, 1 tie

7
8 5

1 3 6
(21,5)
7 (23,4)
(7,1)
1

16 6
4

174
Operations Research

Permanent label of the final network indicates the shortest distance of each node from node 1
(the starting node). In order to identify the route that yields the shortest distance to a particular
node, it is necessary to work backtracking. The shortest route of the nodes from node 1 to each
node is summarized below.
Node Distance Shortest Route Alternative Route
2 5 1-2 -
3 8 1-3 1-4-3
4 7 1-4 -
5 15 1-3-5 1-4-3-5
6 21 1-3-5-6 1-4-3-5-6
5.2.3 Minimum Spanning Tree Model
The minimum spanning tree problem involves in connecting all of the nodes of a network using
as little of the connecting material as possible. For example, the nodes might represent oil
storage tanks, and lines represent pipeline that are used to carry the oil between tanks. The cost
of the pipeline would be proportional to the length of the pipeline used. Hence, the objective
would be to connect all of the tanks using as little pipeline as possible. Similar activities include
designing communication systems by using minimum amount of wiring, designing highway
networks by using minimal amounts of materials etc.
Algorithm of Minimum Spanning Tree
Step 1. Start at any node, (usually, node 1 is used as the starting point), identify the node that
has shortest distance from node 1 and connect it to the node 1 using a line. If a tie occurs, break
it arbitrarily
Step 2. Find the shortest distance form the existing portion of the tree (i.e. the connected nodes)
to a node that is not yet connected. Make the connecting line from previously connected tree to
the new node which has shortest distance.
Step 3. Continue until all nodes have been connected to the tree.
Step 4. To find the total (minimal) length of the connecting distance, sum up their values.
Example:
Consider the following network where the nodes represent fuel storage tanks and the connecting
lines represent possible pipeline connections. The numbers on the lines represent the distance in

175
 Operations Research

meters for a particular pipeline connection. Determine the spanning tree for the network storage
tanks and the amount of pipe that will be needed to make the connections.

22 6
2 26
4
13 18 24

27 33
1 5
20 30
32
18

49 7

3
Arbitrarily choose node 1 as a starting node consider all branches incident on it. They are 1-2 and
1-3 with distance of 13 and 18 respectively. Since 1-2 is the shortest, select this line and connect
it. Now node 1 and 2 are connected.
Next consider all arcs incident either on node 1 or node 2 that connect to other nodes. Such lines
are 1-3, 2-3 and 2-4 with a distance of 18, 20 and 22 respectively. Node 1-3 with a distance of 18
of selected. Now the connected nodes are 1, 2, and 3.
Next consider all braches incident to Node 1, 2 and 3. These are 2-4, 3-4, 3-5 and 3-7 with
distance of 22, 27, 30 and 49. The shortest distance is 22 i.e. 2-4. Make 2-4 part of the network

2 4 6 2 2
13 2

5
1
7 1
18 2
3

A 3
2

176
B
 Operations Research

22
22
2 4 2 4
13 13 6
6

1 18
1 5

3 7
18
5 7
18
C 3

The next step will be to identify the shortest line that is incident to node 3 and 4. These are 3-5,
3-7, 4-5 and 4-6 with a distance of 30, 49, 18 and 26. Since 4-5 is the shortest distance, include it
in the network.
Continue in this manner until all nodes are connected with their shortest distance. The network
solution for the above problem is shown on the following figure
The length of pipe that will be needed for this system can be found by summing the line lengths
(distance): i.e. 18+13+22+18+24+32= 127 meters.

5.3.2 The Maximal Flow Problem

The objective of maximal flow problem is to find the maximum amount of flow of fluid, traffic,
information; etc that can be transported through a capacity limited network. The network is
composed of a source, a sink and connecting arcs of nodes. Flow in each arc is measured by the
amount that can be transported during a unit of time. Flow can be one way only but it is possible
to have two direct arcs between nodes.
Assumption of maximal flow problem

177
Operations Research

 It is assumed that there is a single input node (called a source) and a single output node
(called a sink).
 There is flow conservation, i.e. the flow out of any node is equal to the flow into that
node.
 Junctions or nodes cannot serve as a store, i.e. any flow arriving at a node is transferred
immediately to another location.

Algorithm of Maximal flow model

Rule for Maximal Flow
1. Arbitrarily select a continuous path from the source node to the sink node that has a
positive flow capacity. Flow capacity of a path is determined by the smallest branch
capacity along the path chosen in the direction of flow.
2. In this continuous path determine the arc with the minimum flow capacity.
3. Reduce all the quantities along this path by the amount transported (i.e. minimum of
the path)
4. Repeat step 1- step 3 until all paths are used.
5. To get the maximum flow of the network, add all flows of the path.

Example:
Determine the maximum flow through the system of pipeline. The flow is from node 1 to node 4.

10 2 6

2
1 4

3 6

178
Operations Research

Step 1. Choose any path that will allow a positive flow from the source (node 1) to the sink
(node
4). The possible paths are 1-2-4, 1-2-3-4, 1-3-4 and 1-3-2-4. Suppose arbitrary select path
1-2-4.
The maximum flow of 10 (Node 1-2) is limited by the 6 capacity of node 2-4. Therefore adjust
the capacity of the total path of 1-2-4 by flow of 6. (i.e. reduce by 6)

4 2 6
2

1 4 6
6 6
3

Here, branch 2-4 is fully loaded and no additional flow to this path. But, path 1-2 still accept
additional flow of 4. Therefore, we can use path 1-2-3-4 to pass 2 units through path 1-2-3-4.

2
2 0

0
1
2+6
4
6+2

3
4

179
 Operations Research

The flow capacity of 1-2-3-4 is fully utilized. i.e. no further positive flow is allowed through this
path. But there is a positive flow through path 1-3-4. Therefore a flow of 3 can be assigned to
this path. The resulting network will be as shown below.

2 0
2

2
3+2+6 4 6+2+3
1

0 1

At this point, no additional flow is possible because no path has a positive flow capacity. Hence
the situation is optimal and maximal flow rate is 11 (6+2+3 = 11).

? Dear learner, please discuss on the different type of network models and their application areas.
________________________________________________________________________________
________________________________________________________________________________
________________________________________________________________________________
__

180
Operations Research

CHAPTER SIX
GAME THEORY

Introduction
Unlike a pure monopoly or a perfectly competitive firm, most firms must consider the likely
responses of competitors when they make strategic decisions about price, advertising
expenditure, investment in new capital, and other variables. Although we began to explore some
of these strategic decisions in the last chapter, there are many questions about market structure
and firm behavior that we have not yet addressed. For example, why do firms tend to collude in
some markets and compete aggressively in others'? How do some firms manage to deter entry by
potential competitors? And how should firm’s intake price decisions when demand or cost
conditions are changing, or new competitors are entering the market? To answer these questions,
we will use game theory to extend our analysis of strategic decision making by firms. The
application of game theory has been an important development in microeconomics. This chapter
explains some of this theory and shows how it can be used to understand how markets evolve
and operate and how economic agents should think about the strategic decisions they continually
face. We will discuss how firms can make strategic moves that give them an advantage over their
competitors or the edge in a bargaining situation.
Game theory is perhaps the most important tool in the economists’ analytical kit for analyzing
the strategic behavior. Strategic behavior is concerned with how individuals make decisions
when they recognize that their actions affect, and are affected by, the actions of other individuals
or groups. The Prisoner’s Dilemma is an example of a two-person, non-cooperative,
simultaneous-move, one-shot game in which both players have a strictly dominant strategy A
player has a strictly-dominant strategy if it results in the largest payoff regardless of the strategy
adopted by other players. Nash equilibrium occurs in a non-cooperative game when each player
adopts a strategy that is the best response to what is believed to be the strategy adopted by the
other players. When a game is in Nash equilibrium, neither player can improve their payoff by
unilaterally changing strategies
Dear learner, first, we should clarify what gaming and strategic decision making are all about. In
essence, we are concerned with the following question: If I believe that my competitors are
rational and act to maximize their own profits, how should I take their behavior into account

181
Operations Research

when making my own profit-maximizing decisions? As we will see, this question can be difficult
to answer, even under conditions of complete symmetry and perfect information (i.e., my
competitors and I have the same cost structure and are fully informed about each others' costs,
about demand, etc.). Moreover, we will be concerned with more complex situations in which
firms have different costs, different types of information, and various degrees and forms of
competitive "advantage" and "disadvantage."
The economic games that firms play can be either cooperative or no cooperative. A game is
cooperative if the players can negotiate binding contracts that allow them to plan joint strategies.
A game is non cooperative if negotiation and enforcement of a binding contract are not possible.
An example of a cooperative game is the bargaining between a buyer and a seller over the price
of a rug. If the rug costs $100 to produce and the buyer values the rug at $200, a cooperative
solution to the game is possible because an agreement to sell the rug at any price between $101
and $199 will maximize the gum of the buyer's consumer surplus and the seller's profit, while
making both parties better off. Another cooperative game would involve two firms in an
industry, which negotiate a joint investment to develop a new technology (where neither firm
would have enough know-how to succeed on its own). If the firms can sign a binding contract to
divide the profits from their joint investment, a cooperative outcome that makes both parties
better off is possible. An example of a non cooperative game is a situation in which two
competing firms take each other's likely behavior into account and independently determine a
pricing or advertising strategy to win market share. Note that the fundamental difference between
cooperative and non cooperative games lies in the contracting possibilities. In cooperative games
binding contracts are possible; in non cooperative games they are not. We will be concerned
mostly with non cooperative games. The oligopoly theories we discussed so far are the classical
theory of strategic interaction among firms. Oligopoly economic agents can have been studied by
using the apparatus of game theory. Game theory is concerned with the general analysis of
strategic interaction. It can be used to study:
- Parlor games
- Political negotiation and
- Economic behavior
Thus, in this chapter we will briefly explore this fascinating subject to give you a flavor or how it
works and how it can be used to study economic behavior in oligopolistic markets.

182
Operations Research

6.1 The pay off matrix of a game

Strategic interaction can involve many players and many strategies, but we will limit ourselves to
two-person game with a finite number of strategies. This will allow us to depict the game easily
in a payoff (outcome) matrix. Strategic decisions result in payoffs to the players (Outcomes that
generate rewards for the players). The payoff matrix of a player is a table showing the outcomes
accruing to this player as a result of each possible combination of strategies adopted by his or her
rival(s).
Suppose that -two people (A and B) are playing a simple game and each has two strategies to
play. Let the two strategies of A are top and bottom and that of B left and right. There strategies
could represent economic choices like “raise price” as “lower price” and “don’t advertise” or
“advertise” or they could represent political choices like “declare war” or “don’t declare war”.
The payoff matrix of a game simply depicts the payoffs to each player for each combination of
strategies that are chosen. What will be the outcome of this sort of game?
Consider the following example 1. In this matrix the payoffs to player A are in the first entry and
the payoff to player B are in the second entry in the cells of the box.
A pay off matrix of a game
Player B
Left (don’t adv) Right (advertise)
Top (↑ P) 1,2 0,1
Player A
Bottom 2,1 1,0
(P)

Dear colleague, how can we decide on the best strategy for playing a game? How can we determine
the most likely outcome? We need something to help us determine how the rational behavior of each
player will lead to an equilibrium solution. In the above example, from the viewpoint of person A, it
is always better for him to play bottom. Since his/her payoff from this choice (2 or 1) are always
greater than their corresponding entries in top (1 or 0). Similarly, it is always better for B to play
left. Since 2 and 1 dominate 1 and 0. Thus, we would expect that the equilibrium strategy for A is
to play bottom and B to play left
A dominant strategy is the optimal (best) choice of strategy for each player no matter what the
other player does. Whatever choice B makes, player A will get a higher payoff if he/she plays

183
Operations Research

bottom, so it make sense for A to play bottom. And whatever choice A makes, B will get a higher
payoff if he plays left. Hence, bottom and left dominates the alternatives and we have equilibrium in
dominant strategy.
If there is a dominant strategy for each player, in some game, then we would predict that it would be
the equilibrium outcome of the game. In this example, we would expect an equilibrium outcome in
which A plays bottom, receiving an equilibrium pay off of 2, and B plays left, receiving an
equilibrium pay off of 1.

6.2 Nash Equilibrium

Dominant strategy equilibria are nice when they happen, but they don’t happened all that often. This
is because we may not have dominant strategy equilibrium when the optimal (best) choice of one
depends on what he/she thinks the other will do. If this is the case, we call the equilibrium Nash
equilibrium.
Consider the following example 2.
Nash Equilibrium
Player B
Left Right
1,2 0,0
Top
Player A 0,0 1,2
Bottom

Here when B chooses right, the payoffs to A are 0 or 1. This means that when B chooses left, A
would want to choose top, and when B choose right, A would want to choose bottom. Thus, A’s
optional choice depends on what he thinks B will do and vice versa.
Thus, we will say that a pair of strategies is a Nash equilibrium If A’s choice is optimal given B’s
choice, and B’s choice is optimal given A’s choice. Remember that neither person knows what the
other player will do when he has to make his own choice of strategy. But each person may have
some expectation about what the other person’s choice will be.
Nash equilibrium can be interpreted as a pair of expectations about each person’s choice such that,
when the other person’s choice is revealed neither individual wants to change his behavior.
Therefore,

184
Operations Research

- If A choose top, then the best thing for B is to choose left, since the payoffs to B from
choosing left is 2 and from choosing right is 0.
- And if B chooses left, then A will be better off by choosing top than bottom because (1>0).
Thus (Top, left) is Nash equilibrium. Similarly.
- If A choose bottom, B will choose right and
- If B chooses right, A will choose bottom. Thus (bottom, right) is also Nash equilibrium.
From the above we can conclude that the Nash equilibrium is a generalization of the Cournot
equilibrium for each firm chooses its output level taking the other firm’s choice as being fixed.
 Though Nash equilibrium has certain logic, it has the following problems.
1. The game may have more than one Nash equilibriums. From example 2, bottom and
right can also be another equilibrium.
2. There are games that have no Nash equilibrium of the sort we have been describing at all.
The game on Example 3 has no Nash equilibrium..
Player B
Left Right
0,0 0, -1
Top
Player A 1,0 -1,3
Bottom

The above payoff matrix doesn’t have unique or some Nash equilibrium. And, all of them cannot be
equilibrium outcomes. From the above payoff matrix for example,
- If A chooses top, B will choose left- (top, left)
- If B chooses left, A will choose bottom- (left, bottom)
- If A chooses bottom, B will choose right- (bottom, right)
- If B chooses right, A will choose top- (right, top)
- If A chooses top, B will choose left- (top, left)
Although there is a possibility of no Nash equilibrium in pure strategies there exist Nash
equilibrium in Mixed strategies. A pure strategy is a strategy in which the player makes a
specific choice and sticks to it. Dear student, in the previous section we have seen pure
strategies. Mixed strategies are strategies in which a player makes random choice among two or
more possible actions, based on the set of chosen probabilities.

185
Operations Research

6.3 Mixed strategies

Dear student, so far we have been thinking that each agent as choosing a strategy once and for all (a
one shot game) and sticking to it. This called a pure strategy. Nash equilibrium in mixed strategies
refers to an equilibrium situation in which each agent chooses the optional frequency with which to
play his strategies given the frequency choice of the other agent. Thus, Nash equilibrium in mixed
strategy always exists.
Let’s consider the payoff matrix in Example 3
Player B
Left Right
0,0 0, -1
Top
Player A 1,0 -1,3
Bottom

If A and B follow the mixed strategy, Let player A chooses Top with a probability p and player B
chooses left with a probability q. The computation of mixed strategy payoffs for each player is as
shown below.
Player B
Left(q) Right(1-q) Player A’s expected payoff
Top(p) 0,0 0,-1 0(q)+0(1-q) = 0
Bottom(1-p) 1,0 -1,3 1(q)+ (-1)(1-q)
Player A
Player A’s 0(p)+0(1-p) 0(p)+3(1-p)
Expected payoffs
Thus, the excepted payoff to A from Top is
0(q) +0(1-q) = 0
Thus, the excepted payoff to A from Bottom is
1(q) + (-1)(1-q) = 2q-1
Therefore, Player A chooses bottom if 2q-1 > 0 and chooses Top if 2q-1 < 0.
The same logic applies to player B.
5.3.The Prisoner’s Dilemma
Another problem of Nash equilibrium is that it doesn’t necessarily lead to Pareto efficient outcome.
Consider for example, the game depicted in the table 4. It refers to a situation where two prisoners
who were partners in a crime were being questioned in separate rooms. Each prisoner has a choice

186
Operations Research

of confessing (guilty) in the crime and there by implicating the other or denying (not guilty) that
he/she had participated in the crime.
- If only one prisoner confessed, then he would go free and the authority (judges) will
throw the book (sentenced) at the other prisoner, requiring him to spend 6 months in
prison.
- If both prisoner denied being involved, then both would be held for 1 month on a
technicality (labor work) and
- If both prisoners confessed they would be both held for 3 months. The payoff matrix
for this game is given as
The prisoner, s dilemma
Player B
Confess Deny
-3, -3 0, -6
Confess
Player A -6, 0 -1, -1
Deny

Dear colleague, Put yourself in the position of player A.

- If B decides to deny committing the crime, then you are certainly better off confessing, since
you will get off free.
- Similarly, if B confesses, then you will better off confessing since you will get a sentence of
3 months rather than a sentence of 6 months. Thus, whatever player B does, player A is
better off confessing. The same is true for plays B–he/she is better off confessing. Thus, the
unique Nash equilibrium for this game is for both to confess.
In fact, both players’ confessing is not only a Nash equilibrium; it is dominant strategy equilibrium
since each player has the same optimal (best) choice independent of the other. However, if they
could both just hang tight, they would each be better off. In other words, if they both could be sure
that the other would hold out and both could agree to hold out themselves, they would get a payoff
of -1, which would make each of them better off.
The strategy (deny, deny) is not only Pareto efficient but also Pareto Optimal because there is no
other strategy choice that makes both players better off or either of them with out making the other
worse off. Thus, the strategy (confess, confess) is Pareto inefficient for both. The problem is that
there is no way for the two prisoners to coordinate their action. If each could trust the other, they

187
Operations Research

could both be made better off. This applies to a wide range of economic and political phenomena.
Consider, for example, the problem of arms control. Interpret “confess” as “deploy a new missiles”
and the strategy of “deny” as “don’t deploy”. Note that the payoffs are reasonable. If my opponent
deploys his missile, I certainly want to deploy. But if there is no way to make a binding agreement
not to deploy a missile, we each end up deploying the missile and are both made worse off, which is
Pareto inefficient for both of us. However, if there had been a strong binding agreement that forces
us not to deploy a missile or absolute trust among us, both would have been better off by reducing
the likely hood of human and physical capital loss and using the money in activities that will
enhance economic growth and is not only Pareto efficient but also Pareto optimal for both of us.
Another good example is the problem of cheating in a cartel. Now interpret confess a “produce more
than your quota” and interpret deny as “stick to the original quota”. If one firm (A) thinks the other
firm (B) is going to stick to its quota, it will pay to it to produce more than its quota. And if A thinks
that B will overproduce, then A might as well, too. The prisoner’s dilemma provoked controversy
as to what is a reasonable way to play the game. The answer seems to depend on whether you are
playing a one-shot game or the game is to be repeated an indefinite number of time.

188