Unit I Operation Research Introduction: Characteristics of Operations Research

UNIT I
OPERATION RESEARCH INTRODUCTION
Definitions and Characteristics of Operations Research

Some of the definitions of O R. by leaders and pioneers are as follows:
1. Churchman, Ackoff and Arnoff. “Operations 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 solution to the problems”.
2. Morse and Kimbel. “Operations Research is the scientific method of providing
executive departments with a quantitative basis for decisions regarding the operations under
their control”.
3. Ellis Johnson. “Operations Research is the prediction and comparison of values,
effectiveness and cost of a set of alternative courses of action involving man-machine
systems”.
4. The Operational Research Society of India defines “Operations Research” as the
attack of modern science on complex problems arising in the direction and management of
large systems of men, machines, materials and money in industry, business, government and
defence. The distinctive approach is to develop a scientific model of system incorporating
measurement 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
determine its policy and actions scientifically”.
Considering the above definitions we can define operations research as follows:
“Operations Research is a scientific approach to find optimal solutions to problems of
complex man-machine systems”.
Characteristics of Operations Research
1. Operations Research has a team approach – O.R. is a research carried out by a team
of scientists whose individual members have been drawn from different scientific and
engineering disciplines.
2. Operations Research uses scientific methods, techniques and tools to analyse
executive type problems.
The terms tools, techniques and methods are frequently used interchangeably but in
fact they differ from each other. Scientific tools can be compared with the cutting tools used
in a workshop Scientific techniques indicate how these tools can be used to obtain results.
Scientific methods indicate a plan of using these techniques to achieve the desired objectives.
For example, calculus is a scientific tool. Application of calculus to find the optimal value of
a variable in a mathematical model is a scientific technique and the plan of using
mathematical model to optimize the objective is scientific method.
3. Operations Research aims to help executives to make optimal decisions.
4. Operations Research relies mainly on mathematical model.
5. Operations Research depends largely on electronic computers for the analysis of
mathematical model.
2
Limitations of Operations Research

The limitations of Operations Research are generally the same as for any applied
scientific discipline. Some of the drawbacks crop up because many a time, it has to deal with
people who know very little about its tools and language.
Some important limitations of Operations Research
1. The research nature of undertaking Operations Research is actually a research
of complex industrial and business systems. Hence it is time consuming and the results are
difficult to control and evaluate.
2. The inadequate industrial background of practitioners. Generally the
Operations Research experts have to deal with problems in fields wherein they have no
previous industrial experience. In such situations Operations Research experts have to study
thoroughly the background of the problems. The application of Operations Research without
sufficient background information can only result in failures. The strength of this limitation
will weaken slowly with time as Operations Research experience develops.
3. Poor communication of results. One of the strongest limitations of
Operations Research is the inability to transmit and communicate results to business
executives who have to make decisions.
Generally the solution obtained by an Operations Research team is in the form of
mathematical terminology unknown to managers and executives. This poses a problem of
communication. The Operations Research team or expert has to translate the solution in the
form of procedures, methods and schedules, so that the executives do not face any difficulties
in understanding and executing the same. The Operations Research team generally deals
with problems of complex man-machine systems. The solutions to such problems are equally
complicated. Even if the solution obtained is precise and logical, it may not be feasible.
Feasibility depends on interpretability and workability of the solution. An Industrial engineer
having sufficient knowledge of production processes and procedures can help the
implementation solution to a great extent.
Linear Programming
Introduction
Linear Programming is a formidable Quantitative technique of Decision making.
Such technique is very much useful in the fields of uncertainty, viz., Business and commerce
where decisions are taken on every matter. This technique was introduced for the first time
in 1947 by a Russian Mathematician George B. Dantzig.
The name „Linear Programming‟ consists of the two important terms viz., Linear
and Programming. The term Linear refers to the relationship of the interrelated variables
which is of the form of y = a + bx where x and y are the variables of power one and „a‟ and
„b‟ are constants.
The term programming means planning a way of action in a systematic manner with a
view to achieving some desired optimal results, viz., the minimization, of cost, maximization
of profit etc.
3
Thus Linear Programming is a Mathematical technique which is applied in the form

of a linear formula for arriving at a rational proportion of the variables to be used as inputs to
get the optimum result from a course of action to be planned accordingly.
Definitions
The term Linear Programming has been defined variously by various authors. Some
such definitions are quoted here as under:
According to Dantzig Linear Programming is defined as “a programming of
interdependent activities in a linear structure”.
According to Galton, “Linear Programming is a mathematical technique for
determining the optimal solution of resources and obtaining a particular objective where there
are alternative uses of resources, viz., man material, machinery and money etc”.
From the above definitions it will be clear that linear programming is a mathematical
device of ascertaining the optimal allocation of resources for obtaining the desired objective,
viz., maximization of profit, or minimization of cost where various resources can be used
alternatively.
Types of Problems (where L.P. can be applied)

There are three important types of Problems concerning various fields where linear
programming technique can be applied advantageously. They are:
(1) Problems of allocation
(2) Problems of assignment, and
(3) Problems of transportation.
TYPES OF LINEAR PROGRAMMING PROBLEMS

Linear Programming problems are classified into two types. They are
(1) General or Primal linear programming problem
(2) Duality linear programming problem.
The procedures for determining the desired results under the above two types of
problems are contradicting to each other. These procedures are annotated here as under.
1. General or Primal Linear Programming Problem
Determination of the desired results under this type of problem will involve the
following steps:
Step No. 1. Formulation of the Given Problem
Step No. 2. Solution of the Formulated Problem
Each of the above steps will again require the following sub steps.
Formulation of L.P. Problem

Under this step
(i) Objective function
(ii) Constraint functions, and
(iii) Non – negative functions.
(i) Objective Function. The objective of a problem may be either to maximize or
minimize some result. If it is a case of profit or income or output the objective must be
4
maximization. But if it is a case of loss, cost or input, the objective will be minimization.
For this the rate of profit or cost per variable in issue must be assessed first and then the
number of each variable will be represented in the function through some letters viz. x, y to
be ascertained through the process of solution. As such the objective function will be
presented in the following form:
Z(p) = P1X1 + P2X2 + ......+PnXn
where, Z(p) = Maximum amount of profit
P1, P2, ..., Pn = Rate of profit per different variables to be produced, viz., goods or services
X1, X2, ..., Xn = The number of different variables to be produced under decision.
In case of variables involving cost or loss the objective will be minimization and in
that case the objective function will be formulated as under:
Z(c) = C1X1 + C2X2 + ......+CnXn
where, Z(c) = Minimum amount of cost
C1, C2, ..., Cn = Cost per unit of the variable.
X1, X2, ..., Xn = Different number of the different variable..
(ii) Constraint Function. To accomplish the desirable objective it is necessary to put
some resources, viz., manpower, material, machine or money in the process of production or
performance. But such resources may not be available in unlimited quantity in all the cases.
There are some resources which may be available to a limited extent and thus create
constraints or bottlenecks in the process of performance. There are also some resources
whose availability cannot be obtained below a certain extent and thus compels the
management to procure them in certain larger quantities. However, there might be some
resources, which may be available to the extent just required for the purpose. These
resources, therefore, do not create any obstacle. But the resources which are available up to
or beyond certain limit create constraints in achieving the objective. Thus the objective
function will be adjusted in the light of the given constraints relating to the availability of the
various resources. Thus, the objective function will be flollws by the constraint functions in
the following manner.
Constraints of the
Process – I : a11X1 + a12X2 + ... + a1nXn  b1
Process – II : a21X1 + a22X2 + ... + a2nXn  b2
Process – III : a31X1 + a32X2 + ... + a3nXn = b3
Here a represents the quantity of a particular resource required in particular process,
and b represents the total of quantity of the resources available for a process.
(iii) Non – Negative Function. This function implies that the production or
performance of the variables in issue will never be negative. It will be either zero or greater
than zero but never but never less than zero. This function is therefore represented as under:
X1  0, X2  0, ..., Xn
Thus the formulation phase of a primal linear programming problem will be
constituted as below:
Formulation of L.P.P. (Primal):
Maximize or Minimize Z = C1X1 + C2X2 + ..., CnXn
Subject to constraints:
5
a11X1 + a12X2 + ... + a1nXn  b1

a21X1 + a22X2 + ... + a2nXn  b2
a31X1 + a32X2 + ... + a3nXn = b3
Subject to non-negativity condition that
X1  0, X2  0, ..., Xn  0
Alternatively:
Determine the real numbers X1, X2, ... and Xn such that
a11X1 + a12X2 + ... + a1nXn  b1
a21X1 + a22X2 + ... + a2nXn  b2
a31X1 + a32X2 + ... + a3nXn = b3
X1, X2, ..., Xn  0
and for which expression (Objective Function)
Z = C1X1 + C2X2 + ......+CnXn
may be a maximum or minimum.
Optimal Product Line Problem: An optimal product line problem is one which
needs decision as to how much of „n‟ different products should a firm produce or sell when
each of the products require a particular proportion of various resources, viz., material,
labour, machine hour etc. the supply of which are limited to a certain extent.
Illustration: A firm produces two types of products P and Q through two processes,
viz., Foundry and Machine shop. The number of manhours required for each unit of P and Q
in each of the processes and the number of manhours that can be availed at best in the two
processes are given as follows.
Foundry process Machine process
Product P 10 units 5 units
Product Q 6 units 4 units
Available at best 1000 units 600 units
Net profits expected from each unit of the product are: A – Rs. 50 and B – Rs. 40.
Formulate the problem for solution to arrive at the optimal number of the two products P and
Q to be produced.
Solution: Here the problem obviously involves the maximization of profits. Thus the
formulation of the problem will be made in the following order:
1st Step: Notation
Let Z = Total of maximum possible Net
profit
X1 = No. of product P to be produced
X2 = No. of product Q to be produced
F = Foundry process
M = Machine shop process
6
2nd Step: Decision Table

In this step the given data will be arranged in a table in the following order:
F (process) units M (process) units Net profit per
Product Decision variable
of man hour of man hour unit (Rs.)
P X1 10 5 50
Q X2 6 4 40
Maximum Labour
1000 600 units
hours available
3rd Step: Constitution of the different linear functions

Maximize profit Z =50X1 + 40X2
(ii) Constraint functions
(a) Foundry constraints 10X1 + 6X2  1000
(b) Machine shop constraints 5X1 + 4X2  600
(iii) Non – negative functions: X1  0, X2  0
th
4 Step: Formulation of L.P.P.
Determine the real numbers X1 and X2 such that
10X1 + 6X2  1000
5X1 + 4X2  600
X1X2  0
and for which the objective function Z = 50X1 + 40X2 may be a – maximum.
Diet Problem: A diet problem is one in which decision is taken as to how
much of „n‟ different foods to be included in a diet given the cost of each good and the
particular combination of nutrient each food contains. Here, the objective is to minimize the
cost of diet such that it contains a certain minimum amount of each nutrient.
Illustration: A poultry firm contemplates to procure four special feeds in a
combination which would provide the required vitamin contents and minimize the cost as
well. From the following data formulate the linear programming problem.
Units of vitamins A, B, C in each feed Feed Cost
Feed
A B C (Rs.)
P 4 1 0 2
Q 6 1 2 5
R 1 7 1 6
S 2 5 3 8
Minimum vitamin contents needed per feed mix in units

A – 12
B – 14
C–8
st
Solution: 1 Step: Notations
Let Z = Total of minimum possible costs
X1 = Decision variable for the feed P
7
X2 = Decision variable for the feed Q

X3 = Decision variable for the feed R
X4 Decision variable for the feed S
2nd Step: Decision Table
Decision Vitamins constraints Feed Cost
Feed
variable A B C (Rs.)
p X1 4 1 0 2
Q X2 6 1 2 5
R X3 1 7 1 6
S X4 2 5 3 8
Minimum needed 12 14 18
rd
3 Step: Constitution of the different Linear Functions.
Minimize cost
Z =2X1 + 5X2 + 6X3 + 8X4
Vitamin A constraints
4X1 + 6X2 + X3 + 2X4  12
Vitamin B constraints
X1 + X2 + 7X3 + 5X4  14
Vitamin C constraints
0X1 + 2X2 + X3 + 3X4  8
(iii) Non – negative functions:
X1, X2, X3,X4  0
4th Step: Formulation (Table) of the L.P.P.

(Minimize cost) Z =2X1 + 5X2 + 6X3 + 8X4
Subject to the constraints
4X1 + 6X2 + X3 + 2X4  12
X1 + X2 + 7X3 + 5X4  14
0X1 + 2X2 + X3 + 3X4  8
and subject to the non – negativity condition that
X1, X2, X3,X4  0
Transportation Problem: A transportation problem is one in which decision is taken
over a shipping schedule for particular goods, viz., Pig iron, Petroleum, Cotton, Jute etc. from
each of a number of production centres at different location to each of a number of marketing
centres at different location in such a manner that it would minimize the total shipping cost
subject to the constraints.
(i) that demand at each market will be satisfied and
(ii) that supply at the production centre will not be exceeded.
8
2nd Step: Decision Table.

Marketing Centre constraints
Production Decision No. of units that
M N O P
Centre variables can be supplied
X1 X2 X3 X4
D X1 5 15 7 6 50
E X2 8 7 9 1 60
F X3 15 25 30 40 90
200
No. of units that can be
30 40 60 70 Total of supply
demanded
and demand
3rd Step: Constitution of the different Linear Functions.

(i) Objective function. The objective here is to minimize the cost which can be
represented as thus
Minimize cost Z =5X11 + 15X12 + 7X13 + 6X14 + 8X21
+ 7X22 + 9X23 + X24 + 15X31 + 25X32
+ 30X33 + 40X34
Supply constraints
D = X11 + X12 + X13 + X14 = 50
E = X21 + X22 + X23 + X24 = 60
F = X31 + X32 + X33 + X34 = 90
Demand constraints
M = X11 + X21 + X31 = 30
N = X12 + X22 + X32 = 40
O = X31 + X23 + X33 = 60
P = X14 + X24 + X34 = 70
(iii) Non – negative functions:
X11, X12, X13,... etc.  0
4th Step: Formulation of the L.P.P.

Minimize cost (Z) =5X11 + 15X12 + 7X13 + 6X14 + 8X21
+7X22 + 9X23 + X24 + 15X31 + 25X32
+ 30X33 + 40X34
X11 + X12 + X13 + X14 = 50
X21 + X22 + X23 + X24 = 60
X31 + X32 + X33 + X34 = 90
X11 + X21 + X31 = 30
X12 + X22 + X32 = 40
9
Solution of the Formulated Problems

After a linear programming problem has been properly formulated, the next step is to
attempt at its solution to determine the values of the different decisive variables, viz., X 1, X2,
X3 etc. depicted in the formulation of the said linear programming problem.
Solution of the L.P.P. may be of three types, viz. :
(1) Feasible solution,
(2) Non – Feasible solution, and
(3) Optimal solution.
(1) Feasible Solution: A solution which satisfies the non – negativity conditions of
a general L.P.P. is called feasible solution.
(2) Non – Feasible Solution: A process of solution which does not satisfy the non
– negativity conditions of a general L.P.P. is called Non – feasible solution.
(3) Optimal Solution: A feasible solution which optimizes (minimizes or
maximizes) the objective function of a general L.P.P. is called an optimum solution of a
general L.P.P.
Methods of Solution of L.P.P.
There are two methods of solving a linear programming problem involving
allocations of resources viz.:
(1) Graphic Method, and
(2) Simplex Method.
(1) Graphic Method. A linear programming problem which involves only two
decisive variables, viz, X1 and X2 can be easily solved by Graphic method. But a problem
which involves more than two decisive variables cannot be solved by the Graphic Method.
This is because a graph ordinarily has two axis only, i.e., horizontal and vertical and thus
more than two decisive variables cannot be represented and solved through a graph.
Procedure for Graphic Solution: For solution of a L.P.P. through a graph the
following steps are to be taken up one after another.
1st Step: Formulation of the L.P.P.
2nd Step: Conversion of the constraints functions into the equations and determination
of the values of each of the variables under each equation by assuming the other variable to
be zero.
3rd Step: Drawal of the 1st quadrants of the graph in which only positive values of both
the variables and plotted on the basis of the non – negativity condition, i.e., X1, X2  0.
4th Step: Plotting of each set of points on the graph for the pair of values obtained
under each of the equations and joining them differently by straight lines.
5th Step: Identification of the feasible region through shaded area which satisfies all
the constraints. For “less than or equal to constraints” such region will the below all the
constraint lines but for “greater than or equal to constraints” the said region will lie above all
the constraint lines.
6th Step: Location of the corner points or the extreme points of the feasible region.
7th Step: Evaluation of the objective function at each of the corner points through the
following table.
10
Evaluation Table
Values of the Variables Objective Function Total of Values of
Corner Points
X1 X2 Z = RX1 + RX2 Z
If the objective function relates to maximization, the corner point showing the
maximum value in the above table will give the optimal solution for the values of X1 and X2.
On the other hand, if the objectives function relates to minimization, the corner point showing
the minimum value in the above table will give the optimal solution for the value of X1 and
X2.
Alternative Method of Evaluation
(i) Plot a line for the objectives function assuming any value for it so that it falls within
the shade area of the graph. Such line is known as iso–profit line or ISO – Cost line.
(ii) Move this iso–profit line parallel to itself and farther (closer) from (to) the origin till
it goes completely outside the feasible region.
(iii) Identify the optimal solution as the co-ordinates of that pair on the feasible region
touched by the highest possible profit line (lower possible cost line).
(iv) Read the optimal co-ordinates of X1 and X2 from the graph and compute the profit
or the cost.
Illustration (On Maximization): A firm proposes to purchase some fans and sewing
machines. It has only Rs. 5760 to invest and space for at most 20 items. A fan costs Rs. 360
and a sewing machine Rs. 240. Profit expected from a fan is Rs. 22 and a sewing machine is
Rs. 18. Using graphic method of solution determines the number of fans and sewing
machines he dhows purchase to maximize his profit. Also ascertain the maximum possible
profit he can ear.
Solution: 1st Step: Formulation of the Problem.
Formulation of the problem can be made directly by the following two steps.
(i) Decision Table
Constraints Profit per unit
Decision
Articles Investment
Variables Space Rs.
(Rs.)
Fan X1 360 1 22
Sewing Machine X2 240 1 18

Maximum
 5760  20
Capacity
(ii) Formulation of L.P.P.

Maximize (Profit) Z = 22X1 + 18X2
Subject to Constraints
(1) 360X1 + 240X2  5760
(2) X1 + X2  20 and
11
Subject to non – negativity condition

X1, X2  0
2nd Step: Conversion of constraints into equations and determination of the values of
the ordinates.
(i) 360X1 + 240X2 = 5760
Here, when X1 = 0 X2 = 24
when X2 = 0 X1 = 16
(ii) X1 + X2 = 20
Here, when X1 = 0 X2 = 20
when X2 = 0 X1 = 20
rd th
3 Step to 6 Step
Graphic Representation of the Constraints lines of the L.P.P.
7th Step: Evaluation of the Objective Function

Values of the
Objective Function Total of Values of
Corner Points Co-ordinates
X1 X2 Z = 22X1 + 18X2 Z
P 0 0 22  0 + 18  0 = 0
Q 0 20 22  0 + 18  20 = 360
R 0 12 22  8 + 18  12 = 392 (Max)
s 16 0 22  16 + 18  0 = 352
Hence the company should purchase 8 units of X1 i.e. Fan and 12 units of and X2 i.e.,
sewing machine to make the maximum profit of Rs. 392.
Illustration: (On minimization): A firm produces three different products, viz., R, S
and T through two different plants, viz., P1 and P2 the capacities of which in number of
products per day are as follows:
Product R Product S Product T
P1 3000 1000 2000
P2 1000 1000 6000
The operating cost per day of running the plants P1 and P2 are Rs. 600 and Rs. 400
respectively. The expected minimum demands during any month for the products R, S and T
12
are 24000 units, 16000 units and 4800 units respectively. Show, by Graphic method how
many days should the firm run each plant during a month so that the production cost is
minimized while still meeting the market demand.
Solution
1. Notation
Let Z = objective function (Minimization of cost)
X1 X2 = number of working days of the plants P1 and P2 respectively.
2. Decision Table.
Name of Decision Constraints of the Products Cost per
Plants variables R units S units T units day (Rs.)
P1 X1 3000 1000 2000 600
P2 X2 1000 1000 6000 400
Minimum demands 24000 16000 48000
(iii) Formulation of L.P.P.

Objective Function:
Z(Minimize) = 600X1 + 400X2
Subject to
3000 X1 + 1000 X2  24000
1000 X1 + 1000 X2  16000
2000 X1 + 6000 X2  48000
and X1 + X2  0
4. Conversion of the constraints into equations and determination of the values of the
different sets of ordinates.
(i) 3000 X1 + 1000 X2 = 24000
Let, X1 = 0 then X2 = 24
X2 = 0 then X1 = 8
(ii) 1000 X1 + 1000 X2 = 16000
Let, X1 = 0 then X2 = 16
X2 = 0 then X1 = 16
(iii) 2000 X1 + 6000 X2 = 48000
Let, X1 = 0 then X2 = 8
X2 = 0 then X1 = 24
13
5. Graphic Representation of the Constraints lines of the L.P.P.
6. Evaluation of the Objective Function by corner points.

Co-ordinates Objective Function Total of
Corner Points
X1 X2 Z = 600 X1 + 400 X2 Z
A 0 24 600  0 + 400  24 9600
B 4 12 600  4 + 400  12 7200 (Min.)
C 12 4 600  12 + 400  4 8800
D 24 0 600  24 + 600  0 14400
From the above evaluation table it comes out that the optimal solution lies at the
corner point B, where the total cost is minimum, i.e., 7200. Hence the firm should run the
plant I, for 4 days and plant II for 12 days to minimize its cost and to meet with the expected
demand of the market as well.
14
UNIT II
SIMPLEX METHOD
L.P.P. involving three or more variables cannot be solved by graphical method. Such
problem can be solved by a method mown as Simplex Method. To use this method a L.P.P.
given in canonical form must be converted into a form called standard form in which all
constraints are of equality type.
Inequality constraints are changed into equations by adding or subtracting a non-
negative variable. If the constraint is , a new variable called slack variable is added and if
the constraints is  a new variable, called surplus is subtracted
Ex: If a1x1 + a2x2  b
a1x1 + a2x2 + a3 = b,x3  b0 is the slack variable
If a1x1 + a2x2  b then
a1x1 + a2x2 – a4 = b, x4  b 0 is the surplus variables.
Used slack or surplus variables, a general L.P.P. can be written in the standard form
as follows”
Maximize or Minimize
z = c1x1 + c2x2 + ......+ cn xn ... (1)
a11x1 + a12 x2 +......+ a1n x = b1
a21x1 + a22 x3 +......+ a2n x = b2 ... (2)
...............................................
am1x1 + am2 x2 +......+ amn x = bm
xj = 0, j = 1,2, ....... n ... (3)
Here b1, b2, ...... bm  0
There are m constraints and n variables (n > m)
A Basic Solution to (2) is obtained by putting ( n – m) variables equal to zero and
solving for the remaining m variable. Then variables are called basic variables. The (n – m)
variables which are made to zero are called non basic variables.
If all the basic variables are non-negative them it is called a basic feasible solution. If
all the basic variables are positive, then the solution is called a non-degenerate basic feasible
solution. If one or more of the m basic variable are zero, the solution is called a degenerate
basic feasible.
For a n variable, m constraints, (n > m) L.P.P. we can find n < m basic solutions. By
a well known theory is it can be proved that that one of the basic solution which is feasible
will optimize the given objective function. It will be very difficult to find the n < m different
solution and than identify the optimum solution among them. It will be time consuming.
Simplex method gives us an elegant way of converging to the optimum, solution, starting
from a basic feasible solution we check, whether this solution is optimum if not we identify a
non-basic variable which may improve the solution and introduce that as a new basic variable
acid drop one of the present basic variables. We repeat the process till we get the optimum
solution.
The simplex algorithm is given as follows:
15
Step 1: Write the L.P.P. in the standard form

Step 2: Obtain an initial basic feasible solution, by making (n-m) of the variables
equal to zero let be the m basic variables and let , be their cost
coefficients in the objective function.

Step 3: Compute the net evaluations zj – cj = 1,2,...,1,2,......mj the n variables, by using
the relation.
where yij are the coefficient of the variable xj in the m
constraints.
Step 4: (i) For a maximization problem, if all Zj – cj  0 then the initial basic feasible
solution is an optimum basic feasible solution.
(ii) For a minimisation problem, if all Zj – cj  0 then the initial basic feasible solution
is an optimum basic feasible solution.
(iii) It atleast one Zj – cj < 0 for a maximization problem or Zj – cj > 0 for a
minimization problem go to Step 5.
Step 5: (i) For a maximization problem if there are more than one negative Zj – cj,
then choose the most negative of them. Let it be Zr – cr, for j = r. If yir  0, (i = 1, 2, ...... m)
then there is an unbounded solution to the given L.P.P.
If at least one yir > 0, (i = 1, 2, ...... m) then the non-basic variable xr will become the
new basic variable.
For minimization problem, if zj – cj, is the most positive net evaluation, the xr will
become the new basic variable.
Step 6: Compute the ratios be and chose the minimum
of them.
Let the minimum of these ratios be, Then the variable xk, will leave the set of
basic variables.
The common element ykr, which is in the kth row and rth column is known as the
pivotal element of the simplex table.
Step 7: Convert the pivotal element to unity by dividing its raw by the leading
element. Itself and all other elements in its column to zeroes by making use of the relations.
Step 8: Go to step 3 and repeat the computations procedure until either an optimum
solution is obtained or there is an indication of an unbounded solution.
16
Example 1:
Solve the L.P.P.
Maximize z = 5 x1 + 3 x2
x1 + x2 + 2
5 x1 + 2 x2 10
3 x1 + 8 x2 12
x1, x2  0
Let us first convert the problem to standard form by introducing slack variables.
Since the slack variables do not contribute anything to the objective function, take the cost for
all such variables as zero in the objective function.
Standard form
Maximize z = 5 x1 + 3 x2 + O x3 + O x4 + O x5
Subject to x1 + x2 + x3 = 2
5 x1 + 2 x2 + x4 = 10
3x1 + 8x2 + x5 = 12. xj > 0 j = 1, .... 5
An initial basic solution can be obtained by making the variables x1 and x2 equal to
zero. Then the basic variables are x3 = 2. x4 = 10, x5 = 12.
The constraint equations, the basic feasible solution and the costs of the variables can
be put in the form of a table called simplex table.
xB denotes the set of basic variables
CB the cost corresponding to the basic variables.
5 3 0 0 0
CB XB x1 x2 x3 x4 x5 xBi/y11, y11 > 0
0 x3 = 2  1 1 0 0

0 x4 = 10 5 2 0 1 0
0 x5 = 12 3 8 0 0 1
Net evaluations -5 -3 0 0 0
zj - cj 
Note that the next evaluation corresponding to the variables x1 is evaluated as follows.
zj - cj = (0) (1) + (0) (5) + (0) (3) – 5 = 0 – 5 = –5
Similarly the next evaluations corresponding to the other variables are evaluated. x1
and x2 have negative net evaluations and hence the current solution is not optimum. We have
to find a new basic feasible solution. Among the net evaluations -5 is the most negative.
17
This corresponds to the variable x1. Hence x1 must become a basic variable. To identify the
leaving basic variable, evaluate the ratio
i.e., evaluate They are 2, 2, 4. The minimum of these ratios as 2. They correspond
to x3 and x4. Therefore we can drop either x3 and x4 a basic variable. We can use a tie
breaker and drop x3 as a basic variables. The 1 becomes the pivotal or leading element.
Since it is already unit make the other elements in that column to zero. as a basic variables.
Note that the column corresponding to the leaving variables x 3 has elements 1.0, 0. The new
basic variables xi should have the same elements 1, 0, 0. This can be done by the following
row operations, which is the same as the formula given in Step 7 of the simplex algorithm.
In the new simplex table, the variable x3 will be replaced by x1. Since the pivotal
element is 1 there is no change, in the first row. Row 2 of the new table can be obtained as
follows”.
5R1 = - 10 -5 -5 -5 0 0
R2 = 10 5 2 0 1 0
-5R1+R2 = 0 0 -3 -5 1 0
Similarly the new 3rd row can be obtained by –3R1 + R3. = -3R1 + R3.
-3R1 = - 6 -3 -3 -3 0 0
R3 = 12 3 8 0 0 1
The new simplex =6 0 5 -3 1 0 table then will be
CB XB 5 3 0 0 0
x1 x2 x3 x4 x5
5 x1 = 2 1 1 1 0 0
0 x4 = 0 0 -3 -5 1 0
0 x6 = 6 0 5 -3 0 1
zj - cj 0 2 5 0 0
Since all zj - cj  0 the current solution is optimum, x1 = 2 and x2 = 0 (non-basic
variable x2 has zero value) is the optimum solution and maximum z = 5(2) + 310) = 10.
Example 2
Maximize z = 4x1 + 3x2
Subject to
2x1 + 3x2  6
-3x1 + 2x2  3
2x2  5
2x1 + x2  4
Standard form
18
Maximize z = 4x1 + 3x2 + 0.x3 + 0.x4 + 0.x5 +0.x6

Subject to 2x1 + 2x2 + x3 = 6
-3x1 + 2x2 + x4 = 3
2x2 + x5 = 5
2x1 + x2 + x6 = 4, xj  0, j = 1, 2 ... (6)
A basic feasible solution is x1 = 0, x2 = 0, .x3 = 6, x4 = 5, x6 = 4.
Simplex table
4 3 0 0 0 0 W
CB XB x1 x2 x3 x4 x5 x6 xB1/y11 y11>0 hen we
0 x3 = 6 2 3 1 0 0 0 evaluate
the ratio
xB1./y1j,
0 x4 = 3 -3 2 0 1 0 0 - the
0 x5 = 5 0 2 0 0 1 0 -  radio
x6 = 4  1 0 0 0 1 should
be
evaluate
zj - cj -4 -3 0 0 0 0 d only if
the
element
s yij is positive. Since y21 = -3. y31 = 0, the corresponding ratio‟s need not be evaluated. The
minimum positive ratio is 2 and hence x1 should become the new basic variable in the place
of x6 2 is the pivotal element.
New Simplex table is obtained by the following row operations.
In the above row operations, R1 should not be obtained by the tow operation R4 – R1.
In this case x3 will become negative, which is not feasible.
4 3 0 0 0 0
CB XB x1 x2 x3 x4 x5 x6 xBi/y11, y11 > 0
0 x3 = 2 0  1 0 0 -1 22=1
0 x4 = 9 0 7/2 0 1 0 3/2
19
0 x5 = 5 0 2 0 0 1 0 5  5/2
0 x1 = 2 1 ½ 0 0 0 ½ 9  7/2 = 4
zj - cj 0 -1 0 0 0 2

-1 is the only negative net evaluation, x2 enters as a basic variable. The minimum of
the ratios xB1/y12 is 1  3 is dropped as a basic variable.
 is the pivotal element. The new simplex table is obtained by the following row
operations.
4 3 0 0 0 0
CB XB x1 x2 x3 x4 x5 x6
0 x1 = 1 0 1 ½ 0 0 ½
0 0 0 1 0 13/2
x4 =
0 x5 = 3 0 0 -1 0 1 1
0 x5 = 3/2 1 0 0 0 3/4
zj - cj 0 0 ½ 0 0 3/2
Note that the next evaluation corresponding to the variables x1, x2, x4, x5, are all equal
to zero. For the non basic variables x3 and x6, the net evaluations are positive. Hence the
current solution is the optimum solution.
The optimum solution is x1 = 3/2. x2=1.
Maximize z = 4(3/2) + 3(1) = 9
Example 3
Maximize z = x1 + 3x2 + 2x3
3x1 + x2 + 2x3  7
-2x1 + 4x2  12
4x2 + 3x2 + 8x3  10
x2 , x2, x2  0
Standard form
Minimize z = x1 + 3x2 + 2x3 + 0.4x4 + 0.x5 + 0.x6
Subject to 3x1 - x2 + 2x3 + x4 = 7
- 2x1 + 4x2 + x5 = 12
-4x1 + 3x2 + 8x3 + x6 = 10
xj = 0, j = 1, 2 ... (6)
A basic feasible solution is x1 = 0, x2 = 0, .x3 = 0, x4 = 7, x5 = 12, x6 = 10.
20
Simplex table
4 -3 0 0 0 0
CB XB x1 x2 x3 x4 x5 x6 xB1/y11 y12 > 0 
0 x4 = 7 3 -1 2 0 0 0
0 x5 = 12 -2  0 1 0 0 12  4 = 3 
0 x6 = 10 -4 3 0 0 0 1
zj - cj -1 +3 -2 0 0 0
For a minimization problem, optimization is obtained when all zj - cj  0. If we have

+ive net evaluations, the variable corresponding to the most +ive net evaluation will become
the new basic variable. Here 3 is the only +ive net evaluation. Hence x 2 becomes the new
basic variable. Minimum of the ratios xB1/y12, y12 > 0 = 3. That corresponds to the variable
x5. Hence x2 enters as the basic variable in the place of x5.
4 is the pivotal element. The new simplex table can be obtained by the following row
operations.
New Simplex table

4 -3 0 0 0 0
CB XB x1 x2 x3 x4 x5 x6 xB1/y11 y12 > 0
5/2
0 x4 = 10 0 2 1 ¼ 0
0 x2 = 3 -½ 1 0 0 ¼ 0 -
0 x6 = 1 -5/2 0 8 0 -¾ 1
zj - cj -1 0 -2 0 -¾ 0
x1 enters as a basic variable. In the place of x4. To get the new simplex table, use the
following row operations.
The new simplex table is
4 3 0 0 0 0
0 x3 = 4 0 0 4/5 2/5 0 0
0 x2 = 5 0 1 2/5 1/5 0 0
0 x6 = 11 0 0 10 1 1 1
21
zj - cj 0 0 -12/5 -4/5 -4/5 0
Since all zj - cj  0,
x1 = 4, x2 = 5, x3 = 0 (non-basic variables has the value zero) and minimum z = 4(1) +
5(-3) + (0) (2) = - 11.
Example 4
A company makes products A, B and C which flows through three departments:
Milling, lathe and grinder. The variable time per unit of different products are given below in
hours.
Product Milling Lathe Grinder
A 8 4 2
B 2 3 0
c 3 0 1
The milling, lathe and grinder machines can work for 250. 150 and 50 hours
respectively per week. The unit profit would be Rs. 20, Rs. 6 and Rs. 8 respectively for
products A, B and C. Find how much of each product the company should produce in order
to maximize profit?
Solution
Let the company produce x1 units of A. x2 units of B and x3 of C.
Maximize profit z = 20x1 + 6x2 + 8x3 subject to the following time constraints
For milling machine
8x1 + 2x2 + 3x3  250
For lathe 4x1 + 3x2  150
For grinder 2x1 + x3  50. x1, x2, x3,  0
Standard form
Maximize z = 20x1 + 6x2 + 8x3 + 0.x4 + 0.x5 + 0.x6 Subject to
8x1 + 2x2 + 2x3 + x4 = 250
4x1 + 3x2 + 0.x3 + x5 = 150
2x1 + 0.x2 + x3 + x6 = 50
xj  0, 1,2, ....... 6
A basic feasible solution is
x1 = 0, x2 = 0, x4 = 250, x5 = 150, x6 = 50
Simplex table is
20 6 8 0 0 0 
CB XB x1 x2 x3 x4 x5 x6 xB1/y11 y11 > 0
0 x4 = 250 8 2 3 1 0 0 250  8 = 31.25
0 x5 = 150 4 3 0 0 1 0 150  4 = 37/5
0 x6 = 50 2 0 1 0 0 1 50  2 =25 
zj - cj -20 -6 -8 0 0 0
x1 enters as a basic variable in the place of x6 2 is the pivotal element.

The row operations to form the new table are
22

20 6 8 0 0 0
CB XB x1 x2 x3 x4 x5 x6 xB1/y12 y12 > 0
0 x4 = 50 0 2 -1 1 0 -4 50  2 = 25
0 x5 = 50 0 3 -2 0 1 -2 50  3 = 16.67 
0 x6 = 25 1 0 ½ 0 0 ½ -
zj - cj 0 -6 2 0 0 10

x2 enters as a basic variable in the place of x5 3 is the pivotal element.
The new row operations to get the simplex table are
20 6 8 0 0 0
CB XB x1 x2 x3 x4 x5 x6 xB1/y13 y13 > 0
0 0 0 1/3 1 -2/3 -8/3
0 0 1 -2/3 0 -2/3 -2/3 -
0 x1 = 25 1 0 ½ 0 ½ ½ 
zj - cj 0 0 -2 0 2 6

x3 enters- as a basic variable in the place of x4.
The row operation are
The new table are
20 6 8 0 0 0
0 x4 = 0 -2/3 0 0 1 -2/3 -3
23
0 x2 = 50 4/3 1 0 +1/3 0
0 x3 = 50 2 0 1 0 0 1
zj - cj 4 0 0 0 +2 +8
Since all zj - cj  0, current solution is optimum

Solution is x1 = 0; x2 = 50, x3 = 50
Maximum z = 0 + 6(50) + 8(50) = Rs. 700
Artificial Variables
If an L.P.P. has constraints with  or = sign, then the simplex method will not give us
a basic feasible solution, because we will end up with negative solution, which is not feasible.
In such cases, after converting the inequality to equality by subtracting a surplus variable, add
another variable to the L.H.S. to get a basic feasible solution. Such a variable that is added to
the L.H.S is called artificial variable. For example, if we have a constraint a, x, + a x, > b,,
convert this to equality by subtracting the surplus variable x3.
i.e., a1x1 + a2 x2 – x3 – b1
Add another variable x4 to the LHS. Then we get a1x1 + a2 x2 – x3–x4 = b1. This will
enable us to get a positive basic solution, x4 is called artificial variable. For problems with 
or = sign, we can make use of
i) Big–M method or charmer‟s Penalty method
ii) Two–phase method.
Big–M Method
For problems with  or = sign, we add artificial variables to the constraints. Their
addition causes violation of the corresponding constraints. This difficulty is overcome by
ensuring that the artificial variables will be zero in the final solution. This is achieved by
assigning a very large per unit penalty to these variables in the objective function. Such a
penalty will be designated by – M for maximization problems and + M for minimization
problems. Then the L.P.P. is solved by simplex method as usual. But we must ensure that
the artificial variables should not be present in the optimum solution at non-zero level. If an
artificial variable is present at non-zero level, that indicates the L.P.P. has no solution. This
technique is called Big-M method.
Example
Solve the following L.P.P. by Big-M method
Maximize z = 8x2 - 5x3
Subject to x1 + x3  2
2x1 + x2 + 6x3  6
x1 - x2 + 3x3 = 0
x1, x2, x3  0
The constrains is standard form are

x1 + x3 – x4 = 2
2x1 - x2 + 6x3 + x6 = 6
x1 - x2 + 3x3 = 0
24
Introducing artificial variables x5 and x7 in the first and third equation, we get
x1 + x3 – x4 + x5 = 2
x1 – x2 + 3x3 + x7 = 0
The L.P.P. in standard from will be
Maximize z = 0.x1 + 2x2 – 5x3 + 0.x4 – Mx5 +0x6 – Mx7
Subject to
x1 + 0.x2 – x3 – x4 + x5 = 6
2x1 + x2 + 6x3 + x6 =6
x1 – x2 + 3x3 + x7 =0
xj = 0. j = 1,2,.....7
x1 = 0, x3 = 0, x4 = 0, x5 = 2, x6 = 6 (x6 = 6) = x7 = 0
Simplex table is
0 2 –5 0 –M 0 M 
x3 CB XB x1 x2 x3 x4 x5 x6 x7 xB1/y11 y11 > 0
–M x5 = 2 1 0 1 –1 1 0 0
0 x6 = 6 2 1 6 0 0 1 0 6/6 =1
–M x7 = 0 1 –1 3 0 0 0 1 0/3 = 0 
zj - cj – M– –
M 0 0 0
2M 2 4M+5
enters as a basic variable in the place of x7 3 is the pivotal element.
The row operations to form the new simplex table are
0 2 –5 0 –M 0 M 
–M x5 = 2 2/3 1/3 0 –1 1 0 –1/3 22/3 =3
0 x6 = 6 0 3 0 0 0 1 –2 –
–5 x7 = 0 1/3 –1/3 1 0 0 0 1/3 0
zj - cj – 4/3
–M–2
3M 0 , 0 0 M–
+11/3
–5/3 5/3
enters as a basic variable in the place of x3 1/3 is the pivotal element.

The row operations are
25
The simplex table is

0 2 –5 0 –M 0 –M 
CSince
B XxB5 and x6 have
x1 the same
x2 ratio
x3 2, we
x4 canxdrop
5 x6 x5xor
either 7 x6xas
B1/y
a 11 y11 variable.
basic >0 But preference s
–M x5 = 2 0 1 –2 –1 1 0 0 21=3
0 x6 = 6 0 3 0 0 0 1 –2 63=2
0 x7 = 0 1 –1 3 0 0 0 1 – 
x4 zj - cj 2M+
0 –M–2 M 0 0 M
5
0 2 –5 0 –M 0 –M
CB XB x1 x2 x3 x4 x5 x6 x7 xB1/y11 y11 > 0
2 x2 = 2 0 1 –2 –1 1 0 0
0 x6 = 0 0 0 6 3 –3 1 –2
0 x1 = 2 1 0 1 –1 1 0 1 –
zj - cj 2+
0 0 1 –2 0 M
M
enters as a basic variable in the place of x6 3 is the pivotal element.

The row operations are;
0 2 –5 0 –M 0 –M
CB XB x1 x2 x3 x4 x5 x6 x7
x2 = 2 0 1 0 0 1 1/3 –2/3
0 x4 = 0 0 0 2 1 –1 1/3 –2/3
0 x1 = 2 1 0 3 0 0 1/3 –2/3
zj - cj 0 0 5 0 M 2/3 M–4.3

Since all zj - cj  0, current solution is optimum x1 = 2, x2 = 2, x3 = 0 the optimum
solution.
Maximum z = 2(0) + 2(2) + 0(–5) = 4
Example
Using Big-M method, solve the L.P.P.
Minimize z = 4x1 + x2
3x1 + x2 = 3
4x1 + x2  6
x1 + 2x2 + 3, x1, x2  0
The standard form is
Minimize z = 4x1 + x2 + Mx3 + 0.x4 + Mx5 + 0.x6
26
Subject to
3x1 + x2x3 = 3
4x1 + 3x2 – x4 + x5 = 6
x1 + 2x2 + x6 = 3
xj = 0, j = 1,2,.....6
x1 = x2 = 0, x4 = 0, x3 = 3, x5 = 6, x6 = 2
Simplex table is
4 1 M 0 M 0 
CB XB xB1/y11 y11
x1 x2 x3 x4 x5 x6
>0
M x3 = 3 3 1 1 0 1 0 33=1
M x5 = 6 4 3 0 –1 1 0 3  4 = 1.5
0 x6 = 3 1 2 0 0 0 1 3  1 =3
zj - cj 7M–
4M–1 0 –M 0 0
4
7M–4x3 is the most positive net evaluation. Hence x1 enters as a basic variable in the
place of x3.
4 1 M 0 M 0 
CB XB xB1/y11 y11
x1 x2 x3 x4 x5 x6
>0
M x1 = 1 1 0 0 0
M x5 = 2 0 5/3 –4/3 –1 1 0
0 x6 = 2 0 5/3 1/3 0 0 1
zj - cj 5/3
–
0 M 0 0
M
+1/3
Introduce x2 as a basic variable in the place of x5 5/3 is the pivotal element.
27
4 1 M 0 M 0 
x4 CB XB x xB1/y12 y12 enters
x2 x3 x4 x5 x6
as a 1 >0 basic
4 x1 = 3/5 1 0 3/5 1/5 –1/5 0
1 0 1 0 –
0 x6 = 0 0 0 1 1 –1 1
zj - cj
0 0 1/5 –M– 0
variable in the place of x6 1 is the pivotal element.
4 1 M 0 M 0
4 x1 = 3/5 1 0 0 0
1 x6 = 6/5 0 1 –1/5 0 0 3/5

0 x4 = 0 0 0 1 1 –1 1
zj - cj
0 0 0 –M

Since all zj - cj  0, current solution is optimum. The optimum solution is x1 = 3/5, x2
= 6.
Maximum z = 4(3/5) + 1(6/5) = 18/5
Two–Phase Simplex Method
This method is an alternate method to solve a given L.P.P. in which some artificial
variables are involved i.e. when the constraints are of the type  or = This method involves
two phases.
28
Phase I: Formulate a new problem by replacing the original objective function by the
sum of the artificial variables. The new objective function is then minimized subject to the
constraints of the original problem has a feasible solution, the minimum value of the new
objective function will be zero, which indicates that all the artificial variables are zero. Now
go to phase II. Otherwise, if the minimum value is > 0, the problem has no feasible solution.
Phase II: Use the optimum basic solutions of phase I as the starting solution for the
original problem. In this case the original objective function must be expressed in terms of
the non-basic variables, by elimination technique.
Example
Solve the following L.P.P.
Subject to 3x1 + x2 = 3
4x1 + 2x2  6
x1 + 2x2  3,
x1, x2  0
by two-phase method
Solution
The standard form of the L.P.P. is
Subject to 3x1 + x2 + x3 = 3
4x1 + 3x2 –x4 + x5 = 6
x1 + 2x2 + x6 = 3,
xj = 0, j = 1,2, ..... 6
Here x3 and x4 are artificial variables
Phase I: We have to now minimum u = x3 + x5(sum of the artificial variables)
Subject to the above conditions
A basic feasible solution is x1 = 0, x2 = 0, x4 = 0, x3 = 3, x5 = 6, x6 = 3
The starting simplex table is

0 0 1 0 1 0 
CB XB xB1/y11 y11
x1 x2 x3 x4 x5 x6
>0
1 x3 = 3 3 1 1 0 0 0 33=1
1 x5 = 6 4 3 0 –1 1 0 6  4 = 1.5
0 x6 = 3 1 2 0 0 0 1 31=3
zj - cj 7 4 0 –1 0 0
7 is the most positive net evaluation and hence x1 enters as a basic variable in the
place of x3.
3 is pivotal element
The row operations to get the new table are
29
The new table is
0 0 1 0 1 0 
 CB XB xB1/y12 y12
x1 x2 x3 x4 x5 x6
x2 >0 enters
as a basic
0 x1 = 1 1 0 0 0 3
1 x5 = 2 0 5/3 –4/3 –1 1 0
0 x6 = 2 0 5/3 –1/3 0 0 1
zj - cj 0 5/3 –7/3 –1 0 0
variable in the place of x5 5/3 is the pivotal element.
The new table is
0 0 1 0 1 0
0 x1 = 3/5 1 0 + 0
0 x5 = 6/5 0 1 0
0 x6 = 0 0 0 1 1 –1 1
zj - cj 0 0 –1 0 –1 0

Since all zj - cj  0, current solution is optimum for u. The optimum solution is x 1 = 0,
x5 = 0, i.e. the artificial variables are zero.
Phase II: Now we solve the original problem
Starting with basic solution x1 = 3/5, x2 = 6/5, x6 = 0
We can drop the columns corresponding to x3 and x5 as they are artificial variables,
which are not present in the basic feasible solution.
30
The new starting solution is
0 0 1 0
CB XB xB1/y11 y11
x1 x2 x3 x4
>0
4 x1 = 3/5 1 0 0
1 0 1 0 -
0 x6 = 0 0 0 1 1  
zj - cj
0 0 0
x4 enters as a basic variable in the place of x6 is the pivotal element.
The new table is
0 0 1 0
CB XB x1 x2 x3 x4
4 x1 = 3/5 1 0 0
1 0 1 0
0 x4 = 0 0 0 1 1
zj - cj
0 0 0
For the minimum problem, since all zj - cj  0, current solution is optimum.

The optimum solution is
x1 = 3/5, x2 = 6/5, and min
Example 2
Solve the L.P.P.
Minimize z = 3x1 + x2 – x3
31
x1 + 2x2 + 3x2  11
–4x1 + x2 + 2x3  3
2x1 + x3 = –1
x1, x2 x3  0
The standard form is
Minimize z = 3x1 – x2 – x3
Subject to
x1 + 2x2 + x3 – x4 = 11
–4x1 + x2 + 2x3  3
2x1 + x3 + x7 = 1
xj  0, j = 1,2,.....7
In the given third constraint the R.H.S. is negative. It should be always positive. First
multiply by a –tive sign and then add artificial variable x7
Phase I: Minimize u = x6 + x7 subject to the above conditions

The basic feasible solution is x1 = 0, x2 = 0, x3 = 0, x5 = 0, x4 = 11, x6 = 3, x7 = 1.
The initial simplex table is
0 0 0 0 0 1 1 
CB XB x1 x2 x3 x4 x5 x6 x7 xB1/y13 y13 > 0
0 x4 = 11 1 –2 1 1 0 0 0 11  1 = 11
1 x6 = 3 –4 1 2 0 –1 1 0 3  2 = 1.5
1 x7 = 1 –2 0 1 0 0 0 1 11=1
zj - cj –6 1 3 0 –1 0 0
For the maximization problem, 3 is the most positive net evaluation. Hence x4 enters
as a basic variable in the place of x7 1 is pivotal element.
The row operations to get the new table are
0 0 0 0 0 1 1 
0 x4 = 10 3 2 0 1 0 0 –1
1 x6 = 1 –2 1 0 0 –2 1 –2 1
0 x3 = 1 –2 0 1 0 0 0 1 –
zj - cj –2 1 0 0 –2 0 –3
enters as a basic variable in the place of x6
The row operations to get the new table are;
32
The new table is
0 0 0 0 0 1 1
CB XB x1 x2 x3 x4 x5 x6 x7
0 x4 = 12 –1 0 0 1 –4 3 –5
0 x4 = 1 –2 1 0 0 –2 1 –2
0 x3 = 1 –2 0 1 0 0 0 1
zj - cj 0 0 0 0 0 –1 –1
Since all zj - cj  0, current solution is optimum. Minimum u = 0. The values of the

artificial variables x6 and x7 are equal to zero. We can now go to Phase II.
Phase II: Now we solve the original problem
Minimize z = 3x1 – x2 – x3 taking
x4 = 12, x2 = 1, x3 = 1 as a basic solution. We can drop the columns corresponding to
x6 and x7.
Since all CB XB x1 x2 x3 x4 x5 zj - cj  0, for
the 0 x4 = 12 –1 0 0 1 –4 maximization
problem, the –1 x2 = 1 –2 1 0 0 –2 current solution
is optimum. –1 x3 = 1 –2 0 1 0 0
x1 = 0, zj - cj 1 0 0 0 2 x2 = 1, x3 = 1 is
the optimum solution
Maximum z = 3(0) – 1(1) – 1(1) = –2
Dual Simplex Method

This is another method to solve a L.P.P. having constraints of the hope . The
algorithm for this method is given below:
Step 1: Convert all the constraints into the type  and introduce slack variables in the
constraints and obtain an initial basic solution.
Step 2: Find the net evaluation zj - cj in the starting table
a) For a maximization problem, if all zj – cj  0 and the basic variables are
also > 0, then an optimum basic feasible solution has been obtained.
b) If all z, – c .  0 and atleast one basic variable is negative then to go Step 3.
c) If atleast one, z, – c . < 0, the method is not applicable to the given problem.
Step 3: Select the most negative of the basic variables. Let be the most negative
basic variables. Then leaves as a basic variables.

33
Step 4: a) If all the elements ykj, j = 1, 2, ...... n  0 there is no feasible solution to the
problem.
b) If atleast one ykj is negative, compute the ratio ykj, j = 1, 2, ...... n and choose
the maximum of these. Then the corresponding variables, say enters
as basic variable.
Step 5: Test the new iterated dual simplex table for optimality. Repeat the procedure
until the optimum feasible solution is obtained.
Note: For minimizing problems, in step 2 the net evaluations must be  0. In step 4.
Atleast one ykj must be negative. If is minimum, than xr will become a basic variable.
Example 1
Using dual simplex method solve the L.P.P.
Minimize z = 2x1 + x2 Subject to constraints
3x1 + x2  3
4x1 + 3x2  6
x1 + 2x2  3,
x1 – x2  0
Solution
Convert all the constraints to  type
–3x1 – x2  –3
–4x1 – 3x2  –6
x1 – 2x2  3
Adding surplus variables, the L.P.P. becomes
Subject to –3x1 – x2 + x3 = –3
–4x1 – 3x2 + x4 = –6
x1 + 2x2 + x5 = 3
A basic solution is x1 = 0, x2 = 0, x3 = 3, x4 = 6, x5 = 3. Note that this is not a feasible
solution. We cannot apply simplex method, but we can apply dual simplex method. The
starting dual simplex table is
2 1 0 0 0
CB XB x1 x2 x3 x4 x5
0 x3 = –3 –3 –1 1 0 0
0 x4 = –6 –4 –3 0 1 0
0 x5 = 3 1 2 0 0 1
zj - cj –2 –1 0 0 0
34

Since all zj - cj  0, for the maximization problem, the current solution is optimum,
but not feasible. We can apply dual simplex method since two basic variables x3 = –3, x4 = –
6, are negative x4 = –6 is the most negative variable. It will be dropped as a basic variable.
To find the new basic variable evaluate the ratios
The ratio are The minimum of these ratios is and that corresponds
to variable x2. Hence x2 becomes a new variable in the place of –3 is pivotal value.
The row operations to get the new dual simplex table are
The new dual simplex table is
2 1 0 0 0
0 x3 = –1 –5/3 0 1 –1/3 0 
1 x2 = –2 4/3 1 0 –1/3 0
0 x5 = –1 –5/3 0 0 2/3 1
zj - cj
0 0 0

Since all zj - cj  0, for the minimization problem, the current solution is optimum, but
not feasible. Both x3 and x4 are equal to –1. So we can select any one of them to leave the
basic select x3 = –1 leave the basis.
The corresponds to variable x1. Hence x1 enters as a basic variable in the place of x3
is the pivotal element.

35
The new dual simplex table is
2 1 0 0 0
2 x1 = 3/5 1 0 –3/5 1/5 0
1 x2 = 6/5 0 1 4/5 –3/5 0
0 x5 = 0 0 0 –1 1 1
zj - cj 0 0 –2/5 –1/5 0
Since all zj - cj  0, current solution is optimum and also feasible. The solution is
and minimum
36
UNIT III
TRANSPORTATION PROBLEM
Transportation Problem is a special case L.P.P. and occurs very frequently in practical
life.
Let non-negative quantities a1, a2 ... an of a product be available at m places called
xij + 0, i  1, 2 ,,,,,, m, j = 1,2......m

The data can be given in the form of a table, called transportation table. In this table
there are mn squares, called cells. The per unit cij of transporting from the ith origin to the jth
designation is in the lower right position of the (i, j)th cell. Any feasible solution to the
transportation problem is displayed in the table by entering the value of x ij in the small square
at the upper left position of the (i, j)th cell. The various origin capacities and destination
requirements are listed in the right most columns and bottom respectively. These are called
rim requirements. If the total availability is equal to the total requirement, i.e., if
the problem is called a balanced transportation problem, otherwise it is
called unbalanced. The above is necessary and sufficient condition for the existence of a
feasible to the transportation problem.
Transportation Table
Destinations
D1 D2 DJ Dn Availability
x11 x12 x1i x1n
O1 c13 a1
c11 c13 cij cin
x21 x22 x2i x2n
O2 c13 a2
c21 c22 c2j c2n
Origins
x11 x12 x1i x1n

Oi c13 ai
c11 c12 cij cin
xm1 xm2 xmi xmn

Om am
c13
37
cm1 cm2 cmj cmn

Requirement b1 b2 bJ bn
A set of non-negative variables xij > 0 which satisfies the rim requirements is called a
feasible solution.
For a mn transportation problem, if the number of a feasible is not more then (m+n–
1), then the feasible solution is called basic feasible solution. If the number of feasible
solution is exactly (m+n–1), then it is called non-degenerate basic feasible solution.
Methods to find basic feasible solution (B.F.S) to a transportation problem(T.P)
North West corner rule
Step 1: Start with the North West corner of the transportation table. This cell is (1, 1)
has a1 and b1 as availability and requirement.
If a1 > b1, then assign x11 = min (a1, b1) in the cell (1,1). The requirement at D1 is
satisfied but the capacity of O1 completely exhausted. Score out the first column. Move to
right horizontally to the Second Column and make the second allocation of magnitude x 12 =
min (a1 –x11, –b2) in the cell (1,2).
If a1 > b1, assign x11 = a1 in the cell (1,1). The availability satisfied. Move vertically
below to the nest row to the cell (2,1) until the demand of D1 is satisfied and allocation x21 =
min (a2 –b2–x11,) Score out the first row.
If a1 = b1, the capacity of O1 is completely exhausted as well as the requirements at D1
is completely satisfied. There is a tie for second allocation. Allocation can be made in the
horizontal or vertical cell. make the second allocation of magnitude x12 = min {a1, a13 b2}= 0
in the cell (1,2) or x21 = min {a2, b1 – b1}= 0. Score out either the first column or row.
Step 2
Start from the new north – west corner of the transportation table, satisfying
destination requirements and exhausting the origin capacities, one at a time move down
towards the lower right corner of the transportation table until all the rim requirements are
satisfied.
Example 1.
Determine an initial basic feasible solution to the following T.P. using North West
corner rule.
I II III IV V Availability
A 3 4 8 8 9 20
From B 2 10 1 5 8 30
C 7 11 20 40 3 15
D 2 1 9 14 16 13
Requirement 40 6 8 18 6
Solution
Total availability = Total requirement = 78
This is a balanced transportation problem
Feasible solution exists
38
Assign x11 = min [20, 40] = 20 in the cell (1, 1). Availability at A is exhausted. Score
out the first row.
I II III IV V
20 20
A
3 4 6 8 9
B 2 10 1 5 8 30
C 7 11 20 40 3 15
D 2 1 9 14 16 13
40 20 6 8 18 6
Next allocate x21 = min [30, 20] = 20 in the cell (2, 1). Requirements of I are
satisfied. Score out the first column.
I II III IV V
20 30 - 10
B
2 10 1 5 8
C 15
7 11 20 40 3
D 13
2 9 14 16
20 6 8 18 6
Next allocate x22 = min [10, 6] = 6 in the cell (2, 2). Requirements of II are satisfied.
Score out the second column.
II III IV V
6 10 - 4
B
10 1 5 5
C 15
11 20 40 3
D 13
1 9 14 16
8 18 6
Next allocate x23 = Min [4, 8] = 4 in the cell (2, 3). Availability at B is exhausted.
Score out the second column.
III IV IV
4
B 4
1 5 8
39
C 15
20 40 3
D 13
9 14 16
4 18 6
Next allocate x33 = Min [4, 15] = 20 in the cell (3, 3). Requirements of III are
satisfied. Score out the third column.
IV V
11
C 11
40 3
D 14 16 13
7 6
Next allocate x34 = Min [11, 18] = 11 in the cell (3, 4). Availability at C is exhausted.
Score out the first column.
IV V
11 6
D 13
14 16
7 6
Finally allocate x44 = 7, x45 = 6
Hence a basic solution to the T.P. is
x11 = 20, x22 = 6, x23 = 4, x33 = 4, x34 = 11, x44 = 7, x45 = 6,
Number of allocation is 8, which is equal to m + n – 1 = 4 + 5 – 1 = 8
Hence the solution the solution the solution is a non-degenerate basic feasible solution.
II III IV V
20
A
3 4 6 8 9 20
20 6 4
B
2 10 1 5 8 30
4 11
C
7 11 20 40 3 15
7 6
D
2 1 9 14 16 13
8 18 6
40
The cells which get allocation are called cells. Total transportation cost according to
this allocation = 20(3) + 20(2) + 6(10) + 4(20) + 11940+7(14) + 6(16) = Rs. 878.
Example 2:
Find an initial basic feasible solution to the following T.P. by North West corner
method.
A B C D Availability
I 1 5 3 3 34
Origin II 3 3 1 2 15
III 2 2 2 3 12
IV 2 7 7 4 19
Requirement 21 25 17 17
Solution
Total availability = Total requirement = 80
Allocate x1 = min [34, 21] = 21 in the cell (1, 1). Requirement of A is satisfied.
A B C D
21
I 34 13
1 4 6 8
B 2 3 1 5 15
C 7 2 2 3 12
D 2 7 2 4 19
21 25 17 17
Allocate x12 = min [13, 25] = 13 in the cell (1, 2). Availability of I is exhausted.
A B C
13
I 13
5 3 5
II 3 1 2 15
III 2 2 3 12
IV 7 2 4 19
20 6 8
Next allocate x22 = Min [15, 12] in the cell (2, 2). Requirements of B is satisfied.
Score out the 2nd column.
B C D
12
II 15 3
3 1 2
C 2 3
41
3 12
D
7 2 4 19
12 17 17
Next allocate x23 = Min [3, 17] = 3 in the cell (2, 3). Availability of II is exhausted.
Score out 2nd row.
C D
3
II 3
2 2
III 2 3 12
IV 2 4 19
17 14 17
Next allocate x33 = Min [12, 14] = 12 in the cell (3, 3). Availability of III is exhausted.
Score out 3rd row.
C D
2 17
IV
2 2 19
2 17
Next allocate x43 = 2, x44 = 17
The basic solution is
x11 = 20, x21 = 20, x22 = 6, x23 = 4, x33 = 4, x34 = 11, x44 = 7, x45 = 6,
Since there are m + n – 1 = 4 + 5 – 1 = 4 + 4 – 1 + 7 basic cells, this is a given in the
following transpiration table.
A B C D
21 21
I
1 5 3 3 34
12 3
II
3 3 1 2 15
12
III
0 2 2 3 12
2 17
IV
2 7 2 4 19
8 18 6
42
According to this allocation the total transportation cost.

= 21(1) + (13) (5) + (12) (3) (1) + 4(12) (2) + (2) + (17)
2. Vogel’s Approximation Method (VAM)

Step 1: For each row of transportation table, identify the smallest and next to smallest
costs. Determine the difference between them each row. Display them alongside the
transportation table by enclosing them in parenthesis against the respective rows. Similarly,
compute the difference for each column.
Step 2: Identify the row or column with the largest difference among all the rows and
columns. Let the greatest cost in the ith row. Allocate xij = min (ai, bj) in the cell (i, j) and
cancel the ith row or jth column.
Step 3: Repeat step 1 and step 2 for the reduced transportation table, until all the rim
requirements are satisfied.
Example 1:
Find on initial feasible solution to the following T.P by Vogel‟s approximation
method.
I II III IV V Availability
A 3 4 6 8 9 20
From B 2 10 1 5 8 30
C 7 11 20 40 3 15
D 2 1 9 14 16 13
Requirement 40 6 8 18 6
Solution:
In the first row minimum and next minimum cost are 3 and 4. Their difference is 1.
In the 2nd row the difference between minimum and next minimum is 2 – 1 = 1. In the 3rd
row the difference is 7 – 3 = 4. For the 4th row it is 1. Repeating this for columns, the
difference respective differences are 1, 3, 5, 3, 5. Denoting them in the transportation table
we get.
I II III IV V
A 3 4 6 8 9 20 (1)
B 2 10 1 5 8 22 (1)
C 7 11 20 40 3 15 (4)
D 2 1 9 14 16 13 (1)
40 6 8 18 6
(1) (3) (5) (3) (5)
The maximum of all the difference is 5. That corresponds to the 3rd and 5th columns.
1 is the minimum transportation cost among all the elements of these two columns. First
allocation is made to the cell (2, 3) corresponding to this cost. Allocate x23 = Min [30, 8] = 8.
Cancel the 3rd column. The reduced transportation table and the respective difference of the
least and next least cost are given below.
43
I II IV V
A 3 4 8 9 20 (1)
B 2 10 5 8 22 (1)
C 7 11 40 3 15 (4)
D 2 1 14 16 13 (1)
40 6 18 6
(1) (3) (3) (5)

5 is the maximum difference and the least cost in the corresponding column is 3.
Allocate x35 = Min [15, 6] = 6. Cancel the 5th column and reduce 6 from 15.
The reduce table and the respective difference are:
I II IV
A 3 4 8 20 (1)
B 2 10 5 22 (1)
C 7 11 40 15 (4)
D 2 1 14 13 (1)
40 6 18
(1) (3) (3)
4 is the maximum difference and 7 is the minimum cost in the 3rd row. Allocate x31 =
Min [9, 40] = 9 cancel the 3rd row and reduce 9 from 40.
I II IV
A 3 4 8 20 (1)
B 2 10 5 22 (1)
D 2 1 14 13 (1)
31 6 18
(1) (3) (3)
3 is the maximum difference and I is the minimum cost in the 2nd column. Allocate
x42 = Min [13, 6] = 6 and cancel the 2nd column. Reduce 6 from 13.
The reduced table and the respective difference are:
I IV
A 3 8 20 (5)
B 2 5 22 (3)
D 2 14 13 (12)
31 18
(1) (3)
Allocate x41 = Min [7, 31] = 7 and cancel the row corresponding to D.
We now get
I IV
A 3 8 20 (5)
44
B 2 5 22 (3)
31 18
(1) (3)
Allocate x12 = Min [20, 24] = 20 and cancel the row corresponding to A.
I IV
B 2 5 22
4 18
Allocate x21 = 4, x24 = 18

The basic feasible solution is
II III IV V
20
A
3 4 6 8 9 20
4 8 18
B
2 10 1 5 8 30
9 6
C
7 11 20 40 3 15
7 6
D
2 1 9 14 16 13
40 6 8 18 6
Transportation cost according to this allocation

= (20) (3) + (4) (2) + (8) (1) + (18) (5) + (9) (7) + (6) (3) + (7) (2) + (6) (1)
= Rs. 267
Note: Of all the methods Vogel‟s methods is the best.
ASSIGNMENT PROBLEM
This is a special case of the transportation problem in which the objective is to assign
a number of origins to the equal number of destinations at a minimum cost or a maximum
profit.
Suppose there are n jobs in a factory and the factory has n machines to process the
jobs. Let cij be the cost of assignment when the job i (i = 1.2...n) is processed by the machine
j (j = 1,2...n). The assignment is to be made in such a way that each job can associate with
one and only one machined. Let xij be the variable defined by
45
Hence the assignment problem is to find xij > 0 so as to minimize

subject to the constraints.
This is nothing but a special case of transportation problem.

We are not, going to solve an assignment problem by MODI method. Instead this
problem can be solved by another elegant method called Hungarian method.
HUNGAIRAN METHOD TO SOLVE ASSIGNMENT PROBLEM

Step 1: If the problem is a minimization problem, subtract the minimum cost of each
row of the cost matrix, from all the elements of the respective row. Then modify the
resulting matrix by subtracting the minimum cost of each column from all the elements of the
respective column, to get the starting matrix.
Step 2: Draw the least possible number of horizontal and vertical lines to cover all the
zeros of the starting table. Let the number of these lines be m.
If m = n. the order of the cost matrix, then the optimum assignment can be made. Go
to step 5. If m < n, go to step 3.
Step 3: Determine the smallest cost in the starting table, not covered by the m lines.
Subtract this cost from all the surving (uncovered by lines) elements of the starting matrix
and add the same to all those elements of the starting matrix which are lying at the
intersection of horizontal and vertical lines, thus obtaining the second modified cost matrix.
Step 4: Repeat the steps 1, 2, and 3 until we get m = n.
Step 5: Mark the zeros of the optimum cost matrix In a separate table. Examine the
rows successfully until a row with exactly one zero is found- Enclose this zero inside a circle
and an assignment will be made there. Make a cross (x) in the cells of all other zeros lying in
46
this column of the encircled zero, to show that they cannot be considered for future
assignment. Continue in this manner until all the rows are taken care of.
Step 6: Examine the columns successfully until a column with exactly one unmarked
zero is found. Encircle that zero. An assignment will be made there. Make a cross in the
cells of all other zeros lying in the row of encircled zero. Continue in this way until all the
columns have been taken care of.
Step 7: Repeat steps 5 and 6 successfully until one of the following arises.
i) no unmarked zero is left
ii) there lies more than one unmarked zero in one column or row.
In case (i). The algorithm steps.
In case (ii) encircle one of the unmarked zeros arbitrarily and mark a cross in the cell
remaining zeros in its row and column. Repeat the process until no unmarked zero is left.
Step 8: We now have exactly one encircled zero in each row and each column of the
cost matrix. The assignment schedule corresponding to these zeroes is the optimal
assignment.
Note: In case of maximization problem, make all the elements of the cost matrix
negative and convert it into minimization problem.
Example 1:
Consider for problem of assigning five jobs to five persons. The assignment costs are
given below:
Job
1 2 3 4 5
A 8 4 2 6 1
B 0 9 5 5 4
Persons C 3 8 9 2 6
D 4 3 1 0 3
E 9 5 8 9 5
Determine the optimum assignment schedule.
Solution: From each row of the cost matrix the minimum cost and subtract that from
all the elements of the respective row.
We get
7 3 1 5 0
0 9 5 5 4
1 7 8 0 4
4 3 1 0 4
4 0 3 4 0
Now, subtract the minimum cost of each column from all the elements of that column.
We get
--7 – 3 -- 0 – 5 – 0--
0 9 4 5 4
47
1 7 4 0 4
--4 – 3 – 0 – 0 – 3--
--4 – 0 – 2 – 4 – 0--
Now we draw the minimum possible number of horizontal and vertical lines so as to
cover all the zeros.
--7 – 3 -- 0 – 5 – 0--
0 9 4 5 4
1 7 4 0 4
--4 – 3 – 0 – 0 – 3--
--4 – 0 – 2 – 4 – 0--
Since the minimum of lines used 5 = order of the cost matrix, an optimum assignment
can be made.
Now consider the zeros of the above matrix.
1 2 3 4 5
A O O
B O
C O
D O
E O O
Now examine the row successively to find out a row with exactly one zero and mark
it with a circle and cancel any other zero in the respective columns with a cross. Again
examine the zeros successively to find out the row with only me unmarked zero.
1 2 3 4 5
A O 
B 
C 
D  O
E  O
48
In this table, each row and each column has exactly one encircled zero. Hence the
optimum assignment is achieved. The optimal assignment is
A --> 5, B --> 1, C --> 4, D --> 3. E--> 2
The minimum assignment cost = 1 + 0 + 2 + 1 + 5 = 9
Example 2:
Solve the following assignment problem to minimize total cost of assignment.
6 5 8 11 16
1 13 16 1 10
16 11 8 8 8
9 14 12 10 16
10 13 11 8 16
Solution select the minimum cost from each row and subtract it from the other
elements of the row we get.
1 0 3 6 11
0 12 15 0 9
8 3 0 0 0
2 5 3 0 8
Since all the columns have zero elements, we will get the same matrix for the column
operation. Therefore, the above is the starting matrix. Now we cover the zeros by horizontal
and vertical lines.
--1 – 0 – 3 – 6 – 11--
0 12 15 0 9
--8 – 3 – 0 0 – 0--
0 5 – 3 – 1 – 7
2 5 – 3 – 0 – 8
All the zero are covered by 4 lines which is less then 5, the order of the cost matrix.
Optimal assignment cannot be made. Now modify this matrix by subtracting the least
element among the surviving elements, and adding the same at the inter section of horizontal
and vertical lines.
3 is the least element. The modified matrix is
4 0 3 9 11
0 9 13 0 6
11 3 0 3 0
0 2 0 1 4
49
2 2 0 0 5
Now we cover the zeros by horizontal and vertical lines.
--4 – 0 – 3 – 9 – 11--
0 9 13 0 6
--11 – 3 – 0 3 – 0--
0 2 – 0 – 1 – 4
2 2 – 0 – 0 – 5
Now we have lines covering all the zeros. Hence optimum assignment can be made.
The table of zero is written now.
I II III IV V
A 
B O O
C O O
D O O
E O O
In the cell (1, 2) we have a single zero. Enclose it by a circle. All the other rows have
two zeros. Now check the columns. Column 5 has a single zero. Enclose it and put a cross
for the cell (3, 3).
I II III IV V
A 
B O O 
C O
D O O
E O O
Now each row or column has two zeros. We have encircle a zero arbitrarily. In such
cases the problem has more then one solution. Encircle the cell (2, 4) and score out cell (2, 1)
and (5, 4).
II III IV V
A 
B O 
C O 
D O O
E O O
50
Now 5th row has a zero. Encircle and score out cell (4, 3) cell (4, 1) is the only zero
left. Encircle it.
II III IV V
A 
B O 
C O 
D  O
E  O
Now only one zero has a circle around it in every row or column.
 optimum is achieved. The optimal assignment is
A  III, B  IV, C  V, D  I, E  III
Minimum cost assignment is = 5 + 1 + 8 + 9 + 11 = 34.
51
UNIT IV
GAME THEORY
Meaning, Definition and Characteristics of Game Meaning

The name „Game Theory‟ seems interesting as it given an impressionto layman to
mean a body of knowledge relating to the conduct of the various games played by the people.
Although, to some extent this sense holds good, but in reality it does not relate to the conduct
of any game or play. Lexically, speaking, „Game‟ means some play or gambling between
two or more individuals either with or without some pay offs (i.e., settlement through
payments). But broadly speaking, the term „game‟ refers to any conflicting situation in which
two or more individuals or organizations compete with each other to gain some thing at the
cost of the other or others. The examples of such a situation are competition between the
business units, military battles, election contests between some candidates, gambling, horse-
race, cross word puzzles etc.
Definition
The term game theory may be defined as a conflicting situation between two or more
independent competing entities individuals or organizations) who carry on certain activities
according to a set of rules with a view to gaining some benefits or satisfaction at the cost of
the opponents.
Characteristics
From the above definition, the essential characteristics of a game may be brought out
as under:
1. There must be some interest of conflicting nature that calls for competition between
some persons.
2. There must be some competitors who may be individuals or any organization and
whose number is finite and countable.
3. There must be several alternative courses of action available to each of the
competitors.
4. There must be a set of rules for taking up a course of action and they must be known
to each of the competing participants.
5. The course of actions must be taken independently and simultaneously by each of the
competitors before knowing the line of action taken by his competitors.
6. There must be some outcome of all the different courses of action taken up by the
competitors and all such outcomes must be qualitatively defined and made known to
each of the competitors much in advance of the game.
7. The outcome of the game must be affected by the course of action taken up the
different participants and they must result in a set of payments by one to the other.
8. There is an assumption of rationality that each competitor will act so as to maximize
his minimum gain or minimize his maximum loss.
Assumptions of Game Theory
As pointed out above, the game theory works or certain assumptions which are as
under:
52
1. It is assumed that „as if each competitor behaves rationally and intelligently to

optimize his objective in the game. As such, this theory is also called as the “as if
theory” which implies that as if rational decision makers behave in some well defined
way such as maximizing the minimum gain and minimizing the maximum loss.
2. It is assumed that each participant in the game has available to him a finite set of
possible course of actions.
3. It is assumed that each player is in know of all the information relating to the game he
plays.
4. It is assumed that each player attempts to maximize the gains and minimize the losses
for himself and minimize the gains and maximize the losses for his opponents.
5. It is assumed that each competitor takes his course of action simultaneously with that
of his opponents.
6. It is assumed that each player takes his decision individually in a skilled manner
without any direct communication with his opponents.
7. It is assumed that the pay offs (payments to be made in settlement of a game) are
already fixed in relation, to the different courses of actions of the players i.e., the
participants.
Definition of the Terms used in the Game Theory
For a clear understanding of the game theory it is necessary to have full knowledge of
the various terms technically used in such theory. These terms are defined here as under.
(i) Game. It refers to any field of conflicting situation to which the competitors are
attracted to optimize their objective in terms of profit, benefit, expenses or losses etc.
at the cost of their opponents.
(ii) Play. It refers to a course of action taken up by a competitor in a game.
(iii) Player. It refers to a competitor or participant who takes active part in a game.
(iv) Pay-off. It means quantitative measure of satisfaction which a player gains or losses
at the end of play. This is usually expressed in the form of rupees or numbers.
(v) Strategy. It refers to the skill or predetermined rules by which a competitor selects
his/her course of action from a list of course of actions, during a game.
(vi) Pure Strategy. It refers to a decision taken in advance of all the plays to choose a
particular course of action. It is usually represented by a number with which a course
of action is associated.
(vii) Mixed Strategy. It refers to a decision taken in advance of all plays to choose of
action for each play in accordance with some fixed proportions.
(viii) Pay-off matrix. It refers to a table of pay-offs or outcomes of different strategies of a
game. The following is the specimen of a pay off matrix:
Pay-off matrix
Minimizing player Y
Y1 Y2 Y3 Y4
X1 10 15 20 25
Maximizing Player X X2 5 6 0 8
X3 7 9 1 3
53
(ix) Optimal Strategy. It refers to the best possible strategy or a plan of action adopted by
a competitor that gives him the maximum possible benefit. In terms of the pay off
matrix beyond, it also refers to the most possible reduced from of the pay off matrix
beyond which the strategies can not be altered.
(x) Maximin. It refers to the maximum of the minimum values of a row. It is usually
indicated by a box .
(xi) Minimax. It refers to the minimum of the maximum values of a column. It is usually
indicated by a circle O.
(xii) Value of game. It means the sum of the pay-offs expected by the players at this level
of their optimum strategy. In a game with saddle point(s), the pay off at the saddle
point o is called the value of the game and it is obviously equal to the maximum and
minimax of the game. In other cases, it is the total of the pay-offs expected by the
players at the level of their optimum strategies. The value of a game for a winning
player is positive whereas for a losing player it is negative.
(xiii) Saddle Point. It refers to the cell(s) of a pay-off matrix that contain(s) the element
equal to the maximum of the minimum values of the rows and the minimum of the
maximum values of the columns. For convenience sake, it is indicated by the mark o
. In a two persons zero-sum game the element at the saddle point is the value of the
game. Thus, there exists a saddle point in a pay-off matrix, where the maximum is
equal to the minimax. There can be more than one saddle point in a particular game.
A saddle point is also, otherwise called as the equilibrium point of a pay-off matrix.
Rules of determining a saddle point

1. Select the minimum element of each row of a pay-off matrix and put them, in a
column to the right. Also put a circle O over all such minimum elements selected
including the ones that appear in the pay-off matrix.
2. Select the maximum element of each column of the pay off matrix and put them in a
row at the bottom of the table. Also put a box over all such maximum elements
selected including the ones that appear in the pay-off matrix.
3. Select the cell(s) that contain(s) the element(s) marked with both a circle and a box o
. The cells, thus selected constitutes the saddle point(s) and the element of such
cell(s) is the value of the game.
Illustration 1: Determine the saddle-point(s) of the following matrix:
Pay-off Matrix
B
B1 B2 B3 B4
A
A1 5 15 7 8
A2 13 12 10 20
A3 25 30 15 40
A4 60 50 10 8
Solution: Following the rules of saddle point the given matrix is marked as under
54
Marked-up Matrix
From the above marked up matrix we find that the element 15 in the cell (3, 3) is
covered by both the marks o and is equal to both the maximum and minimax values shown
in the row minimum column and column maximum row of the matrix. Hence, this cell, (3, 3)
constitutes the saddle point and the value of the game is 15. This indicates that the player A
will gain Rs. 15 if he takes up the A3 course of action. This also indicates that the player B
will loose the same Rs. 15 if he adopts the B3 course of action when A adopts the A3 course
of action.
Illustration (on multi saddle points). Find the saddle point(s) from the following pay
off matrix.
Minimizing Player Y
Y
Y1 Y2 Y3
X
Maximizing Player X
X1 –2 15 –2
X2 –5 –6 –4
X3 –5 20 –8
Solution: Applying the saddle point rules the given matrix is marked as under:
55
From the above marked matrix we find that there are two cells (1, 1) and (1, 3) that
contain the element – 2 covered by both the mark o which is equal to both the maximum
and minimax indicated therein. This shows that there are two saddle points in the given game
viz., cell (1,1) and cell (1, 3), where
(i) the best strategy for the player X is X1
(ii) the best strategy for the player Y is both Y1 and Y3.
(iii) the value of the game is – 2 for X and +2 for Y.
Game of Dominance
A game of dominance is one to which the rules of dominance can be applied to reduce
the size of its pay off matrix for making it convenient for solution. In the pay off matrix of
such a game each element of one or more rows (columns) is less than or equal to the
corresponding elements of some other rows (columns) or their averages. A row containing
all the bigger and the equal elements, and a column containing all the smaller and the equal
elements is called a dominating row and column respectively. On the other hand, a row
containing all the smaller and equal elements and a column containing all the all the bigger
and equal elements is called a dominated row or column respectively. According to the rule
of dominance all the dominated rows and columns of a pay off matrix are to be deleted to
reduce its size to an easily workable form. The rules of dominance read as follows.
Rules of dominance
1. If the strategy of a player dominates over the other strategy in all conditions then the
latter strategy (dominated row or column) can be ignored as it will not affect the solution
in any way.
2. If all the elements in a row are less than or equal to the corresponding elements in
another row or the average of the corresponding elements of some other rows, then that
row is dominated and hence, is liable to be deleted.
3. If all the elements in a column are greater than or equal to the corresponding elements
in another column or the average of the corresponding elements of some other columns,
then that column is dominated and hence is liable to be deleted.
4. If any row (column) dominates the average of some other rows (columns) then one of
the rows (columns) thus averaged may be deleted from the matrix.
5. Such rules of dominance can be applied to both the problems of pure strategy and
mixed strategy to reduce the size off a pay of matrix to 2  2 as far as possible beyond
which it need not be reduced.
Illustration: Indicate, which of the following pay of matrices represent (i) game of
dominance and (ii) game of non-dominance. Also, reduce the game of dominance to the
possible extent by applying the rules of dominance.
Pay-off Matrix
Player B
(a) B
B1 B2 B3
A
A1 2 4 6
Player A
A2 10 3 8
A3 4 7 3
56
(b) B
B1 B2 B3 B4
A
A1 7 6 8 9
A2 –4 –3 9 10
A3 3 0 4 9
A4 10 5 –2 0
Solution: (i) The pay off matrix given under (a) represents a game of non-dominance
since it does not contain any row (column), each of the elements of which is less than or
equal to the corresponding elements of other row (column) or to the average of corresponding
elements of any other row (column). Hence, such a matrix cannot be reduced by the rules of
dominance.
(ii) The pay off matrix given under (b) represents a game of dominance as it
contains some dominated rows and columns which can be deleted to reduce its size of 2  2
matrix as under:
In the first instance: We observe that each of the elements in the row A3 is less than
or equal to its corresponding element in the row A1. Thus, row A3 is dominated by row A1.
Hence we can delete the row A3 and get the matrix reduced as under:
B
B1 B2 B3 B4
A
A1 7 6 8 9
A2 –4 –3 9 10
A4 10 5 –2 0
In the second instance: We observe that each of the average values of row A2 and
row A4 is less than the corresponding value of the row A1 as under:
B1 B2 B3 B4
A1 7 6 8 9
(A2 +
3 1 3.5 5
A4)/2
Thus, the row A2 and A4 involved in the average are dominated by the row A1.
Hence, we can reduce any of these rows say A2 and get the matrix reduced as under.
B
B1 B2 B3 B4
A
A1 7 6 8 9
A4 10 5 –2 0
In the third instance: Now, we observe that each of the elements in the column B1 is
greater than the corresponding element in the column B2. Thus, B1 is dominated by B2.
Hence, we can delete the columns B1 and get the matrix reduced as under.
57
B
B2 B3 B4
A
A1 6 8 9
A4 5 –2 0
In the 4th Instance: Now, we see that there is no such column, each of the elements
of which is greater than or equal to the corresponding element in another column or the
corresponding average elements of some other column. Thus, the above matrix cannot be
reduced any further. Hence, the possible reduced form of the given matrix (optimal strategy
of the two persons A and B) is given as above.
Note: Now, we can compare the two matrices (original one and the reduced one) with
reference to their maximum value of the minimum of the rows and the minimum value of the
maximum of the columns and see that both of them produce the same result as under:
Original Matrix
B Row
B1 B2 B3 B4
A minimum
A1 7 6 8 9 6  Maximum
A2 –4 –3 9 10 –4
A3 3 0 4 9 0
A4 10 5 –2 0 –2
Column
10 6 9 10
Max.

Minimax
Reduced Matrix
B Row
B2 B3 B4
A minimum
A1 6 8 9 6  Maximum
A4 5 –2 0 –2
Column
6 8 9
Max.

Minimax
Graphic Method
Under this method the given pay off matrix is reduced to 2  2 matrix by means of
graphs and then the same is solved by any other method to arrive at the probable values of the
mixed strategy of the players and the value of the game.
The applicability of this method, however, is subject to the following conditions.
(1) The pay off matrix must be of the order of 2  n or m  2.
58
(2) There must not be a saddle point in the problem.

(3) There must not be the possibility of reducing the matrix any further by the rules of
dominance.
Steps: This method involves the following steps to be taken up in turn.

1. Represent the probabilities of the the player with only two strategies by p1 and p2
respectively the sum of which is equal to one (i.e., p1 + p2 = 1).
2. Represent the two pure strategies I and II of a player by two parallel lines drawn one
unit apart on the Y-axis and mark a natural scale on each of them upto the maximum
value given in the matrix.
3. Represent the probability line by P–P with a unit length drawn on the X–axis,
touching the points of origin of the two parallel lines drawn as above.
4. Put the dots for each of the coordinated values of the different strategies of the other
player, B at the proper points of the two respective parallel lines. This will be put as
under:
Value of e11 on the line 1 and value of e21 on the line II.
Join the coordinated dots thus put by straight lines and represent them as the
respective pay off lines.
5. In case of a 2  n problem, mark the lower boundary of the pay off lines by a thick
line and mark the highest point of intersection on this line by P to indicate the
optimum value of P1 (maximin).
But in case of m  2 problems, mark the upper boundary of the pay off lines by a
thick line and mark the lowerst point of intersection on this line by P to indicate the
optimum value of p1 (minimax).
6. Take up the two lines that pass through the point P to represent the two optimum
strategies for the other player and thereby form a reduced matrix of 2  2 order.
7. Solve the reduced matrix so obtained by any method (preferably arithmetic method)
and obtain the different values of the optimal solution.
Note: In case, there are more than two lines passing through maximum/minimax point
P, select any two lines for forming the reduced matrix whose sum of the vertical oddments is
equal to the sum of the horizontal oddments.
Illustration: Reduce the following game to 2  2 by graphic method and solve the
same to find the relevant optimal results.
B
1 4 –2 –3
A
2 1 4 5
Solution: The pay off matrix thus given is :
B
B1 B2 B3 B4
A1 1 4 –2 –3 –3
A A2 2 1 4 5  Maximin
59
2 4 4 5

Minimax
1. Tests: From the above matrix it is observed
(i) that the game is in 2  n order
(ii) that there is no saddle point in the problem, and
(iii) that the matrix cannot be reduced further by any rule of dominance.
Since all the above conditions are satisfied, we can now proceed with the graphic
method to reduce the game to (2  2) as under:
2. Here, the player A has two strategies viz., A1 and A2. Representing the
probability of his strategies by p1 and p2 respectively, we get the coordinated values of the
various pay off lines as under:
A1 A2
For the pay off line B1 1 2
For the pay off line B2 4 1
For the pay off line B3 –2 4
For the pay off line B4 –3 5
3. Representing the strategy A1 by the scalled line A1, the strategy A2 by the
scalled line A2 the probability line of a unit length by P - P, and drawing the pay off lines
with their coordinated values as above we get the graphic representation of the game as
follows:
Graphic Representation of the Game
4. Analysis: From the above graph it is revealed that the point P indicates the
point of maximum when the two pay off lines B1 and B2 intersect each other. These two lines
represent the two optimum strategies for B. Thus, the 2  n game is reduced to 2  2 game as
under:
60
B
1 2
1 1 4
A 2 2 1
5. Solving the above 2  2 matrix by the arithmetic method, we get,

B
1 2 Oddment Propn.
1 1 4 1 ¼(p1)
A 2 2 1 3 ¾ (p2)
Oddment 3 1 4
Propn. ¾ ¼
(q1) (q2)
6. Interpretation: thus, the various optimal results of the game are as under:
Optimum strategies for A :
Optimum strategies for B :
Value of the game =
Maximum gain for A =
and Maximum loss for B =
Illustration: Reduce the following game to 2  2 by graphic method and find the
various optimum results.
B
–6 7
4 –5
A –1 –2
–2 5
7 –6
Solution: The pay off matrix given is:
B
1 2 Row min.
1 –6 7 –6
2 4 –5 –5
A 3 –1 –2 –2
Maximum
4 –2 5 –2
61
5 7 –6 –6
Col. Max. 7 7
Minimax
1. Test: From the above matrix it is noticed.

(i) that the game is in m  2 order.
(ii) that there is no saddle point in the problem
(iii) that the matrix can not be reduced further by any rule of dominance.
Since, all the above criteria are satisfied we now proceed we now proceed with the
graphic method to reduce the game to (2  2) as under:
2. Here, the player B has two strategies viz., B1 and B2. Representing the
probabilities of his strategies by p1 and p2 respectively we get the coordinated values of the
various pay off lines as follows:
B1 B2
For the pay off line A1 –6 7
For the pay off line A2 4 –5
For the pay off line A3 –1 –2
For the pay off line A4 –2 5
For the pay off line A5 7 –6
3. Representing the strategy B1 by the scalled line B1, the strategy B2 by the
scalled line B2, the probability line of a unit length by P – P and drawing the pay off lines
with their coordinate values as above, we get the graphic representation of the game as under:
62
4. Analysis: From the above graph it is obvious that the point P indicates the
minimax where, the two pay off lines A4 and A5 intereset each other. These two lines
represent the two optimum strategies for A. Thus, the m  2 game is reduced to 2  2 game
as follows:
B
1 2
5. Solving the 4 –2 5 above 2  2 matrix by the
arithmetic method we get, A 5 7 –7
B
1 2 Oddment Propn.
4 –2 5 13 13/20(q1)
A 5 7 6 7 7/20 (q2)
Oddment 11 9 20
Propn. 11/20 9/20
(p1) (p2)
6. Interpretation: Thus, the various optimal results of the game are:
(i) Optimal strategies of A:

63
of B:
(ii) Value of the game :
(iii) Maximum gain for A
Minimum loss for B

64
UNIT V
NETWORK ANALYSIS
(CPM and PERT)
Meaning
Network is a series of related activities which result in some products or services
which in turn contribute to the predetermined goals of an organization. Then, What is
Network Analysis? „Network analysis may be defined as a technique applied in sequencing
problems of a particular project. It is suited to projects of non-routine and non-repetitive
nature‟. The techniques used here, are of two types, viz., CPM (Critical Path Method) and
PERT (Programme Evolution and Review Technique).
Sequencing Problem
Sequencing problems are those which are concerned with minimization of some key
factors of production, such as time and cost.
Project
A project is a work which is proposed to be accomplished within a stipulated time and
a budgeted cost. It may be a Building, a Dam or any other work involving heavy amount of
resources.
Network Diagram
Network diagram is a pictorial presentation of the various events and activities
relating to a particular project. It is drawn purely on the basis of two important things. viz.,
event and event and activity which are explained as under:
(A) Event
An event refers to the starting or ending point of an activity and it requires no time. It
is represented by a figure within a circle, viz: (1), (2), (3), (4) and so on. An event cannot be
achieved until all activities preceding it have been completed. In a network diagram there
will be one starting event and one ending event.
(Example of a Network Diagram)

Types of Event
An event may be classified into various types, viz.: 1. Simple event, (2) Node, (3)
Burst, (4) Head event and (5) Tail event.
These are explained briefly as under:
(1) Simple event. A simple event is one which takes place after or before one
activity takes place.
(2) Node. A node is an event which takes place after the completion of two or
more activities at a time. In other words, it is the ending event of two or more activities.
65
(3) Burst. A Burst is an event immediately after which two or more activities
start. In other words, it is the beginning event of two or more activities.
(4) Head Event. A Head event is an event where an activity ends. Thus, if an
activity, say, „B‟ begins after the event (2), the event (2) will be the tail event.
(5) Tail Event. A Tail event is an event where an activity begins. Thus, if an
activity, say, „B‟ begins after the event (2), the event (2) will be the tail event.
It is to be noted that an activity will have two events, viz., : Head Event and Tail
Event, explained as above.
(B) Activity
An activity refers to some action, the performance of which requires some time. It is
represented by an arrow mark, viz., . Every Activity is named by a capital letter, viz. : A,
B, C, D etc., put above the arrow mark. Further, the time taken by each activity is indicated
just below its arrow mark, viz.: An activity may be of following types.
Types of Activities
(i) Preceding Activity. It is that activity which is completed before an event takes place,
viz.: (3). Here, Activity “C” is a preceding activity of the event (3).
(ii) Succeeding Activity. It is that activity which begins after an event has taken place
viz.: (3) Here, activity which is performed along with another activity of the event (3).
(iii) Concurrent Activity. It is that activity which is performed along with another
activity at the same time, viz.:
Here activity “F” is a concurrent activity of the activity „G‟.

(iv) Dummy Activity. It is that activity the performance of which requires no time, nor
any resources. It is a false activity indicated by the broken lines. Such activities are shown in
a network diagram only to designate a precedence relationship. i.e., to show that the activities
and the events in a network diagram are in the proper order. They simply help in getting a
network diagram complete in an orderly manner, i.e., in sequencing a network diagram.
66
(v) Critical Activity. It is that activity which falls along the critical path. Any delay
caused in the performance of such activity will cause delay in the completion of the project in
toto and hence such activities are to be completed strictly according to the time schedule.
Such activities are also called bottleneck activities.
(vi) Non-critical Activity. It is that activity which falls on a path other than the critical
path. Delay in the performance of such activities does not cause delay in the completion of
the project. Such activities are otherwise called as non-bottleneck activities.
Requirements for a Network Diagram
The drawal of a network diagram and its proper interpretation need the following
information and a clear understanding of the following terms as well:
(A) Information Needed
(i) Sequencing of the activities, i.e., order of the activities in which they are to be
performed.
(ii) Estimate of the times, to be consumed by each of the activities.
(B) Concept of the Terms Related to the Network
(i) Path. A path is a chain of activities that begins with the starting event and ends with
ending event of a particular project. It is indicated through a Network.
(ii) Critical Path. It is that path that runs through a network with the maximum length of
time. In other words, this path indicates the minimum possible time required for the
completion of a project. It does not remain confined to any particular path. With the
readjustment and improvement of the mode of operation of the different activities, any path
may prove to be the critical path.
(iii) Earliest Starting Time (EFT): It is the earliest possible time at which an activity
starts after an event. It is shortly indicated as Est and is put at the beginning point of an
activity above its arrow line. The earliest starting time of the initial activities is zero. But the
earliest starting time of the other activities is the greatest of the Earliest Finishing times of
their respective immediately preceding activities.
(iv) Earliest Finishing Time (EFT) It is the earliest possible time at which at which an
activity finishes. Such time is determined by adding the estimated time of an activity to its
earliest starting time. It is shortly indicated as Eft and is put at the ending point of an activity
above its arrow line. The Eft of a preceding activity ordinarily succeeding activity but in case
of a Node event (where there are two or more activities preceding the event) the greatest of
the Efts of its preceding activities becomes the Est of its succeeding activity.
Note: It is to be noted that the process of indicating the Ests and Efts above each of
the arrow lines runs from the starting point to the ending point of every path of a network and
such process is popularly known as forward-process.
(v) Latest Finishing Time (LFT): It is the latest possible time at which an activity can
be finished without delaying the project beyond its stipulated time. It is shortly indicated as
Lft and is shown at the ending point of an activity below its arrow line. The marking of such
Lfts begins with the ending event of of a network and the process of such marking runs from
the right to the left of the network. Such process is popularly known as Backward process.
The greatest of the Efts of the ending activities becomes the Lfts for all those ending
67
activities concerned. But the Lft of all other activities will be equal to the smallest of the Lst
values of their immediately succeeding activities.
(vi) Latest Starting Time (LST). It is the latest possible time at which an activity can
start without disturbing the overall time fixed for the project. It is shortly indicated as the Lst
and is put at the beginning of any activity just below its arrow line. It is determined by
deducting the estimated time of an activity from its Latest finishing time thus already fixed.
The Lst of an activity ordinarily becomes the Lft for all its immediate preceding activities.
But at the point of a „Burst‟ event (where there are two are more activities succeeding the
event) the smallest of the Lsts of the succeeding activities becomes Lft of a preceding
activity.
(vii) Latest Event or Occurrence Time (ET). It is the latest possible time at which an
event can occur. It is shortly indicated as LT and is put just above or below an event. It is
equal to the Lft of the activities proceeding to the event. However, the LT of the starting
event of a project is zero.
(viii) Earliest Event or Occurrence Time (ET). It is the earliest possible time at which an
event can occur. It is shortly indicated as ET and is put just above or below an event along
with the LT. The ET of the starting event of a project is zero but the ET of all other events is
equal to the greatest of the Efts of its preceding activities.
(ix) Slack. The term „Slack‟ can be associated with both an event and an activity as well.
In relation to an event, a slack is the difference between its latest and earliest event time.
Thus, slack of an event = LT – ET. In relation to an activity, a slack will be synonymous to a
float, i.e.: (Lst – Est). But, slack is ordinarily associated with an event and in such case an
activity will have two slacks, viz.: Head Slack (i.e., slack of its head event) and Tail Slack
(i.e., slack of its tail event) where,
Head Slack or Head Event Slack
= LTH – ETH
(i.e., Latest occurrence time and Earliest occurrence time of the Head Event)
and, tail Slack or Tail Event Slack
= LTT – ETT
(i.e., Latest occurrence time and Earliest occurrence time of the tail Event)
A Slack can be positive or negative depending upon the latest and earliest occurrence
time of an event.
(x) Float. The term float refers to the slack of an activity. It is classified into three
types, viz., (1) Total Float, (2) Free Float, and (3) Independent Float. These are briefly
identified as under:
1. Total Float it is the time by which an activity other than the critical activity can be
manipulated without delaying the overall time of the project. This can be easily calculated by
taking the difference between the Lst and Est of an activity.
Thus,
Total Float = (Lst – Est) or (Lft – Eft)
It can give either positive or negative value. But it should be noted that there will be
no float for a critical activity for there is no question of delaying such an activity.
68
2. Free Float. It is that portion of the total float within which an activity can be
manipulated without affecting the float of its subsequent activities. It is calculated by
deducting the head event slack from the total float. Thus,
Free Float = Total Float – Head Event Slack or Head Slack, Such float may be either
positive or negative in value.
3. Independent Float, It is that part of the total float within which an activity can be
delayed for its start without affecting the floats of its preceding activities. It is found out by
subtracting the Tail event slack from the free float.
Thus,
Independent Float = Free Float – Tail event Slack or Tail Slack Independent float can
be positive only and where its working obtains negative value the same should be made up to
zero. However, Total float and Free float can have either positive or negative value. Positive
float implies that there are idle time and resources for the activities which could be utilized
otherwise. Negative float indicates that the activities concerned are short of time and
resources. Both positive and negative floats are undesirable.
Illustration 1: Draw a network diagram from the following information.
B and C follow A, E follows B, F follows D and C and G follows E and F. Where A, B, C,
D, E, F and G are the activities requiring 3, 7, 4, 2, 7, 4 and 7 days respectively.
Solution: Network Diagram
Here, the arrow marks represented by A, B, C, D, E, F and G are the different activities and
the circles represented by 1, 2, 3 etc. are the events.
Illustration 2: Draw a Network diagram from the following.
Activity Name of the Pre-requisite Estimated
Event Event Activity Activity Time
1 2 A None 4
1 3 B None 2
1 4 C None 2
1 5 D None 2
2 8 E A –
3 7 F B 3
4 6 G C 2
5 8 H D –
7 8 I F&J 4
8 7 J G 3
69
Solution:
Illustration 3: From the following information draw up a Network diagram and

indicate the Est, Lst, Eft and Lft of the various activities.
Name of the Activity Pre-requisite Activity Estimated Time (days)
A None 2
B None 3
C None 4
D A 6
E B 7
F C 5
G D&E 8
H B 9
I H&F 5
Solution: Network Diagram
Illustration 4: From the following table of information find out the following:
a) Critical path and non–critical paths through a Network diagram.
b) Latest event times, Earliest event time and slacks of the events.
c) Different slacks of the activities.
70
d) Different floats of the activities.

Activity: 1 – 2, 1 – 3, 2 – 3, 2 – 4, 3–4, 4–5.
Estimated Time (days): 20, 25, 10, 12, 6, 10
Solution: Network diagram
a) Different Paths:
(i) A  D  F requiring 20 + 12 + 10 = 42 days
(ii) A  C  E  F requiring 20 + 10 + 6 + 10 = 46 days (Maximum)
(iii) B  E  F requiring 25 + 6 + 10 = 41 days
Thus, the critical path is A  C  E  F requiring maximum number of days, i.e.,
46 days and the non–critical paths are:
1) A  D  F requiring only 42 days and
2) B  E  F requiring only 41 days
(b) Latest event time, Earliest event Time and Slacks of the Events
Slacks
Events LT ET
LT – ET
1 0 0 0
2 20 20 0
3 30 30 0
4 36 36 0
5 46 46 0
LT  Latest Event Time, ET  Earliest Event Time
(c) Slacks of the Activities

Head Slack Slacks
Activities
LTH - ETH LT – ET
A 20 – 20 = 0 0–0=0
B 30 – 30 = 0 0–0=0
C 30 – 30 = 0 20 – 20 = 0
D 36 – 36 = 0 20 – 20 = 0
71
E 36 – 36 = 0 30 – 30 = 0
F 46 – 46 = 0 36 – 36 = 0
LTH  Latest Occurrence Time of the Head Event
ETH  Earliest Occurrence Time of the Head Event
LTT  Latest Occurrence Time of the Tail Event
ETT  Earliest Occurrence Time of the Tail Event
(d) Calculation of Different floats of the Activities
Earliest Latest Total Float Independent
Free Float
Activities Lst–Est or Float
Start Finish Start Finish (TF–HS)
Lft–Eft (FF–TS)
A 0 20 0 20 0 0–0=0 0–0=0
B 0 25 5 30 5 5–0=5 5–0=5
C 20 30 20 30 0 0–0=0 0–0=0
D 20 32 24 36 4 4–0=4 4–0=4
E 30 36 20 36 0 0–0=0 0–0=0
F 36 46 36 46 0 0–0=0 0–0=0
Where, TF = Total float, HS = Head Slack
FF = Free Float, TS = Tail Slack
Lst = Latest starting time
Est = Earliest Starting time
Lft = Latest finishing time
and Eft = Earliest finishing time
Critical Path Method (CPM)

Critical Path Method (CPM) is one of the two important quantitative techniques of
Network analysis. This technique was introduced for the first time in the year 1956 at the E.I.
due Port Nemours and Co. of America to aid in planning and scheduling of the projects. It
uses two time and two cost estimates for each activity. One time and Cost estimate is made
for the normal situation and the other time and cost estimate is made for the crash or critical
situation. However, this technique operates with an assumption of precise known time and it
does not make use of the statistical analysis in determining the time estimates. It is a
graphical technique which involves the drawal of a Network diagram and its analysis to
indicate the critical path, non-critical paths, slacks and different floats of the various activities
explained earlier. It has the potential scheduling of a project in minimum possible time and
cost in the light of the given constraints.
As explained earlier, critical path is the longest path which can be easily found out
from among the different possible paths indicated through the network diagram with their
respective times. It is the path which takes the maximum time to complete a project. In other
words this path indicates the minimum time that will be required to complete a project. In the
Illustration 4 analysed above the path A  C  E  F requiring (20 + 10 + 6 + 10) 46 days
in toto is the critical path for completing the project. This is because this path is the longest
of all the time different path indicated in the network diagram in the said illustration 4 viz.,
1st Path: A  D  F requiring 20 + 12 + 10 = 42 days
72
2nd Path: A  C  E  F requiring 20 + 10 + 6 +10 = 46 days

3rd Path: B  E  F requiring 25 + 6 + 10 = 41 days
It is to be highlighted here, that all the activities, viz., A, C, E and F which are
associated with such critical path are called the critical activities or bottleneck activities. All
other activities which are not associated with the critical path viz. B and D in the said project
are non-critical activities. Non-critical activities can be delayed to the extent of the float
available without affecting the overall time stipulated for the project as they have a certain
amount of spare time. But the critical activities can never be delayed as they do not have any
float or spare time available.
Importance of Critical Path

Importance of the critical path and the analysis of the network diagram in the light of
the critical path can be highlighted as under:
(1) The critical path indicates the activities which must be performed without any
delay in order that the project may be completed within the stipulated time.
(2) The critical path indicates the activities which must be performed more rapidly in
order that the completion time of the entire project may be reduced.
(3) The critical both also indicates the activities, the performance of which can be
delayed to a certain extent.
(4) The critical path determined in the first instance may also aid in the determination
of another path as critical through advance planning and improvement.
(5) The critical path draws attention of the management to the important facts, spots
out the potential bottlenecks and avoids unnecessary pressure on other paths that do not help
in the earlier completion of the project.
(6) On the basis of the different floats of the different activities found out the critical
path makes it possible to make the following adjustments in the project.
a) Reduction of time estimates of the critical activities.
b) Elimination of some non-critical activities.
c) Introduction of some more resources.
d) Transfer of some resources from non-critical activities to critical activities.
e) Drawal of the revised network diagram to reduce to the completion time of the
project.
Limitation of CPM
In spite of the above importance, the CPM suffers from the following limitation:
a) It spite of the above assumption of a precise known time for each activity which
may not be true in real situation.
b) It does not make use of the statistical analysis in the determination of the time
estimates for each activity.
c) It requires repetition of the evaluation of the entire project each time a change is
introduced to the network. This is a very difficult and cumbersome process.
73
d) It cannot serve as a dynamic controlling device as it was introduced as a static

planning model.
Illustration 5: The following table gives the information relating to the various
activities concerning a project.
Time Estimated
Name of the Activity Pre–requisite Activity
(in days)
A None 2
B A 3
C A 4
D B and C 6
E None 2
F E 8
From the above information relating to the project:

(i) Draw up a Network diagram
(ii) Determine the critical path
(iii) Ascertain the minimum completion time of the project.
(iv) List the critical and non-critical activities.
(v) Find out the total float, Free float and independent float for the various
activities
Solution: (i) Network diagram

(ii) Determination of the critical path:
Different paths are:
(a) A  B  D requiring 2 + 3 + 6 = 11 days
(b) A  C  D requiring 2 + 4 +6 = 12 days
(c) E  F requiring 2 + 8 = 10 days
From the different paths cited above, the longest path A  C  E requiring the
maximum time 12 days is the critical path for the project.
(iii) The time of 12 days required for the critical path is the minimum time required
for the completion of the project.
(iv) List of the critical activities:
i. A
ii. C
iii. D
List of the None-critical activities
i. B
ii. E
iii. F
(v) (a) Table showing the Slacks of the Events
Slacks
Events LT ET
LT – ET
1 0 0 0
74
2 2 2 0
3 6 6 0
4 4 2 2
5 12 12 0
(b) Table showing the slacks of the activities

Head Slack Slacks
Activities
LTH - ETH LT – ET
A 2–2=0 0–0=0
B 6–6=0 2–2=0
C 6–6=0 2–2=0
D 12 – 12 = 0 6–6=0
E 4–2=2 0–0=0
F 12 – 12 = 0 4 – 2 =2
(iv) Computation of the Different Floats
Earliest Latest Total Float Independent
Free Float
Activities Lst–Est or Float
Start Finish Start Finish (TF–HS)
Lft–Eft (FF–TS)
A 0 2 0 0 0 0–0=0 0–0=0
B 0 5 3 6 1 1–0=1 1–0=1
C 2 6 2 6 0 0–0=0 0–0=0
D 6 12 6 12 0 0–0=0 0–0=0
E 0 2 2 4 2 2–0=2 0–0=0
F 2 10 4 12 2 2–0=2 2–2=0
Allocation and Levelling of Resources (Through CPM)

Resources allocation or scheduling of resources means the allocation of the available
resources to various activities. Levelling of resources, on the other hand, means allocation of
resources to various activities in such a manner that the variation in resources requirements
are minimized and the resource demand is evened out.
The task of allocation of resources and their levelling is of vital importance and it
involves various types of problem viz. (i) Resources levelling and (ii) Limited Resource
allocation. These problems are well tackled with the help of Critical Path Method. For this
the following steps to be taken up:
(A) Steps for Levelling of Resources

1) Draw up a list of the resources required
2) Draw up a profile for each of the resources
3) Determine the periods of the highest and lowest demand.
4) Alter the times of „start‟ and „finish‟ of the non-critical activities in order of their
higher floats and repeat this process till the peaks are lowered down to the level of the
troughs as far as possible without affecting the overall time limit of the project.
75
(B) Steps for Tackling the Problem of Limited Resources

1) Allocate the resources to the various activities serially through a Network diagram.
2) When there is competition among the several activities for the same resources give
preference to the activity having maximum float for the said float period.
3) If the floats of the competing activities appear to be equal, postpone the activity
requiring less time for its float period.
4) Reschedule the non-critical activities if possible to free the resources for scheduling
the critical activities.
The following illustrations will show how the problems of Resource levelling and
limited resources allocation are tackled through Critical Path Method.
Illustration 6: (On Resource Levelling) From the data given below, make the
levelling and limited of the resources as far as possible and work out the aggregate resources
requirements day by day
Time Labour
Activity Events Estimates requirements (in
(in days) units)
A 1–1 4 9
B 1–3 2 3
C 1–4 2 6
D 1–5 2 4
E 2–8 0 0
F 3–7 3 8
G 4–6 2 7
H 6–7 3 2
I 7–8 4 1
J 5–8 0 0
Solution: Net problem

76
From the above network diagram it appear that the longest path or critical path of the
project is
C  G  H  I requiring
2 + 2+ 3 + 4 = 11 days
and the non-critical activities are A, B, D and F.
Thus the minimum number of days required for the project is 11 days during which
the resources will be used up.
1. Diagram Showing the Serial Allocation of Resources (1st Instance)
The figures put above the arrow marks of the activities indicate the units of labour
resources required during each day for the part of each activity.
The above diagram shows a wide variation shows a wide variation between the peaks
and troughs of the labour allocations i.e., 24 to 1. So it is necessary to smooth out the
allocation of resources by tackling the non-critical activities viz., A, B, D and F on the basis
floats calculated as under:
Calculation of the Floats
Total
Activity Est Lst Eft Lft
Float
1 2 3 4 5 6
A 1–2 0 7 4 11 7
B 1–3 0 2 2 4 2
C 1–4 0 0 2 2 0
D 1–5 0 2 2 4 2
F 3–7 2 4 5 7 2
G 4–6 2 2 4 4 0
H 6–7 4 4 7 7 0
I 7–8 7 7 11 11 0
The above table shows that the non-critical activity „A‟ has the maximum float and
that it can be postponed for 7 days without disturbing the overall completion time of the
77
project. Thus, to reduce the magnitude between the peaks and troughs the allocation of the
resources will be made as under:
II. Diagram Showing the Reallocation of Resource
From the above modified diagram it appears that the magnitude of resource allocation
between the peaks and troughs has been reduced from 1 to 24 to 2 to 15. This indicates that
we can make further modifications in the resource allocation to reduce such wider magnitude.
According to the float table we can delay the start of the activity „B‟ (1-3) by 2 days without
imparting the overall completion time. Thus the third modifying diagram of resource
levelling will be as follows:
III. Diagram Showing the Levelling of Resource

78
The diagram drawn above shows that the resource allocation has now been perfectly
levelled up which of course may not be possible in all the cases. However, the magnitude of
allocation between the peaks and trough can be reduced in the manner.
Illustration 7: (On limited resources). From the following data relating to a project
show how the project is to be scheduled to be completed as soon as possible assuming that
only 10 units of labour can be made available a day:
Data
Time Labour
Activity Events
Estimates requirements per day
A 1–2 1 7
B 2–3 4 1
C 2–4 3 4
D 2–5 2 3
E 2–6 2 5
F 3–9 2 8
G 4–7 2 2
H 5–7 3 6
I 6–8 0 0
J 7–8 1 9
K 8–9 3 10
Solution: Net Diagram
Figures indicated within the brackets above the arrow marks are the units of
manpower required per day. The above diagram shows that the manpower requirements on
various days do not remain within the constraint of 10 units of labour. Hence, it needs
scheduling in the following manner:
79
Statement Showing the Floats

Time
Total
Activity estimat Est Lst Eft Lft
Float
e
1 2 3 4 5 6 7
A 1–2 1 0 0 1 1 0
B 2–3 4 1 4 5 8 3
C 2–4 3 1 1 4 4 0
D 2–5 2 1 1 3 3 0
E 2–6 2 1 5 3 7 4
F 3–9 2 5 8 7 10 3
G 4–7 2 4 4 6 6 0
H 5–7 3 3 3 6 6 0
I 6–8 0 3 7 3 7 4
J 7–8 1 6 6 7 7 0
K 8–9 3 7 7 10 10 0
The above net work diagram shows that there are two critical paths for the project as
under:
(i) A  C  G  J  K requiring (1 + 3 + 2 + 1 + 3) = 10 days
(ii) A  D  H  J  K requiring (1 + 2 – 3 + 1 + 3) = 10 days
Thus, the critical activities are A, C, D, G, H, H and K, which must be started at their
earliest start time unless the constraint of manpower (i.e., 10 units per day) prevents in the
matter.
Further, the activity I (6–8) is a dummy activity which is there simply to facilitate the
sequencing and that the activity K (8–9) cannot be started till activities E (2–6) and I (6–8)
are completed. Hence, all critical activities excepting the activity K (8–9) should be started at
their earliest start time. Activity E ( 2–6), although has a float of 4 days cannot start on the
6th day nor earlier than the 8th day in view of the constraint of 10 units of labour available on
a day. Activity I (6–8) being a dummy activity can be ignored totally irrespective of its float
of 4 days. Both the activities B (2–3) and F (3–9) have floats of 3 days for which the starting
time of each of them can be delayed for 3 days. Thus, the activity B (2–3) can start from the
5th day instead of the 2nd day. But the activity F (3–9) can start only from the 10th day instead
of the 6th and the 9th day in view of the shortage of the manpower beyond 10 units on a day.
In view of the above arrangements in the light of the given constraints the last critical
activity K (8–9) can be started only from the 12th day instead of the earlier 8th day. As such
the overall completion time of the project will be increased to 14 days from 10 days. Thus,
the project will be scheduled in the revised network diagram as follows:
80
Revised Network Diagram

(Showing the Scheduling of the project)
The above diagram keeps the manpower requirements within the constraint of 10
units of labour on a day.
Time Cost Analysis (Through CPM)
The execution of a project needs some time and fund both of which are very precious
for the management. But the relation between the two is obviously an inverse one i.e., more
the time less is the fund and less the time more is the fund required for a project. Thus, it will
be a critical task for the manager to decide over the time-cost relationship concerning his
project. If he considers the time to be more precious then he should plan out to complete the
project within the shortest possible time by reducing the estimated time and investing more
funds in the project. This type of execution of work is called as the execution of a work on
Crash basis or hurry up basis, the manager must see that the basis thus adopted proves most
economical for the management. For this he has to make comparative analysis of the various
paths in the light of the time-cost trade off concept. Critical path method helps a lot in taking
a wise decision in such matter. The following illustration will make the point in issue more
clear.
Illustration 8: From the following data determine the minimum possible time and
maximum possible find that proves most economical to complete a project in issue.
Timed Estimate Cost Estimate
Name of the
Normal Crash Normal Crash
Activity
(Days) (Days) Rs. Rs.
A 1–2 3 2 8,000 10,000
B 1–3 5 3 15,000 20,000
C 1–4 4 3 25,000 32,000
D 2–5 2 1 4,000 6,000
E 3–5 3 2 6,000 8,000
F 4–6 9 8 45,000 60,000
G 5–7 8 6 35,000 50,000
H 3–6 7 6½ 30,000 40,000
81
I 6–7 9 7 35,000 42,000

Total 2,03,000 2,68,000
Solution: The above problem involves the time cost analysis which can be solved
through CPM as under:
(a) Network Diagram
Figures above the arrows represent cost and below the arrows represent time required
on normal basis.
(b) Different paths and their required time and cost.
(i) A  D  G require 13 days and Rs. 49,000
(ii) B  E  G require 16 days and Rs. 56,000
(iii) B  H  I require 21 days and Rs. 80,000
(iv) C  F  I require 22 days and Rs. 1,05,000
Thus, Path C  F  1 in the critical path.
(c) Table Showing the Possible Reduction in Time and Increment in Cost under Crash
Programme
Name of the Total of Incremental
Critical Reducible Time Incremental Cost per day
Activities Cost (Rs.) (Rs.)
C 1 7,000 7,000
F 1 15,000 15,000
I 2 7,000 3,500
Reduction in time required only in case of critical activities that fall on the critical
path, viz., C  F  I. Since the non-critical activities have floats and reduction of time in
respect there of would not reduce the overall time of the project.
(D) Table showing the Time-Cost Analysis (Table given below:)
Activities
Critical Time required Cost required
On Normal On Crash
Path (in days) (Rs.)
basis basis
All None CF1 22 203000
A, B, C, D, 210000
I CF1
E, F, G, H (+ 7000 for I)
82
A, B, D, E, CF1 217000
C, I
F, G, H BH1 (+ 7000 for C)
243500
C, F, I
A, D, E,
fully and CF1
G, H
B for 1 day
From the above analysis table it transpires that the minimum possible times is 18 days
and maximum possible cost in Rs. 2,34,500 to complete the project most economically.
Note: Activity „B‟ could have been reduced by 2 days to complete the path BHI
within 17 days but it is not possible because even if all the activates are done on crash basis
the total project time comes to 18 days and the critical path happens to be C  F  I. Hence
the activates „B‟ can be worked out on crash basis only for a day with an additional cost of
Rs. 2,500 and the activity „F‟ can be worked out for 1 day with additional cost of Rs. 15,000.
Programme Evaluation and Review Technique (PERT)
Programme Evaluation and Review Technique (PERT) is another quantitative
technique of Network analysis. This technique was introduced for the first time in the year
1958 at the E.I due Pnrt Nemous & Co. of U.S.A. for use in defence projects. Now it is being
used as a formidable instrument by the project planners and controllers in the determination
of the expected total time a project is likely to take for its completion and for analysing and
administering the large complex projects to be performed in some sequence. It is defined as
an approach of multiple time estimates in which the following four time estimates are
employed through statistical analysis.
(i) The most optimistic time (to)
(ii) The most pessimistic time (tp)
(iii) The most likely time (tm)
(iv) The most expected time (te)
Optimistic time(to): It refers to that shortest possible time within which an activity
can be performed when everything goes perfectly without any disturbance. It is the most
ideal time which rarely occurs in practice and thus its chance of taking place may be 1%.
Pessimistic time (tp): It refers to that longest possible time which is likely to take
place to complete an activity when all sorts of complicacies and unusual delays take place. It
is another extreme time which rarely takes place in actual practice and thus its chance of
occurrence is also 1%.
Most likely Time(tm): It refers to that time which is most likely to take place to
complete a particular activity or a project. This is the best possible estimate of time which
normally occurs in the completion of an activity.
Expected Time (te): It refers to that time within which an activity is expected to be
accomplished. It is determined for each activity on the basis of the three estimated times
aforestated, viz., optimistic time, most likely time and pessimistic time. It is calculated as a
weighted average of these three types of activates. The weights assigned for the purpose are:
For optimistic time – 1
83
For pessimistic time 1

For most likely time 4
Thus, the expected time of activity is calculated with the following formula.
It is to be noted here that under CPM the precise known time for each activity was
assumed but under PERT it is assumed that the time to perform an activity is uncertain and as
such here we calculate the expected time for each activity on the basis of the three time
estimates cited above, viz., optimistic time, most likely time and pessimistic time. After the
expected times for each activity are estimated in the above manner, determination of all other
relevant facts, viz., Different paths of performance including the critical path, minimum time
required to complete a project, earliest and latest start and finish times of the activates,
earliest and latest times of occurrence of events, different, floats and slacks etc. are found out
through Network diagram just on similar lines of that of the critical path method.
In addition to the above, PERT also enables us to determine the probabilities of
completing an activity and the entire project within a given time. For this PERT allows us to
make use of the following statistical devices to ascertain the values of the various relevant
factors as follows:
1. Standard Deviation of an Activity
2. Variance of an Activity
3. Standard Deviation of a Path
4. Variance of path:
where,
a1, a2, a3, ... and an represent different activates along a particular path and (p)
represents a particular path.
The above formula for variance of a path is based upon the assumptions that the
variance of sum of random variables equals the sum of the variance of each random variable
of independent nature.
84
It is to be noted that for finding the standard deviation and variance of the entire
project the standard deviation and variance of the critical path are to be found out as per the
above formulae (3) and (4) respectively.
5. Probability of a Path being completed within a given time.
This is given by
P(p) = Table value of the standard normal variate (z) as per the
cumulative normal distribution table (vide the Appendix
at the end of the book.)
z=
where
Z = Standard normal variate of a path
Xp = Give time of the path
ap = Mean or length of the given path
(p) = Standard deviation of the path
6. Probability of the Project being completed within a given time:
Pw = Products of the probabilities of the different independent
paths
i.e., = Product of the probability of critical and that of the other
independent paths.
= z(cp)  z(oip)
The following illustrations will make the procedure of PERT analysis more clear.
Illustration 9: From the following information relating to the various activities of a
project find the expected time (te) for the different activities.
Optimistic time Most likely time Pessimistic
Activities (days) (days) Time (days)
to tm tp
A 2 4 6
B 6 8 10
C 1 5 15
D 1 5 9
E 6 8 10
F 5 7 8
Solution: Expected time or

85
The expected time of the above six activities will be calculated as under:
Formulation
Expected Time
Activity
(in days)
A (2 + 4  4 + 6)/6 4
B (6 + 4  8 + 10)/6 8
C (1 + 4  5 + 15)/6 6
D (1 + 4  5 + 9)/6 5
E (6 + 4  8 + 10)/6 8
F (5 + 4  7 + 8)/6
Illustration 10: From the following data relating to a project.

a) Find the expected duration, standard deviation and variance of each activity.
b) Draw up the project network diagram and trace out all the possible paths from it.
c) Find the expected project length.
d) Find the probability that the project will be completed within 4 weeks.
e) Calculate the Est, Eft, Lst and Lft values for each activity. Also find the total float
for the activities.
Activity Name of Duration

Event Event Activity Optimistic Most likely Pessimistic
1 2 3 4 5 6
1 2 A 1 1 7
1 3 B 3 5 7
1 4 C 2 2 8
2 5 D 1 1 1
3 5 E 3 6 9
4 6 F 2 5 8
5 6 G 4 6 14
6 7 H 6 8 10
5 7 I 3 7 11
Solution: (a) Computation of Expected duration, Standard Deviation and Variance of

each activity.
Standard
Expected duration
deviation
Name of (week)
Variance
the
()2
Activity
A (1 + 4  1 + 7)/6 = 2 (7 – 1)/6 = 1 1
86
B (3 + 4  5 + 7)/6 = 5 (7 – 3)/6 = 67 45
C (2 + 4  2 + 8)/6 = 3 (8 – 2)/6 = 1 1.00
D (1 + 4  1 + 1)/6 = 1 (1 – 1)/6 = 0 0
E (3 + 4  6 + 9)/6 = 6 (9 – 3)/6 = 1 1.00
F (2 + 4  5 + 8)/6 = 5 (8 – 2)/6 = 1 1.00
G (4 + 4  6 + 14)/6 = 7 (14 – 4)/6 = 1.67 2.79
H (6 + 4  8 + 10)/6 = 8 (10 – 6)/6 = .67 .45
I (3 + 4  7 + 11)/6 = 7 (11 – 3)/6 = 1.33 1.77
(b) Project Network Diagram (on the basis of the Expected duration i.e. te)
Possible Paths
(i) A  D  I requiring 2 + 1 + 7 = 10 weeks
(ii) A  D  G  H requiring 2 + 1 + 7 + 8 = 18 weeks
(iii) B  E  I requiring 5 + 6 + 7 = 18 weeks
(iv) B  E  G  H requiring 5 + 6 + 7 + 8 = 26 weeks
(v) C  F  H requiring 3 + 5 + 8 = 16 weeks
(c) Expected project length is 26 weeks, i.e., the length of the longest path or the
critical path i.e. B  E  G  H.
(d) Probability that the project will be completed within 30 weeks.
For this we must find in the following manner, the product of the probabilities of the
independent paths, viz., (1) The critical path, i.e., B  E  G  H and (2) A
DI
Thus, probability of completing the C.P. within 30 weeks in given by
where,
Zcp = Standard normal variate of the critical path.
X= Given time = 30 weeks
acp = Mean or length of the critical path = 26 weeks and
(cp) = Standard deviation of critical path.
87
(cp) =
=
= 2.16
Therefore,
Table value of Z at 1.85 = .297684

Further, probability of the other independent path A  D  I being completed with
30 weeks.
where,
Z(vip) = Standard normal variate of the other independent path.
X= Given time = 30 weeks
a(vip) = Mean or length of the other independent path i.e., 10 weeks
(cip) = Standard deviation of the other independent path.
=
Thus, Z(vip) =
Table value of Z at 12.04 approximately = 1. Thus the probability of the project

being completed within 30 weeks.
or P(w) = Z(cp)  Z(vip)
P(w) = .96784  1 = .96784
Therefore, the probability of the project not being completed within 30 weeks.
or q(w) = 1 – Pw
= 1 – .96784 = .03216
(e) Calculation of the Est, Eft, Lst, Lft and total floats of the various activities
Expected
Activity duration Est Eft Lst Lft Floats
tE.
(1–2) A 2 0 2 8 10 8
88
(1–3) B 5 0 5 0 5 0
(1–4) C 3 0 3 10 13 10
(2–5) D 1 2 3 10 11 8
(3–5) E 6 5 11 5 11 8
(4–6) F 5 3 8 13 18 10
(5–6) G 7 11 18 11 18 0
(6–7) H 8 18 26 18 26 0
(5–7) I 7 11 18 19 26 8
Merits and Demerits of PERT
Having thus analysed, the merits and demerits of PERT as Quantitative Technique
can be summed up as follows:
Merits:
1. It enables a manager to understand easily the relationship that exists between the
activities in a project.
2. It enables a manager to know in advance where the trouble may occur, where more
supervision may be needed and where resources may be transferred to keep the project on
schedule.
3. It compels the manager to plan carefully and study how the various activities fit in the
project.
4. It draws attention of the management to the critical activities of the project.
5. It suggests areas of increasing efficiency, decreasing cost and maximizing profits.
6. It enables the use of statistical analysis.
7. It makes possible a forward looking type of control.
8. It compels the management for taking necessary action at the right time without any
let up.
9. It provides upto date information through frequent reporting, data processing and
programme analysis.
10. It helps in formulating a new schedule when the existing ones cannot meet the
situation.
11. It helps in minimizing delays and disruptions by scheduling the time and budgeting
the resources.
12. It helps in coordinating the various parts of the project and expediting the mode of
operation for completing the project in time.
13. It permits more effective planning and control.
Demerits:
1. It does not lay any emphasis on the cost of a project but time only.
2. It does not help in routine planning of the recurring events.
3. Errors in time estimates under the PERT make the network diagram and the critical
path etc. meaningless.
4. In the calculation of probability under the PERT it is assumed that a large number of
independent activities operate on critical path and that the distribution of total time is
normal. This may not hold good in a peculiar situation.
89
5. For effective control, PERT requires, frequent up-to-date information and revision in
calculation which may be quite costly for the management.
6. It does not consider the matter of resources required for various types of activities of a
project.
Comparison between CPM and PERT
The critical path method and programme Evaluation and Review Technique have
some characteristics in common. These are depicted as under:
1. Both CPM and PERT are quantitative techniques of Network analysis.
2. Both CPM and PERT are used by the management, as tools for taking decisions
relating to a large complex project.
3. Both CPM and PERT technique involve drawal of Network diagram and its analysis
on various scores.
Contrasts between the CPM and PERT
CPM and PERT contrast each other on the following points:
CPM PERT
1. It is a deterministic model under 1. It is a probabilistic model under
which result is ascertained in a which result is estimated in a
manner of certainty manner of probability
2. It deals with the activities of 2. It deals with the activities of
precise well known time uncertain time.
3. It is used for repetitive jobs like 3. It is known for non-repetitive
residential construction. jobs like planning and scheduling of
research programmes.
4. It is activity oriented in as much 4. It is event oriented in as much as
as its results are calculated on the its results are calculated on the basis
basis of the activities. of events.
5. It does not make use of Dummy 5. It makes use of Dummy
activities. activities to represent the proper
sequencing of the activities.
6. It deals with costs of a project 6. It has nothing to do with cost of
schedules and their minimisation. a project.
7. It deals with the concept of 7. It does not deal with concept of
crashing. crashing.
8. Its calculation is based on one 8. It finds out expected time of
type of time estimation that is each activity on the basis of three
precisely known. types of estimates, viz., optimistic
time, pessimistic time and most
likely time.
9. It can be used as a control 9. It is used as an important control
devices as it requires repetition of device as it assists the management
the entire evaluation of the project in controlling a project by constant
each time the changes are review of the delays in activities.
introduced to the network.
90
10. It does not make use of the 10. It makes use of the statistical
statistical devices in the devices, viz., standard deviation
determination of the time estimates. variance, probability (Z), and
normal distribution table in the
determination of probabilities of
completing or not completing a
project or a path within a given
time.

Unit I Operation Research Introduction: Characteristics of Operations Research

Uploaded by

Copyright:

Available Formats

Unit I Operation Research Introduction: Characteristics of Operations Research

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Unit I Operation Research Introduction: Characteristics of Operations Research

Uploaded by

Copyright:

Available Formats

UNIT I

OPERATION RESEARCH INTRODUCTION

Definitions and Characteristics of Operations Research

Limitations of Operations Research

Thus Linear Programming is a Mathematical technique which is applied in the form

Types of Problems (where L.P. can be applied)

TYPES OF LINEAR PROGRAMMING PROBLEMS

Formulation of L.P. Problem

a11X1 + a12X2 + ... + a1nXn  b1

2nd Step: Decision Table

3rd Step: Constitution of the different linear functions

Minimum vitamin contents needed per feed mix in units

X2 = Decision variable for the feed Q

4th Step: Formulation (Table) of the L.P.P.

2nd Step: Decision Table.

3rd Step: Constitution of the different Linear Functions.

4th Step: Formulation of the L.P.P.

Solution of the Formulated Problems

Sewing Machine X2 240 1 18

(ii) Formulation of L.P.P.

Subject to non – negativity condition

7th Step: Evaluation of the Objective Function

(iii) Formulation of L.P.P.

5. Graphic Representation of the Constraints lines of the L.P.P.

6. Evaluation of the Objective Function by corner points.

Step 1: Write the L.P.P. in the standard form

coefficients in the objective function.

Step 6: Compute the ratios be and chose the minimum

Maximize z = 4x1 + 3x2 + 0.x3 + 0.x4 + 0.x5 +0.x6

For a minimization problem, optimization is obtained when all zj - cj  0. If we have

New Simplex table

The new simplex table is

zj - cj 0 0 -12/5 -4/5 -4/5 0

x1 enters as a basic variable in the place of x6 2 is the pivotal element.

The new simplex table is

0 0 0 1/3 1 -2/3 -8/3

0 0 1 -2/3 0 -2/3 -2/3 -

The new table are

Since all zj - cj  0, current solution is optimum

The constrains is standard form are

enters as a basic variable in the place of x3 1/3 is the pivotal element.

The simplex table is

enters as a basic variable in the place of x6 3 is the pivotal element.

The new simplex table is

4 x1 = 3/5 1 0 3/5 1/5 –1/5 0

variable in the place of x6 1 is the pivotal element.

The new simplex table is

1 x6 = 6/5 0 1 –1/5 0 0 3/5

The starting simplex table is

The new table is

The new table is

The new starting solution is

x4 enters as a basic variable in the place of x6 is the pivotal element.

The new table is

For the minimum problem, since all zj - cj  0, current solution is optimum.

x1 = 3/5, x2 = 6/5, and min

Phase I: Minimize u = x6 + x7 subject to the above conditions

The new table is