Are you sure?
This action might not be possible to undo. Are you sure you want to continue?
[
]
Production Engineering Department
Shantilal Shah Engineering College
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 2
CERTIFICATE
This is to certify that Mr./Miss
_______________________________________________________________
Student of semester  VIII Production
Engineering,Roll No. _______ of S.S.Engineering College has
Satisfactorily accomplished his/her term work by
submitting this file of ______________________________________
on Date: ___________
Examinor Head of the Dept
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 3
INDEX
Sr.
No.
Name of Experiment Page
From To
Date of
Start
Date of
Complet
ion
Sign
1.
OPERATION RESEARCH
2.
STUDY OF LINEAR
PROGRAMMING METHODS,
MODELS FORMULATION &
ITS ROLE IN OPERATION
RESEARCH
3.
STUDY & IMPORTANCE OF
SIMPLEX METHOD IN LINEAR
PROGRAMMING
4.
STUDY OF DUAL PROBLEM,
ITS INTERPRETATION &
SENSITIVITY ANALYSIS
5.
SPECIAL CASE OF SIMPLEX
METHOD
6.
PROJECT EVALUATION AND
REVIEW TECHNIQUES &
CRITICAL PATH METHOD
(PERT/CPM)
7.
GAMES THEORY
8.
INVENTORY MANAGEMENT
9.
QUEUING THEORY
10.
REPLACEMENT THEORY
11.
STUDY ABOUT SEQUENCING
PROBLEM.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 4
EXPERIMENT.NO: 1
OPERATION RESEARCH
HISTORICAL DEVELOPMENT
It is generally agreed that Operation Research (OR) came into existence as a discipline
during World War II. However, a particular model and technique of Operation Research
can be traced back much earlier. The term Operation Research was coined as a result of
research on military operations during this war. Since the war involved strategic and tactical
problems, which were so complicated, that to expecting adequate solution from individuals
Operation Research specialists in a single discipline was unrealistic. Therefore, groups of
individuals who collectively and physical science were formed as special units within the
armed forced to deal with strategic and tactical problems of various military operations.
Such groups were first in England and the United States. One of the groups in England
came to be known as Blackett‟s Circus. This group under the leadership of Prof.
P.M.S.Blackett was attached to the Radar Operation Research unit and assigned the
problem of analyzing the coordination of Radar equipment at gun sites. Following the
success of this group, such mixed team approach was adopted in other allied nations.
A key person in post war development of Operation Research was George B Dantzig. In
1947, he developed linear programming and its solution methods known as simplex
method. Besides linear programming, many others were well developed before the end of
1950‟s.
During the 1950‟s, there was substantial progress in the applications of Operation Research
techniques for civilian activities along with a great interest in the professional development
and education in Operation Research. Many college and universities introduce Operation
Research in their curricula. They were generally schools of engineering, public
administration, business management, applied mathematics, economics, computer services,
etc. Today, however service organization such as banks, hospitals, libraries, airlines,
railways, etc. recognize the usefulness for in improving their efficiency. In 1948, an
Operation Research was formed in England which later changed the name to Operation
Research Society of America (ORSA) was found in 1952 and its journal, Operation
Research, first published in 1953. In the same year, the Institute of Management Science
(TIMS) was founded as an International Society to identify, extend and unify scientific
knowledge pertaining to management. Its journals, Management, Science, first appeared in
1954.
In India, Operation Research came into existence in 1949 when an Operation Research unit
was established at Regional Research Laboratory, Hyderabad. At the same time, Prof. R S
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 5
Verma, also ser up an Operation Research team at Defence Science Laboratory to solve
problems of store, purchase and planning. In 1933, Prof. P C Mahalanobis established an
Operation Research team in the Indian statistical institute, Calcutta to solve to solve
problems related to national planning and survey. The Operation Research Society of India
was founded in 1957 and started publishing its journal OPSEARCH. In same year, India
along with Japan became in London. The other member of IFORS was USA, UK,
FRANCE and what was when WEST GERMANY.
Because of OR‟s multi disciplinary character and application in varied fields, it has a good
future provided people devoted to OR study can help meet the needs of society. Some of the
problems the area of hospital management, energy conservation, environmental pollution
etc. which have been solved by OR specialists is an indication that OR can also contribute
towards the improvements in the pattern of social life and areas of global need. However, in
order to make the future of OR more bright, it specialists have to make good use of the
avenues open to them.
SIGNIFICANCE OF OPERATIONS RESEARCH
As the term implies OR involves research in military operations. This indicates the
approach as well as the area of its application. The OR approach is particularly useful in
balancing conflicting objectives (goals OR interests) where there are many alternatives
courses of action available to the decision makers.
In view of any problem situation involving the whole system, the decision maker, whatever
his specialization will need help and it is in the attempt to provide this assistance that OR
has been developed. OR attempts to resolve the conflicts of interest among various sections
of the organization and seeks the optimal solution which may not be acceptable to one
department but is in the interest of the organization as a whole. Further, OR is concerned
with providing the decision maker with decision aids (or rules) derived from:
 A total system orientation.
 Scientific methods of investigation.
 Models of reality generally based on quantitative measurement technique.
Thus, successful application of Operation Research technique for solving a problem must
involve:
 Constructing mathematical, economic and statistical model of the problem under
study to treat situations of complexity and uncertainly. This helps to view the problem
in it‟s entirely.
 Analyzing the relationships among different variables and parameters associated with
the problem so as to determine consequences of decision alternative.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 6
 Suggesting suitable measures of desirability in order to evaluate the relative merit of
decision alternatives.
OPERATION RESEARCH APPROACH
The future of Operation Research approach to any decision and control problems can be
considered under the following methodologies:
1. INTERDISCIPLINARY APPROACH: Interdisciplinary teamwork is essential
because while attempting to solve a complex management problem one person may not
have the complete knowledge of all aspect (such as economic, social, political,
psychological, engineering etc.) of it. This means we should not expect a desirable
solution to managerial problems. Therefore, a team of individuals specializing in that
field in order to arrive at an appropriate and desirable solution of the problem. However,
there are certain problem situations, which may be analyzed even by one individual.
2. METHODOLOGICAL APPROACH: “Operation Research is the application of
scientific methods, techniques and tools to problem involving the operation of system so
as to provide those in control of operations with optimum solutions to the problems”.
The scientific method consist of observing and defining the problems; formulating and
testing the hypothesis; and analyzing the results of the test. The data so obtained are
then used to decide whether the hypothesis is accepted, then the results should be
implemented. Otherwise, an alternative hypothesis has to be formulated.
3. WHOLISTIC APPROACH: Arriving at a decision, an operation research team
examines the relative importance of all confliction and multiple objectives and validity
of the claims of various departments of the organization from the perspective of the
whole organization.
4. OBJECTIVE APPROACH: An Operation Research approach seeks to obtain an
optimal solution to the problem under analysis. For this a measure of desirability is
defined based on the objectives of the organization. A measure of desirability so defined
is then used to compare alternative course of action with respect to their out come.
SCIENTIFIC METHOD IN OPERATION RESEARCH
The most important feature of Operation Research is the use of the scientific method and
building of decision models. There are three phases of the scientific method on which
Operation Research practice is based.
1. JUDGMENT PHASE: This phase includes
a) Identification of the reallife problem.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 7
b) Selection of an appropriate objective and the values of various variables related to
these objectives.
c) Application of the appropriate scale of measurement, i.e. deciding the measures of
effectiveness.
d) Formulation of an appropriate model of the problem, abstracting the essential
information‟s so that a solution at the decision maker‟s goals can be sought.
2. RESEARCH PHASE: This phase is the largest and longest between other two phases.
However, the remaining two also equally important as they provide the basis for a
scientific method. This phase utilize:
a) Observations and data collection for a better understanding of the problem.
b) Formulation and experimentation to test the hypothesis on the basis of additional
data.
c) Analysis of the available information and verification of the hypothesis using pre
established measures of desirability.
d) Predictions of various results form the hypothesis.
e) Generalization of the result and consideration of alternative method.
3. ACTION PHASE: This phase consists of making recommendation for implementing
the decision by an individual who is the position to implement results. This man must
be aware of the environment in which the problem occurred objective, assumption and
omissions of the model of the problem.
MODELS OF OPERATION RESEARCH
PHYSICAL MODELS
These models provide a physical appearance of the real object under study either reduced in
size/scaled up. Physical models are useful only in design problems because they are easy to
observe, build and describe. Problem such as portfolio selection, media selection,
production scheduling etc. cannot be analyzed with a physical model. Physical model are
classified into the following two categories:
1. Iconic Model: Iconic models retain some of the physical properties and
characteristics of the system they represent. An iconic model is either in an idealized
formula scaled version of the system. In other words, such models represent the
system as it is by scaling it up Operation Research down.
The iconic models are simple to conceive, specific and concrete. An iconic model is
used to describe the characteristics of the system rather than explanatory. This
means such models are used to represent a static event and characteristics, which are
not used in determining/predicting effects due to certain changes in actual system.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 8
For example, color of an atom does not play any vital role in the scientific study of
its structure.
2. Analogue Model: These models represent a system by the set properties different
from that of the original system does not resemble physically. After the problem is
solved, the solution is reinterpreted in terms of the original system.
For example, the organizational chart represents the state of formal relationships
existing between members of the organization. Map in different colors may
represent water, desert and other geographical features. These models are less
specific and concrete but easier to manipulate and more general than iconic models.
SYMBOLIC MODELS:
These models use letters, numbers and other symbols to represent the properties of the
system. These models are also used to represent relationship, which can be represented in a
physical form. Symbolic models can be classified into two categories:
1. Verbal Models: These models describe a situation in written Operation Research
spoken language. Written sentences, books, etc. are examples of a verbal model.
2. Mathematical Models: These models involve the use of mathematical symbols,
letters, numbers, and mathematical operators (+, , *, /) to represent relationship
among various variable of the system to describe its properties Operation Research
behavior.
The solution to such models is then obtained by applying suitable mathematical
techniques. Symbolic models are precise and abstract and can be manipulated by
using laws of mathematics.
BASIC OPERATIONS RESEARCH MODELS
There is no unique set of problems, which can be solved by using Operation Research
models Operation Research techniques. Several Operation Research models/techniques can
be grouped into some basic categories as given below:
1. ALLOCATION MODEL: Allocation model are used to allocate resources to activate
in such a way that some measure of effectiveness is optimized. Mathematical
programming is the broad term for the Operation Research technique solves allocation
problems.
When the solution values Operation Research decision variables for the problem are
restricted to being integer values / just zeroone values, the problem is classified as an
integer programming problem Operation Research a zeroone programming problem
respectively.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 9
The problem having multiple, conflicting and incommensurable objective functions
subject to linear constraints is called a goalprogramming problem. If the decision
variables in the linear programming problem depend on chance, then such a problem is
called a stochastic programming problem.
2. INVENTORY MODELS: Inventory models deal with the problem of determination
of how much to order at a point in time and when to place an order. The main objective
is to minimize the sum three conflicting inventory costs: the cost of holding Operation
Research carrying extra inventory, the cost of storage Operation Research delay in the
delivery of items when it is needed, a cost of ordering or setup. These are also useful in
dealing with quantity discounts and selective inventory control.
3. WAITING LINE MODELS (QUEUING): These models have been developed to
establish a tradeoff between costs of providing service and the waiting time of a
customer in the queuing system. Constructing a models entails describing the
components of the system: arrival process, queue structural and service process and
solving for the measure of performance – average length of waiting time, average time
spent by the customer in the line, traffic intensity, etc. of the waiting system.
4. COMPETITIVE MODELS (GAME THEORY): These models are used to
characterize the behaviors of two/more opponents who compete for the achievement of
conflicting goals. These models are classified according to several factors such as
number of competitors, sum of loss and gain and type of strategy which would yield
him the best / the worst outcomes.
5. NETWORK MODELS: These models are applied to the management of largescale
projects. PERT/CPM technique help in identifying potential trouble spots in project
through the identification of the critical path. These techniques improve project co
ordination and enable the efficient use of resources. Network methods are also used to
determine timeclose trade off, resource allocation and updating of activity times.
6. SEQUENCING MODELS: These models are used whenever there is program in
determining the sequence in which a number of tasks can be performed by a number of
service facilities such as hospital, plant etc. in such a way that some measure of
performance, for example, total time to process all the jobs on all the machines, is
optimized.
7. REPLACEMENT MODELS: These models are use when one must decide the
optimal time to replace equipment for one reason Operation Research the otherfor
instance, in the case of equipment whose efficiency deteriorates with time Operation
Research fails immediately and completely. For example, in case of an automobile the
user has his own measure of effectiveness. So there will not be single optimal answer
for everyone even if each automobile gives exactly the same service.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 10
8. DYNAMIC PROGRAMMING MODELS: Dynamic programming may be
considered as an out growth of mathematical programming and involves the
optimization of multistage decision processes. The method starts by dividing a given
problem into stages Operation Research subproblems and then solves that subproblem
sequentially until the solution to the original problem is obtained.
9. MARKOVCHAIN MODELS: These models are used for analyzing a system that
changes over a period of time among various possible outcomes/states. The models
while dealing with such systems describe transitions in terms of transition probabilities
of various states. These models have been used to test brandloyalty and brand
switching tendencies of consumers, where each system state is considered to be a
particular brand purchase. These have also been used in reliability analysis, where the
states of the system are the various levels of performance of the equipment being
monitored.
10. SIMULATION MODELS: These models are used to develop a method to evaluate
the merit of alternative courses of action by experimenting with a mathematical model
of the problems where various variables are random. That is, these provide a means of
or generating representative samples of the measures of performance variable. Thus,
repetition of the process by using the simulation model provides an indication of the
merit of alternative course of action with respect to the decision variable.
GENERAL METHODS
In general, the following three models are used solving Operation Research models. In all
these methods, values of decision variables are obtained that optimize the given objectives
function.
1. ANALYTICAL METHOD (DEDUCTIVE): In this method, classical optimization
techniques such as calculus, finite difference and graphs are used for solving and
Operation Research model. In this case, we have a general solution specified by symbol
and we can obtain the optimal solution in a noniterative manner. For example, in
inventory models, in order to calculate economic order quantity the analytical method
requires that the first derivative of the mathematical expression
TC = (D/Q) Cp + (Q/2) Ch
Where, TC = Total variable inventory cost;
Cp = Ordering cost per;
Q = size of an order;
D = annual demand;
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 11
Ch = carrying cost per time period.
Be taken and equated to zero as the first step toward identifying optimum value of Q* =
sqrt (2DCp/Ch). This is because of the concept of maximum and minimum for
optimality. Here, it may be noted that the validity of the conclusion depends only on the
validity of the mathematical approach that is being used.
2. NUMERICAL METHODS (ITERATIVE): When analytical methods fail to obtain
the solution of a particular problem due to its complexity in terms of constraints/number
of variables, then numerical methods is used to get the solution. In this method instead
of solving the problem directly, a general is applied to obtain a specific numerical
solution.
The numerical methods starts with a solution obtained by trail and error and as set of
rules for improving it towards optimally. The solution so obtained is then replaced by
the improved solution and the process of getting improved solution is repeated until
such improvement is not possible/ the cost of further calculation cannot be justified.
3. MONTE CARLO METHOD: This method is based upon the idea of experimenting
on a mathematical model by inserting into the model specific values of decision
variables at different points of time and under different conditions and then observing
their effect on the criterion chosen for variables. In this method, random samples of
specified random variables are drawn to know what is happening to the system for a
selected period of time under different conditions. The random samples form a
probability distribution that represents the real life system and from this probability
distribution, the value of the desired random variable can be estimated.
APPLICATION AND SCOPE OF OPERATON RESEARCH
Some of the industrial/ government/ business problems which can be analyzed by Operation
Research approach have been arranged functional area wise as follows: 
1. FINANCE AND ACCOUNTING:
 Divided policies, investment and portfolio management, auditing balance sheet,
cash flow analysis.
 Breakeven analysis, capital budgeting, cost allocation and control, financial
planning.
 Claim and complaint procedure, public accounting.
2. MARKETING:
 Selection of product – mix, marketing planning, exports planning
 Sales effort allocation and assignment.
 Advertising and media planning.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 12
3. PURCHASING, PROCUREMENT AND EXPLORATION:
 Optimal buying and recording under price quantity discount.
 Bidding policies.
 Transportation planning.
 Vendor analysis.
 Replacement policies.
4. PRODUCTION MANAGEMENT:
 Facilities Planning
 Manufacturing
 Maintenance and project scheduling
5. PERSONNEL MANAGEMENT:
 Manpower planning, wage/salary administration
 Skills and wages balancing
 Scheduling of training programmers
6. TECHNIQUES AND GENERAL MANAGEMENT:
 Decision support system and MIS; forecasting
 Organizational design and control
 Project Management, strategic planning
7. GOVERNMENT:
 Economic planning, natural resources, social planning, energy
 Urban and housing problem
 Military, police, pollution control etc.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 13
EXPERIMENT.NO: 2
STUDY OF LINEAR PROGRAMMING METHODS, MODELS
FORMULATION & ITS ROLE IN OPERATION RESEARCH
GUIDELINES ON LINEAR PROGRAMING MODEL
FORMULATION
The usefulness of linear programming as a tool for optimal decision making on resource
allocation is based on its applicability to many diversified decision problems. The effective
use and application require, as a first step, the mathematical formulation of the LP model
when the problem in words.
STEPS OF LP MODEL FORMULATION
Step 1:  Define Decision Variable:
a) Express each constraint in words. For this first see, whether the constraint is of form,
(at least than). Operation Research of the for (no longer than) Operation Research
= (Exactly equal to)
b) Then express the objective function in words.
c) Step 1(a) and 1(b) should then allow you to verbally identify the decision variable.
Step 2:  Formulate the Constraints:
Formulate all the constraints imposed buy the resource availability and express them as
linear equality Operation Research inequality in terms of the decision variables defined in
step 1.
Step 3:  Formulate the Objective Function:
Define the objective function. That is, determine whether the objective function is to be
maximized /minimized. Then express it as a linear function of decision variables multiplied
by their profit/cost contributions.
INTRODUCTION
An optimal as well as feasible solution an LP problem is obtained by choosing among
several values of decision variables X
1
, X
2
,…., X
n
the one set of values that satisfy the
given set of constraints simultaneously and also provides the optimal values of the given
objective function.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 14
GRAPHICAL SOLUTION METHODS OF LP PROBLEMS
While obtaining the optimal solution to the LP problem with the graphical method, the
statement of the following theorems of linear programming is used.
1. The collection of all feasible solutions to an LP problem constitutes a convex set whose
extreme points correspond to the basic feasible solutions.
2. There are finite numbers of basic feasible solutions within the feasible solution space.
3. If the optimal solution occurs at feasible solutions on the system Ax = b, x>0, is a
convex polyhedron, then at least one of the extreme points gives an optimal solution.
4. If the optimal solution occurs at more than one extreme point, then the value of
objective function will be the same for all convex combinations of these extreme points.
EXTREME POINT ENUMERATION APPROACH
This solution method for an LP problem divided into five steps.
Step 1: State the given problem in the mathematical form as illustrated previously.
Step 2: Graph the constraints, by temporarily ignoring inequality sign and decide about the
area of feasible solutions according to the inequality sign of the constraints.
Indicate the area of feasible solutions by a shaded area, which forms a convex
polyhedron.
Step 3: Determine the coordinate so the extreme points of the feasible solution space.
Step 4: Evaluate the values of the objective function at each extreme point.
Step 5: Determine the extreme point to obtain the optimum value of the objective function.
SOME SPECIAL CASES IN LINEAR PROGRAMMING
1. ALTERNATIVE OPTIMAL SOLUTION: So far we have seen that the optimal
solution of any linear programming problem occurs at extreme point of the feasible
region and the solution is unique, i.e. no other solution yields the same value of the
objective function. However, in certain cases a given LP problems may have more than
one optimal solution yielding the same objective function value.
There are two conditions that should satisfy in order that an alternative optimal solution
exists.
 The given objective function is parallel to a constraint that forms the boundary of
the feasible solutions region. In other words, the slope of the objective function is
same as that of the constraint forming the objective function is same as that of the
constraint forming the boundary of the feasible solutions region.
 The constraint should form a boundary on the feasible region in the direction of
optimal movement of the objective function. In other words, the constraint should
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 15
be an active constraint. Graphically an active constraint is one that passes through
one of the extreme points of the feasible solution space.
2. AN UNBOUNDED SOLUTION: When the value of decision variables in linear
programming is permitted to increase infinitely without violation the feasibility
condition, then the solutions said to be unbounded. The objective function value can be
increased infinitely. However, an unbounded feasible region may yield some definite
value of the objective function.
3. AN INFEASIBLE SOLUTION: If it is not possible to find a feasible solution that
satisfies all the constraint, then LP problem, I said to have an infeasible solution /
alternatively inconsistency. Infeasibility depends mainly on the constraints and has
nothing to do with the objective function.
LINEAR PROGRAMMING WITH THE HELP OF SIMPLEX
METHOD
STANDARD FORM OF LP
The use of the simplex method to solve an LP problem requires that the problem be
converted into its standard form. The standard form of the LP problem should have the
following characteristics:
 All the constraints should be expressed as equations by adding slack operation
research surplus and artificial variables.
 The right hand side of each constraint should be made nonnegative, if kit is not;
multiplying both sides of the resulting constraint by –1 should do this.
 The objective function should be of the maximization is expressed as:
For your ready reference the standard for of the LP problem is expressed as,
Optimize (Max/Min) Z = C
1
X
1
+ C
2
X
2
+….+ C
n
X
n
+ 0S
1
+….+ 0S
m
Subject to the linear constraints
A
11
X
1
+ A
12
X
2
+ …+ A
1n
X
n
+ S
1
= B
1
A
21
X
1
+ A
22
X
2
+ … + A
2n
X
n
+S
2
= B
2


A
m1
X
1
+ A
m2
X
2
+ … + A
m
nX
n
+S
m
= B
m
And X
1
, X
2
,…, X
n
, S
1
, S
2
,…,S
m
0
In matrix notation the standard form is expressed as:
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 16
Optimize Z = CX + 0S
Subject to the linear constraints
AX + S = B and X, S 0
Where C = (C
1
,C
2
,…,C
n
) is the row vector;
X = (X
1
,X
2
,…,X
n
), B = (B
1
,B
2
,…,B
m
)
S = (S
1
,S
2
,…,S
n
) are columns vectors and
A
11
A
12
… A
1n
A
21
A
22
… A
2n
… … …
A
m1
A
m2
A
mn
is the (m x n) matrix of coefficients of variables X
1
,
X
2
,…, X
n
in the constraints.
There types of additional a variables, namely
1. Slack variable (S)
2. Surplus variables (S)
3. Artificial variables (A)
TWO PHASE METHOD
In the first phase of the method the sum of the artificial variables is minimized subject to
the given constraints to get a basic feasible solution of the LP problem. The second phase
minimizes the original objective function starting with the basic feasible solution obtained
at the first phase. Since the solution of the LP problem is completed in two phases, this is
called the twophase method.
STEPS OF ALGORITHM
PHASE 1:
Step 1:
a) If all the constraints in the give LP problem are (<) type, then phase II can be directly
use to solve the problem. Otherwise, a sufficient number of artificial variables are
added to get a basis matrix (identity matrix).
b) If the given LP problem is of minimization, then convert it to the maximization type by
the usual method.
Step 2: Solve the following LP problem by assigning a coefficient of –1 to each artificial
variable and zero to all other variable and zero to all other variables in the objective
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 17
function and with the basic feasible solution X
1
= X
2
= … = X
n
= 0 and A
i
= b
1
; i =
1,2,…,m
Maximization Z* = E (1) A1
Subject to the constraints
Z* = E a
ij
x
j
= A
j
= b
i
; i = 1,2,…,m; and x
j
, A
j
> 0
Apply the simplex algorithm to solve this LP problem. The following three cases may arise
at optimally:
 Max Z* = 0 and at least one artificial variable is present in the basis with position
value. Then no feasible solution exists for the original LP problem.
 Max Z* = 0 and no artificial variable is present in the basis. Then the basis consists of
only decision variables (Xjs) and hence we may move to phase II to obtain an optimal
basic feasible solution to the original problem.
 Max Z* = 0 and at least one artificial variable is present in the basis at zero value.
The feasible solution to the above LP problem is also a feasible solution to the
original LP problem. Now orders to arrive at the basic feasible solution we may
proceed directly to phase II operation research else eliminate the artificial basic
variables and then proceed to phase II.
THE BIG – M METHOD
The big – M method is another method of removing artificial variables from the basis. In
this method, we assign coefficients to artificial variables, under sizable from the objective
function point of view. If objective function Z is to be minimized, then a very large positive
price is assigned to each artificial variable. Similarly, if Z is to be maximized then a very
large negative price is assigned to each of these variables. The penalty will be designated by
= M for a maximization problem and +M for a minimization problem where M>0. The big
– M method for solving an LP problem can be summarized in the following steps:
Step 1: Express the LP problem in the standard form by adding surplus variables and
artificial variables. Assign a zero coefficient to surplus variables and a very large
positive number +M to artificial variables in the objective function.
Step 2: The initial basic feasible solution is obtained by assigning zero value to original
variables.
Step 3: Calculate the values of C
j
– Z
j
in last row of the simplex table and examine these
values:
 If all C
j
– Z
j
0 then the current basic feasible solution is optimal.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 18
 If for a column, k, C
k
– Z
k
is most negative and all entries in this column are negative,
then the problem has an unbounded optimal solution.
 If one or more C
j
– Z
j
< 0 then select the variables to enter into the basic with the
largest negative C
j
– Z
j
value. That is, C
k
– Z
k
= Min(C
j
– Z
j
:C
j
– Z
j
< 0)
The column to enter is called key or pivot column.
Step 4: Determine the key row and key element in the same manner as discussed in the
simplex algorithm of the maximization case.
LINEAR PROGRAMMING FORMULATION OF DUAL PROBLEM
The term „dual‟ in a general sense implies two/double. The concept of duality is very useful
in mathematics, physics, and statistics and engineering.
In context of linear programming duality implies that each linear programming problem can
be analyzed in two different ways but having equality solutions. Each LP problem stated in
its original form has associated with another linearprogramming problem, which is unique
based on the same data. In general it is immaterial which of the two problems is called
primal / dual since the dual is primal.
SYMMETRICAL FORM
Suppose the primal LP problem is given in the form,
Maximize Z
x
= C
1
X
1
+ C
2
X
2
+…+ C
n
X
n
Subject to the constraints
A
11
X
1
+ A
12
X
2
+…+ A
1n
X
n
+ S
1
= B
1
A
21
X
1
+ A
22
X
2
+…+ A
2n
X
n
+ S
2
= B
2

A
m1
X
1
+ A
m2
X
2
+…+ A
mn
X
n
+S
m
= B
m
And X
1
,X
2
,… ,X
n
0
Then corresponding dual LP problem is defined as:
Minimize Z
y
= B
1
Y
1
+ B
2
Y
2
+…+ B
m
Y
m
Subject to the constraints
A
11
Y
1
+ A
12
Y
2
+…+A
1n
Y
n
+ S
1
= B
1
A
21
Y
1
+ A
22
Y
2
+…+ A
2n
Y
n
+S
2
= B
2

A
m1
Y
1
+ A
m2
Y
2
+…+ A
mn
Y
n
+ S
m
= B
m
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 19
And Y
1
, Y
2
,…,Y
n
0
Therefore above pair of LP problem can be expressed in the general LP model form as:
PRIMAL DUAL
n
Max Z
x
= E C
j
X
j
j=1
Subject to
n
E A
ji
X
j
B
i
; i = 1,2,…,m
j=1
And X
j
0; j = 1,2,..,m
n
Min Z
y
= E B
i
Y
i
j=1
Subject to
n
E A
ji
Y
i
C
j
; j = 1,2,…,n
j=1
And Y
i
0; i = 1,2,..,m
A summary of the general relationship between primal and dual LP problem is given in
table below:
IF PRIMAL THEN DUAL
Objective is to maximize Objective is to minimize
Variable X
j
Constraint j
Constraint i Variable Yi
Variables X
j
unrestricted in sing Constraint j is = type
Constraint i is = type sign Variable Yi is restricted in sign
type constraints type constraints
0 variables inequality constraints
inequality constraints 0 variable
ECONOMIC INTERPRETATION
In the primal LP model we may define each parameter as follows:
Z = return
X
j
= units of variable j
C
j
= return per unit of variable X
j
A
ij
= requirement of resource I by per unit of variable j
B
i
= units of resource i
The new parameters introduced in the dual problem are Z
y
and Y
i
. Since Z
x
= Z
y
so Z
y
is
also in terms of “return”. The interpretation associated with the dual variables Y
i
, i =
1,2,…,m is discussed below:
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 20
Rewriting the primal LP problem in terms of new terminology
PRIMAL:
n
Maximize Z
x
= E { return / units of variable X
j
}(units of variable X
j
)
j=1
Subject to the constraints
E{ units of resource I / units of variable X
j
}(units of variable) unit of resource i;
i = 1,2,…,m and X
j
0
DUAL:
n
Minimize Z
y
= E (units of resource i) Y
i
, subject to the constraints
j=1
n
E { units of resource i/ units of variable X
j
} Y
i
{return / units of variables X
j
}
j=1
i = 1,2,…,n and Y
i
0
TRANSPORTATION MODEL
One important application of linear programming has been in the area of physical
distribution of resources, from one place to another to meet a specific set of requirements. It
is easy to express a transportation problem mathematically in terms of an LP model, which
can be solved by simplex method. Since transportation problem involves a large number of
variables and constraints it takes a very long time to solve it by the simplex ways, involving
transportation algorithm, namely the stepping stone method, MODI method have been
involved for this purpose.
The transportation algorithm discussed in this chapter is applied to minimize the total cost
of transporting a homogeneous commodity from one plane to another. However, it can also
be applied to the minimization of some total value / utility, for example, financial resources
are distributed in such a way that the profitable return is maximized.
TRANSPORTATION PROBLEM
The transportation problem applied to situation in which a single product is to be
transported form several sources to several sinks. In general, let there be m sources, S
1
,
S
2
,…,S
m
having Ai units of supplies or capacity respectively table transported among n
destination D
1
,D
2
,…,D
n
with B
j
units of requirement respectively. Let C
ij
be the cost of
shipping one unit of commodity from sources i to destination j for each route. If X
ij
represents the units shipped per route from source i to destination j, the problem is to
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 21
determine the transportation schedule so as to minimize the total transportation cost
satisfying supply and demand conditions. Mathematically the problem can be stated as:
m n
Minimize (total cost) = E E C
ij
X
ij
i=1 j=1
Subject to the constraints
n
E X
ij
= A
i
, i = 1,2,…,m
j=1
m
E X
ij
= B
j
, j = 1,2,…,n
i=1
And X
ij
0 for all i and j.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 22
EXPERIMENT.NO: 3
STUDY & IMPORTANCE OF SIMPLEX METHOD IN
LINEAR PROGRAMMING
GRAPHICAL SOLUTION
INTODUCTION
An optimal as well as feasible solution to an LP problem is obtained by choosing among
several values of decision variables x
1
, x
2
, …., x
n
. The one set of values that satisfies the
given set of constraints simultaneously and also provide the optimal values of a given
objective function.
For LP problems that have only two variables it is possible that the center set of feasible
solution can be displayed graphically by plotting linear constraints to locate a best optimal
solution. A technique used to identify the optimal solution is called the graphical solution
technique for an LP problem with two variables.
DEFINITION
Basic Solution:
In a system of m equations and n unknowns where n>m,
AX= b and XT belongs to R
n
When A is a (m x n) matrix of rank m.
Let B be any x sub matrix formed by linearly independent columns of A, then the solution
is obtained by setting nm variables, not associated with the columns of B, equal to zeroes
and solving the resultant system is called a Basic Solution to the given system of equations.
The nm variables whose values did not appear in the solution are called nonbasic
variables and the remaining m variables are called basic variables.
Solution:
Solution values of decision variables xj(j=1,2,…n) which satisfy the constraint of general
LP model is called the solution of the LP model.
Feasible Solution:
Solution values of decision variables xj (j=1,2,…n) which satisfy the constraints and non
negativity condition of general LP model are said to constitute the feasible solution to that
LP model.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 23
Basic Feasible Solution:
A feasible solution to an LP problem, which is also the basic solution, is called the Basic
Feasible Solution. That is, all basic variables assume nonnegative values. Basic Feasible
Solution are of two types:
 Degenerate: A basic feasible solution is called Degenerate if at least one basic
variable possesses zero value.
 NonDegenerate: A basic feasible solution is called Nondegenerate if all m basic
variables are positive and nonzero.
Optimum Basic Feasible Solution:
A basic feasible solution which optimizes the objective function of the given LP models
called an Optimum Basic Feasible Solution.
Unbounded Solution:
A solution which can be increased operation research decreased the value of objective
function of LP problem indefinitely is called Unbounded Solution.
This solution method for an LP problem is divided into five steps:
Step1: State the given problem in the mathematical form.
Step2: Graph the constraints, by temporarily ignoring the inequality sign and decide about
the area of feasible solution according to the inequality sign of constraints. Indicate
the area of feasible solutions by a shaded area which forms a convex polyhedron.
Step3: Determine the coordinate so the extreme points of the feasible solution space.
Step4: Evaluate the value of the objective function at each extreme point.
Step5: Determine the extreme points to obtain the optimum value of the objective function.
SIMPLEX METOD:
STANDARD FORM OF LP:
The use of the simplex method to solve an LP problem requires that the problem be
converted into its standard form. The standard of the LP problem should have the following
characteristic.
1. All the constraints should be expressed as equations by adding slack, operation
research surplus and artificial variables.
2. The right hand side of each constraint should be made nonnegative, if it is not, this
should be done by multiplying both sides of the resulting constraints by1.
3. The onjective function should be of the maximization is expressed as:
Optimize (max/min) Z = C
1
X
1
+ C
2
X
2
+…..+ C
n
X
n
+ 0S
1
+….+ 0S
m
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 24
Subject to linear constraints,
A
11
X
1
+ A
12
X
2
+…..+ A
1n
Xn
+ S
1
= B
1
A
21
X
1
+ A
22
X
2
+…..+ A
2n
Xn
+S
2
= B
2

A
m1
X
1
+ A
m2
X
2
+…..+ A
mn
X
n
+ S
m
= Bm
And X
1
,X
2
,……..,X
n
,S
1
,S
2
0
In matrix notation the standard form is expressed as:
Optimize Z = CX+0S
Subject to linear constraints
AX+S = b and X,S 0
Where C=[C
1
,C
2
,…..C
n
]is the row vector.
X=[X
1
,X
2
,….X
n
], B=[B
1
,B
2
,…..B
m
]T
S= [S1, S2, …. Sn] are column vectors.
A
11
A
12
A
1n
A
21
A22 A2n
  
A
m1
A
m2
A
mn
Is the matrix of coefficients of variables X
1
, X
2
, …, X
n
in the constraints. Three types of
additional variables namely
1. Slack variables [S]
2. Surplus variables [S]
3. artificial variables [A]
STEPS IN SIMPLEX METHOD:
Step1: Formulation of LPP
Step2: Convert constraints into equality form.
Step3: Construct the starting simplex table.
Step4: Test optimality by analysis.
Step5: Find “incoming” and “outgoing” variables and rewrite the table as per given set of
rules.
Step6: Repeat step 4 onwards again till optimum basic feasible solution is obtained.
THE BIG  M METHOD
Big M method is another method of removing artificial variables from the basis. In this
method, we assign coefficients to artificial variables, undesirable from the objective
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 25
function point of view, if objective function Z is to be minimized, then a very large positive
price is assigned to each artificial variable. Similarly, if Z is to be maximized, then a very
large negative price is assigned to each of this variable. The penalty will be designated by
=M for maximization problem and +M for a minimization problem, where M>0. The big M
method for solving and LP problem can be summarized in the following steps:
1. Express the LP problem in the standard form by adding surplus variables and artificial
variables. Assign a zero coefficients to surplus variables and a very large positive
number +M andM to artificial variables in the objective function.
2. The initial basic feasible solution is obtained by assigning zero value to original
variables.
3. Calculate the values of CjZj in fast row of the simplex table and examine these values.
a) [a] If all C
j
Z
j
>0, then the current basic feasible solution optimal.
b) [b] If for a column, k, C
k
Z
k
is most negative and all entries in this column are
negative then the problem has an Unbounded Optimal Solution.
c) [c] If one / more C
j

Zj
<0 then select the variable to enter into the basic with the
largest negative C
j
Z
j
value. That is,
d) C
k
Z
k
= Min{C
j
Z
j
; C
j
Z
j
<0}
e) The column to be entered is called the key / pivot column.
4. Determine the key row and key element in the same manner as discussed in the
simplex procedure of the maximization case.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 26
EXPERIMENT.NO: 4
STUDY OF DUAL PROBLEM, ITS INTERPRETATION &
SENSITIVITY ANALYSIS
DEFINATION
The dual is an auxiliary LP problem defined directly and systematically from the original of
primal LP model. In most LP treatments, the dual is defined for various forms of the primal
depending on the types of the constraints, the signs of the variables and the sense of
optimization.
The general standard for the primal is defined as
n
Maximize or Minimize Z = ¯ C
j
X
j
j =1
Subject to
n
¯ A
ij
X
j
= B
i
, i = 1,2,……,m
j=1
X
j
0, j = 1,2,….,n
CONSTRUCTION OF DUAL PROBLEM
Following steps are as under:
 If the primal problem has its objective function of maximization its all constraints must
have type of inequality. Like that, the primal problem having its objective function to
be minimized should have all its constraints type of inequality.
 The nature of optimization is changed, i.e. if objective of primal problem is to be
maximized, the objective of dual problem is to be minimized and vice versa.
 Right hand constants of primal problem become objective function coefficients of dual
problem. This demands that for each constraint in primal problem there is a separate
dual variable. And vice versa objective function coefficients of primal problem become
right hand constants of dual problem giving a constraint of each variable in the primal
problem. For example, if primal problem ha m number of constraints and n number of
variables then its dual problems has n number of constraints and m number of variables,
i.e. if primal is of order m x n the dual is of order n x m.
 Inequalities of constraints are reversed.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 27
 Row coefficients of the matrix of primal problem become column coefficients of the
matrix of dual problem in the same order i.e. 1
st
row of primal become 1
st
column of
dual.
 Primal as well as dual variables are nonnegative.
INTERPRETATION AND PROPERTIES OF DUAL
Following are a few important interpretation and properties of dual problem:
 The objective function value of any feasible solution to the primal will be less than or
equal to the objective function value of every feasible solution to the dual i.e. ZZ
min
.
 If either the primal or the dual has an unbounded solution, the solution to the other
problem is infeasible.
 If both primal and dual problems have feasible solution then both also have optimum
solution and Z
max
= Z
min
.
 If the primal has a feasible solution but the dual has no feasible solutions, then the
primal does not have a finite optimum solution. Likewise, if the dual has a optimum
solution.
The above conclusions are derived as the statements of dual theorems.
Some of the important of PrimaDual are:
 The index row coefficient under the slack variables in primal optimal solution is the
optimal values of the dual variables.
 The index row coefficient under variable in the primal optimal gives the difference
between the left and right hand sides of the associated optimal dual corresponding dual
constraints.
DUAL SIMPLEX METHOD
In simplex method we start with initial feasible solution and through stages of iteration
along simplex algorithm we improve the solution till optimal solution is one, in terms of
algorithm for maximization in which index row coefficients are nonnegative i.e. dual
variables are feasible. Now, from our understanding of PrimalDual relations, we know that
index row coefficients are the values of dual variables and hence it suggests dual feasible
solution. There are situations where a primal solution may be infeasible, but corresponding
variables indicate that dual is feasible. Thrust solution is primal infeasible but optimum. In
a dual simplex method, initial solution is infeasible but optimum, and through iteration, it
reaches feasibility at which stage it also reaches true optimum.
To summaries the procedure:
 The solution associated with a basis is optimal if all index row coefficients are 0.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 28
 The basic variable having the largest negative value is the departing variable and its row
is the key row.
SENSITIVITY ANALYSIS
Management‟s interest in optimization is not confined to obtaining optimum solutions for
the given set of data. In fact, management for its decisionmaking in changing situations has
interest in analyzing the present optimal solutions for projecting its future picture. Such an
investigative analysis is called sensitivity analysis. As this analysis is based on the optimal
solution presently obtained, it is called postoptimally analysis.
Sensitivity analysis will seek answers to the following questions: (these questions are put in
relation to a productmix problem for profit maximization with the given set of constraints
due to resources. However, the analysis will draw general conclusion.)
…. if objective coefficient of a particular basic variable in optimal solution changes, does
the present solution remain optimal? If objective coefficient of a particular nonbasic
variable in optimal solution changes, does the present solution remains optimal?
…. when will happen to the present optimal solution if a certain resources is augmented or
curtailed?
…. what will happen to the present optimal solution, if a product is added or dropped?
CHANGE IN THE OBJECTIVE FUNCTION COEFICIENT
Variation in the objective function coefficient may belong to a basic variable or to a non
basic variable. We shall take the two cases separately. Consider a change in the objective
function coefficient of the nonbasic variable in the optimal solution. Any change in the
objective function coefficient of the nonbasic variable will affect only its index row
coefficient and not others. Here X
3
is nonbasic variable. Any decrease in its coefficient
will not affect the present solution.
Now consider a change in the objective coefficient of the basic variable in the optimal
solution. Here, it affects the index now coefficients of all the variable. Hence, as soon as the
index row coefficient of basic variable becomes positive, it leaves the solution, and that of
nonbasic variable becomes negative, it qualifies for entry into the solution. In either case,
the present optimal solution changes.
CHANGES IN RIGHT HAND SIDE CONSTANTS
The right hand side constants denote present level of availability of resources. When this is
increased or decreased, it will have the effect on objective function. Besides, it may also
change the basic variables in the solution.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 29
EXPERIMENT.NO: 5
SPECIAL CASE OF SIMPLEX METHOD
Special applications of simplex methods are:
1 1) ) Transportation technique.
2 2) ) Assignment model.
1. TRANSPORTATION TECHNIQUE
This is a special case of linear programming. The aim of this technique is that within a
given time period, a single homogenous specified commodity to be transported form „m‟
number of sink on destination available units so that, total cost of transportation is
minimum. Here size and location of plant are decided. The technique helps to determining
optimal solution.
Consider „m‟ number of source and „n‟ number of destination, also consider i
th
source of
availability of A
i
number of units of a product and n number of destinations as j
th
destination requires B
j
number of units for the same product. Let the cost associated with
transportation is C
ij
and decision variable is X
ij
. Let mathematical statement of problem,
m n
Minimize Z = E E C
ij
X
ij
………………… (1)
i=1 j=1
equation (1) is called objective function.
m
Subject to E X
ij
= B
j
i=1
n
& E X
ij
= A
i
j=1
These are called constraints of the function.
For a balanced problem, suppose there are 4 sources and 3 destinations, minimize,
Z = C
11
X
11
+ C
12
X
12
+ C
13
X
13
+ C
21
X
21
+ C
22
X
22
+ C
23
X
23
+ C
31
X
31
+ C
32
X
32
+ C
33
X
33
+
%S C
41
X
41
+ C
42
X
42
+ C
43
X
43
Subject to X
11
+ X
12
+ X
13
= A
1
X
21
+ X
22
+ X
23
= A
2
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 30
X
31
+ X
32
+ X
33
= A
3
X
41
+ X
42
+ X
43
= A
4
X
11
+ X
12
+ X
13
+ X
14
= B
1
X
21
+ X
22
+ X
23
+ X
24
= B
2
X
31
+ X
32
+ X
33
+ X
34
= B
3
For balance problem,
A
1
+ A
2
+ A
3
+ A
4
= B
1
+ B
2
+ B
3
It can be represented in matrix as under:
Origin Destination Availability
1 2 3
1 X
11
C
11
X
12
C
12
X
13
C
13
A
1
2 X
21
C
21
X
22
C
22
X
23
C
23
A
2
3 X
31
C
31
X
32
C
32
X
33
C
33
A
1
4 X
41
C
41
X
42
C
42
X
43
C
43
A
4
Requirement B
1
B
2
B
3
TECHNIQUES OF FEASIBLE SOLUTION
1 1) ) NorthWest Corner Method: It may be recalled that feasible solution is the one,
which satisfies the given set of constraints. Here first, the northwest corner is
selected and the availability of with source utilized and the row is eliminated. Again
after A
th
row is utilized next new northwest corner is selected and its availability
and requirement is satisfied, and this is done until all the requirement is satisfied, and
as the allocation done, cost of respective allocation is known and total cost of
transportation is calculated.
2 2) ) Vogel‟s Approximation Method: This method is an improved algorithm to obtain
better feasible solution. The process is to evaluate penalty number for each row and
column. It is given, as the difference between the lowest cost coefficients is each
row and column. The penalty number physically amounts to cost penalty associated
with by not accepting that row or column for assigning an allocation to its lowest
cost cell. Select largest of these penalty numbers and determine the row or column
to which an allocation has to be make. Thus row or column selected against largest
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 31
penalty cost will be assigned more possible units and its lowest cost. The process
will be repeated for each allocation to be made after eliminating row or column.
3 3) ) Inspection Method: If size of problem is small, inspection method is used. Here
dummy row or column will be allocated. Select lowest cost cell and assign more
feasible units to it. Repeat this until A
i
< B
j
or A
i
> B
j
and until availability and
requirement become equal.
DIFFERENT METHODS FOR OPTIMAL SOLUTION
Before any other method is used for optimally test a check has to be made over number of
allocations assigned in the solution; which must be equal to (m + n) having nondegenerate
solution.
1 1) ) Stepping Stone Method: Any empty cell in the feasible solution gives nonbasic
variables. Each vacant on unfilled cell has X
ij
= 0. We never know if any one or
more of them may become basic variables there by reducing the transportation cost
or improving solution further. Each empty cell must be evaluated. Opportunity cost
of an empty cell is the net increase or decrease in the total transportation cost if a unit
is assigned to it. As sum conditions are to be honored, one unit from the present
allocation in the row as well as column of this empty cell has to be deducted and
some unit has to be added to the present allocation in the column as well as row at
which deduction took place and whole orthogonal path is completed.
2 2) ) Modi Method: This method is also called modified method. In this method, we
generate a column under beading of U
i
and V
j
of respectively right of the solution
and row of at the bottom of the solution. Here number of variables „m‟ and „n‟ at U
i
and V
j
are determined. Select any one of the U
i
variables and assign any arbitrary
value (normally zero) to it and determine rest of them the relationship.
U
i
+ V
j
= C
ij
After all the values U
i
and V
j
are determined, opportunity costs of empty cells are
evaluated using the relationship.
D
ij
= C
ij
– (U
i
+ V
j
)
2. ASSIGNMENT MODEL
Consider that there are „n‟ persons having varying degrees of proficiency to accomplish „n‟
different tasks in an organization. These „n‟ tasks are to be assigned to each one of „n‟
persons in such a manner that the effective performance of the organization as a whole is
maximized. This is called assignment problem. The objective function to be optimized
may be effective performance (more) total cost of finishing „n‟ jobs or total time of
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 32
completing „n‟ jobs or any other suitable criteria for objective function depending upon the
assignment problem.
MATHEMATICAL STATEMENT OF ASSIGNMENT MODEL
With „n‟ tasks and „n‟ agents, let C
ij
be the cost of its agent, performing j
th
task.
n n
Min E E C
ij
X
ij
i=1 j=1
n
Sub to E X
ij
= 1
j=1
The statement appears similar to that of transportation techniques and X
ij
takes value either
0 or 1.
X
ij
= 1 (if task is assigned to agent 1)
X
ij
= 1 (if task is not assigned to agent 1)
TECHNIQUE FOR ASSIGNMENT MODEL
Hungarian method follows steps as under:
 Add as many dummy rows on dummy column to make matrix square. This
corresponds to balancing of transportation problem. In a problem of assignment, each
of the origin has availability of a unit and each of destinations has requirement of a
unit.
 Take the first row and subtract its smallest cost element from all the elements. Repeat
this for all the rows.
 Similar do for column. In this there will be at least one zero element in each row and
column.
 Starting with the row that has only one zero give assignment there and cross mark
there.
 If there are more than one zero, start with column and find the solution.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 33
EXPERIMENT.NO: 6
PROJECT EVALUATION AND REVIEW TECHNIQUES &
CRITICAL PATH METHOD (PERT/CPM)
In real world time can‟t be known with total certainty. Particularly there are several
projects involving research, innovation, and design, development where there is no previous
experience or record of handling such activities. For such activities forecasting of time
estimates will involve uncertainty. This necessitates probabilistic approach and is popularly
known as PERT. This is its one and only feature which distinguishes it from CPM.
Time t for an activity is assumed top have its probability density p(t) following Beta
distribution in PERT; thus it is
P(t) = k (tt
o
)
o
(t
p
t)

where t
o
= the optimistic time,
t
p
= the pessimistic time,
o. = exponents values of which are decided on the basis of the mode and mean ab
of the distribution
Optimistic time means estimated duration of activity where all factors concerning the
activity operate favorably. Although it is easier said than obtained, optimistic time
presumes that everything goes all right for the activity right from the word „getsetgo‟.
Pessimistic time means estimated duration of activity where in all factors concerning the
activity operate unfavorably. Both of these estimates are qualified guesses and are
additionally consider the concept of most likely time, t
m
that is the mode of the distribution.
Figure 1 shows beta distribution followed by activity times, where T
E
is the average time in
which activity is expected to be
completed. The value of T
E
depends on how close the values
of t
o
and t
p
are relative to t
m
.
Expected time of an activity
T
E
= ( t
o
+ 4t
m
+ t
p
) / 6
Since the actual time of an activity
is likely to vary form mean value
(T
E
), we need the variance of the
activity time distribution. Being
unimodal distribution, it spread
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 34
may be approximated to six times the standard deviation value. Thus,
6 = t
p
 t
o
Having three time estimates for all the activities in a project, we can compute average time
and variance for each activity. Considering average time as actual time, the critical path is
found. Project time will then be the sum of all the activity times in the critical path. But
the activity time is random variable, sum of all such times is also a random variable, and we
must give average time of the project and its variance. The variance of the project duration
is the sum of all the variances of the activities in the critical path.
PERT & CPM
Attempt is often made to bring out difference between PERT and CPM, besides that of
timeestimates being probabilistic in PERT and deterministic in CPM. In fact, all other
aspects are common. Historically PERT came up from different problem area. This was in
relation to plan and accelerate development of the Polaris ballistic missile.
When activity durations are uncertain, attention is principally focused on vents of activity.
This is because events are identified. For example, to develop a process for synthetic
cryogenic rubber. Here terminal pointevent is known, though activity duration is not
known with certainty. It is in this context of more familiarity of events involved and less
certainty of activities. The PERT is called eventoriented. In fact, the difference is not
organic but that of focus.
EXPECTED DURATION OF CPMPERT
By calculating expected activity time for each activity from three times estimates, PERT
network is prepared and critical path is found on that. As each activity time has standard
deviation and variance, the project duration obtained is also expected duration having
standard deviation and variance. The critical path length is the summation of activities on
the path. As the variance of a sum of independent activities is equal to the sum of their
residual variance, the variance of critical path duration can be calculated,
V
T
= variance of critical path = variances of critical activities
V
T
= V
t1
+ V
t2
+ V
t3
V
T
= variance of any activity in PERT = ()
2
Where is the standard deviation of each activity time
E = (t
p
– t
o
) / 6
S
T
= standard deviation of the project = uV
T
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 35
Variance V
T
standard deviation S
T
of the project gives the measure of uncertainty of the
expected duration of the project T
E
.
CRASHING
Being more practical on scheduling problems, know if the activities involved in the project
can be reduced in duration by providing extra resources. These added resources will
increase the cost of completing the activity, but other advantage out weigh this increased
cost, the activity should be expedited, or crashed. The duration of the project will be
reduced, only if duration of one or more critical activities is reduced. If there are activities
using similar resources and if these activities have slack, they can be carried out at their
normal or most efficient pace, with allocation of less resource, so that part of these can be
diverted to the activity on the critical path without additional cost. In general, this is not
possible for all activities and thus we use the concept of criticality to apply extra resources
to project activities selectively, so that maximum reduction of project time is achieved with
the least additional cost. The increased cost of an activity and is stated in terms of
additional expenditure required for reduction in activity duration of one unit time period.
The total cost of a project is not merely the sum of all direct costs of all the activities in the
project, but other indirect expenses, which depend on the duration of project, must also be
included.
Thus, project schedules influence two types of costs – direct cost of all activities in the
project, which increases when activities are expedited and indirect cost which decreases
when the project duration is reduced.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 36
Figure 2a gives the relation of direct cost of the activity versus activity duration, presuming
linear relationship. On the other hand, relation of indirect costs with respect to project
duration assumed linear is given in figure 2b.
If activity „A‟ has the direct cost defined by the slope o
A
any reduction of the duration of
the activity by time T
A
will involve additional cost of (T
A
) * o
A
. As against that it will
reduce the indirect cost of the project by (T
A
) * . Thus, crashing of activity A is
beneficial only if
(T
A
) *  > (T
A
) * o
A
i.e.,  > o
A
It, therefore, suggests that it is possible to have an optimum schedule – the lower – cost
schedule, which strikes a balance between direct costs and indirect costs.
For simplicity, time cost relationships is approximated as linear one. For better
understanding, activity timecost relationship is further elaborated. Initially, the time
estimate is so set that the resources required are used in the most efficient way to
accomplish the task at minimum cost. These are known as normal time and normal cost.
Suppose there is an activity, which requires three men for its completion in normal time. In
order to expedite it, every time one more man may be added and reduction in time and
increase in cost may be noted and plotted on graph. Every time the slope will be steeper.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 37
Finally it reaches to the point, where increase in number of men does not reduce the time
further, thus this is the shortest time in which this activity can be completed. This is called
crash time and cost against it as crash cost. By joining the crash and normal points, the line
is accepted as entire relationship. Beyond normal time, if the activity is delayed, the cost
may remain constant for small duration but will increase if the activity is excessively
delayed. Figure 3 shows the activity time cost relationship.
There is no reason to believe that every contractor should have some normal and crash
points for a selected activity. Thee tow points will depend on experience of the ken,
available equipments, teamwork, administrative ability that differ from organization to
organization.
The CPM model finds the least cost
schedule after verifying the possible
crashing of the project. To begin
with, a preliminary schedule is
generated with each activity at its
normal point. This is the maximum
length schedule. Then activities on
critical part are marked as critical
activities. For each of these critical
activities, direct cost versus reduction
in activity duration relation
established. At the same, relation
between project duration versus
indirect cost of project is also known.
If for one or more activities increase in
direct cost is less than the saving in the
indirect cost, then a less expensive schedule can be obtained. While considering critical
activities, the activity that has the highest potential for affecting reduction in duration are
the lowest additional cost is chosen first. The improvements are made in steps. New
schedules are generated as long as activities can be crashed with the net reduction in total
project costs.
NETWORK MODELS
Example: Routine problem on telecom networks.
At a given time some links are available, from the available ones, select the
cheapest/fastest/non reliable/shortest path from a given origin to a destination.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 38
REPRESENTATION
Label the modes 1,2,…..,m.
Label the links 1,2,…..,m.
DISCRETE FORMULATION
For a minimally connected network, n = m – 1.
Each link may change attributes [cost, capacity, etc.].
Each node may have same attributes [demand, capacity, etc.].
Put 1 for end mode.
Put –1 for beginning mode.
A useful concept is to define flows on links of a network.
Let X
j
be the row on link j, if C
j
is the cost of flow on j.
OBJECTIVE
EX
j
= 1 j leaving
EC
j
= 1 j entering
General minimum cost flow problem (also called optimum distribution problem)
Min EC
j
X
j
Subject to EC
ij
X
j
= B
i
for each i j = arcs, i = nodes.
Is X
j
C
ij
The shortest path problem is a special case of this with
Bi = 1 for the origin
Bi = 1 for the destination
Bi = 0 for all other nodes
C
ij
= cost / time for traveling link j, ij = 0, ij = 1.
The bound constraints are treated separately. This C
p
has integer solution (if the data is
integer). This is because of the constraint matrix.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 39
EXPERIMENT.NO: 7
GAMES THEORY
INTRODUCTION
In business and economics literature, the term „games‟ refers to the general situation of
conflict and competition in which two or more competitions are engages din decision
making activities in anticipation of certain outcomes over time. The competitors are called
players. A player may be an individual, a group of individuals, or an organization. A few
examples of the competitive and conflict situations where techniques of theory of games
may be used to resolve them are; two or more candidates contesting and election with the
objective of winning with more votes; advertising and other marketing campaigns between
competing business firm; contractors filling bid to win business contract etc theoretically
theory of games provides mathematical models that can be quite useful in explaining
interactive decisionmaking concept. But as a practical tool, theory of games is limited in
scope. However, it is one type of decisionmaking technique where on competitor‟s choice
of course of actions does determined after taking in to account all possible alternative
courses of action of another competitor engaged in the competition.
The models in the theory of games can be classified depending upon the following factors:
1. NUMBER OF PLAYERS: If a game involving only two players, then it is called a
two – person game. However, if the numbers of players are more than two, the game is
referred to as nperson game.
2. SUM OF GAINS & LOSSES: If in a game the gains to one player are exactly equal to
the losses to another player, so that sum of the gains and losses equal zero, then the
game is said to be a zerosum game of it said to be nonzero game.
3. STRATEGY: The strategy for a player is the list of all possible actions that the he will
take for every payoff that might arise. Here it is on necessary that players have definite
information about each other‟s strategies.
The particular strategy by which a player optimizes his gains or losses without knowing the
competitor‟s strategies is called optima strategy. The expected out come per play when
players follow their optimal strategy is called the value of game. Generally, players in a
game employ two types of strategies as given below:
1. PURE STRATEGY: It is decision rule, which is always used by the player to select
the particular course of action. Thus, each player knows in advances of all strategies
out of which he always selects only one particular strategy irrespective of the strategy
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 40
others may choose and the objective of the players is to maximize gains or minimize
losses.
2. MIXED STRATEGY: When both the players are guessing as to which course of
action is to be selected on a particular occasion with some fixed probability, it is mixed
strategy game.
TWO – PERSONS – ZERO – SUM – GAME
A game with only two players, say player A ands B is called Two – persons  Zero – Sum –
Game, if say player A‟s gain is equal to the loss of player B, so that total sum is zero. The
payoffs in terms of gains or losses, when players select their particular strategies, can be
represented in the form of a matrix, called the payoff matrix.
ASSUMPTIONS OF THE GAME
1. Each player has available to him a finite number of possible courses of action. The
list may not be same for each player.
2. Player A attempts to maximize gains and Player B minimize losses.
3. The decisions of both players are made simultaneously and announce simultaneously
so that neither player has an adventives resulting from direct knowledge of the other
player‟s decision.
4. Both the players know not only possible to themselves by also of each other
PURE STRATEGIES
GAMES WITH SADDLE POINT
The selection of an optimal strategy by each player without the knowledge of the
competitor‟s strategy is the basis problem of playing games. Since the payoff for either
player provides all the essential information, therefore only one player‟s payoff table to
required evaluating the decisions. By convention, the payoff table for the player whose
strategies are represented by (say player A) rows is constructed. Now the objective of the
study 1 to know how these players select their respective strategies so that they may
optimize their payoff. Such a decision making criterion is referred to as the minimax –
maximin principle. Such principle I pure strategies game always leads to the best possible
selection of a strategies game always leads to the best possible selection of a strategy for
both players.
RULES TO DETERMING THE SADDLE POINT
The readers are advised to follow the following steps in order to determine the saddle point
in the payoff matrix.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 41
1. Select the minimum element in each row of the payoff matrix and write them under
row minima heading. Then select a largest element among these elements and
enclose it in a rectangle.
2. Select the maximum elements in each column of the payoff matrix and write them
under column maxima heading. Then select a lowest element among these elements
and enclose it in a circle.
3. Find out the elements that are same in the circle as well as rectangle and mark the
position of such elements in the matrix. This element represents the value of the
games and it is called the saddle point.
MIXED STRATEGIES
There are certain payoffs where pure strategies do not lead to saddle point and hence it is
not possible to decide optimal strategies. There is no stable solution of the payoff matrix.
The shifting from one strategy to another in case of unstable solution results in selection of
the strategies by both players on random basis. The two players have to decide their plans
for playing the game and assign probability values X and Y by which they will play their
respective strategies. These plans are referred as mixed strategies.
ARITHMETIC MODEL
This method enables us to decide odd elements of each strategy in mixed strategies. From
odd elements, we can calculate the optimal frequency of each strategy and also the value of
the game.
Consider the following payoff matrix:
B
1
B
2
A
1
A
2
a
11
a
21
a
12
a
22
1
= a
11
– a
12
2
= a
21
– a
22
3
= a
11
– a
21
4
= a
12
– a
22
Odd elements of a given row are obtained by the numerical difference of elements in
another row. Similarly, oddments of a given column are obtained by the numerical
differences of elements in another column.
Optimal frequency of the strategy A
1
, X
1
=
1
/ (
1
+
2
)
Optimal frequency of the strategy A
2
, X
2
=
2
/ (
1
+
2
)
Optimal frequency of the strategy B
1
, Y
1
=
3
/ (
3
+
4
)
Optimal frequency of the strategy B
2
, Y
2
=
4
/ (
3
+
4
)
If total of oddments of rows is equal to the total of oddments of columns, the value of the
game, V will be
V
b1
= a
11
X
1
+ a
21
X
2
….. value if B plays 1
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 42
V
b2
= a
12
X
1
+ a
22
X
2
….. value if B plays 2
V
a1
= a
11
Y
1
+ a
12
Y
2
….. value if A plays 1
V
a2
= a
21
Y
1
+ a
22
Y
2
….. value if A plays 2
METHOD OF SUB – GAMES
This is applicable to any m x 2 or 2 x n game. The whole game is split into several 2 x 2
subgames, which are then solved by pure strategy, arithmetic method or algebraic method
as applicable. From amongst these subgames, a subgame which gives maximum game
value m x 2 game (where row player has more than 2 strategies) or minimum game value in
2 x n games (where column player has more than 2 strategies) is selected as the solution.
METHOD OF MATRICES
Previously we discussed arithmetic method as applicable 2 x 2 payoff matrix. In fact, it can
also be extended 3 x 3 matrix, as under, which is self – explanatory.
B
1
B
2
B
3
A
1
a
11
a
12
a
13
A
2
a
21
a
22
a
23
A
3
a
31
a
32
a
33
Differences of column 1 & 2 Differences of column 2 & 3 Oddments
1
= a
11
– a
12
4
= a
12
– a
13
o
1
2
= a
21
– a
22
5
= a
22
– a
23
o
2
3
= a
31
– a
32
6
= a
32
– a
33
o
3
Differences of row 1 & 2 Differences of row 2 & 3 Oddments
7
= a
11
– a
21
10
= a
21
– a
31

1
8
= a
12
– a
22
11
= a
22
– a
32

2
9
= a
13
– a
23
12
= a
23
– a
33

3
If sum of row oddments is equal to sum of column oddments, optimum frequencies are:
X
1
= o
1
/ (o
1
+ o
2
+ o
3
); X
2
= o
2
/ (o
1
+ o
2
+ o
3
); X
3
= o
1
/ (o
1
+ o
2
+ o
3
).
Y
1
= 
1
/ (
1
+ 
2
+ 
3
); Y
2
= 
2
/ (
1
+ 
2
+ 
3
); Y
3
= 
3
/ (
1
+ 
2
+ 
3
).
If sum of row oddments is not equal to sum of column oddments, the methods of matrices
fail.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 43
LINEAR PROGRAMMING METHOD FOR GAME PROBLEM
As remarked in graphical method for 2 x n or m x 2 games, the characteristic of linear
relations can be used for solving the games by simplex method of linear programming.
This method is applicable for more general m x n games between two players. It also
comes handy when oddments of rows are not equal to oddments of columns. The game
problem can be solved by simplex algorithm and the value of the game obtained for new
matrix will be reduced by the constant added to get the value of the game for original
matrix. Frequency values of strategies in both matrices remain same.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 44
EXPERIMENT.NO: 8
INVENTORY MANAGEMENT
INTRODUCTION
Inventory is material in stock and idles at a given point of time that will be used to satisfy
some further demand of that material. Inventory decreases as demand process operates and
increases as replenishment process operates. Demand may be random function and perhaps
normally not under control of decisionmaker, while replenishment in monitored by the
decisionmaker. The replenishment is the input process to the inventory system and gives
the rate at which material is added to the inventory system. The source of replenishment
may be either production or purchase of the material. The rate at which material is
withdrawn from the inventory is the output process of the inventory. The output is always
in response to demand. The output rate equals to demand rate unless the policy decision
provides otherwise or in event of inventory level decreasing to zero level, when the
situation is described as stock out or outofstock condition.
The mathematical models of inventory system under different operating conditions involve
mainly twodecision variable i.e. how much to order and when to order, when source of
replenishment is purchase of how much to produce and when to produce, when production
is the source or replenishment. With the associated cost of material and its storing,
inventory system operating cost. It is therefore necessary to appreciate what benefits are
derived from having inventory and then derive the concept of optimum inventory. The
following benefits of inventory control, which resemble with its function, are self
explanatory:
1. Inventory enables in decoupling production system from demand fluctuation. If demand
increases inventory of finished and semifinished goods take care of it. It demands
decreases the inventory of finished and semifinished goods could be built up.
2. Inventory enables to derive advantages of planning optimum lot size in production as a
tradeoff against setting time.
3. Inventory enables the sequencing of different components for optimum machine loading
and setup costs.
4. Inventory enables to have uniform labor force resulting labor force resulting in good
labors relation. In production of the product with seasonal demand.
5. Inventory enables to sustain customer goodwill, as it is possible to honors the orders by
committed dates.
6. In the absence of adequate, inventory a situation my come where “rushorder” has to be
made for which, in general, very high price as to be paid.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 45
All these advantages justify the policy of having large inventory. However, there are
considerations, which warrant such a policy. These considerations are as under:
1. Inventory means locked up capital, which could be used for alternative purposed.
2. Locked up capital involves interest.
3. Inventory being physical goods in likely to deteriorate on storage.
4. It involves the risk of fire ad theft. Alternatively, if involves insurance cost. This
suggests that although inventory, because it has to be only adequate. Thus, excess
inventories are undesirable. This calls for controlling the inventories in most profitable
way.
INVENTORY MODELS
By inventory model, we mean to represent the inventory state, operating condition and costs
involved through mathematical relationships, so as to fine Economic Order Quantity when
source of procurement is purchase or to fine Economic Batch Quantity when source of
procurement is manufacturing.
Model 1:
The model is Deterministic. Single item model i.e. it ignores all uncertainties and deals only
with one item. It is characterized by: (1) demand is constant, at uniform rate and known
with certainty (2) lead time and other system parameters such as costs are constant,
independent of replenishment quantity and are known with certainty (3) replenishment is
infinite and (4) shortage of stock is not permitted. A production system needs R units of a
purchased material over a schedule period of time T. the rate of usage is uniform i.e. R/T. if
all R units are purchased at the beginning of the schedule period T, the inventory position is
as shown in fig, which is the graphical presentation of inventory level v/s time.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 46
T
O
u
a
n
t
i
t
y
i
n
s
t
o
c
k
Time, t
As shortage is not permitted, the replenishment has to commence at b. As replenishment is
instantaneous, it is shown by the vertical line and the next cycle commences.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 47
Thus, for the schedule period T there is average inventory of R/2 units, which involves
holding cost. There is one order cost. Therefore, total associated cost for order R units:
TCR (R) = C
1
(R/2) T + C
2
…..(1)
Instead, if q units are ordered at the beginning and the order of the same quantity is
repeated R/q times, the position will be shown in following figure.
tc
O
u
a
n
t
i
t
y
i
n
s
t
o
c
k
tc tc
Time, t
The total associated cost in schedule period T will then be R/q times cost associated with
each order.
TAC (q) = [C
1
* (q/2) * t
c
] R/q
TAC (q) = C
1
* (R/2) * t
c
+ C
2
(R/q)…………(2)
Comparing equations (1) and (2), it can be seen that by not ordering the full quantity in the
beginning the costcomponent due to holding decreases, as t
c
< T, and costcomponent due
to reordering increases R/q times. So, there has to be certain value of q at which TAC(q) is
min. This value of q is optimum ordersize of Economic Ordering Quantity designated as
qo. Number of purchasecycle in schedule time T = R/q = T/ t
c
.
Therefore equation (2) can be rewritten as:
TAC (q) = C
1
* (R/2) * (Tq/R) + C
2
(R/q)
TAC (q) = C
1
* (T/2) * q + C
2
(R/q)…………..(3)
From above equation TAC(q) depends only on q. Two individuals cost component
(C
1
T/
2
)*q and C
2
R/q, as well as TAC(q) are shown in the following figure.
To determine q
o
we differentiate TAC(q) in equation (3) with respect to q, for maximum
and minimum TAC(q). Thus,
d(TAC)/d(q) = C1T/2 – C2R/q2 = 0
So, q
o
= \2C
2
R / (C
1
/T)
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 48
C R/q
q
C
o
s
t
2
C Tq/2 1
TAC (q)
Second differentiation of TAC is Positive. Hence, the value of q is for minimum TAC, and
corresponding value of q is optimum.
So, q
o
= \2C
2
R / (C
1
/T).…………(4)
This is called optimum lot size, also known as economic lot size. R/T is the rate of demand,
in units per unit time.
Designating it as r = R/T equation (4) becomes,
q
o
= \2C
2
r/C
1
If i is the interest rate on capital per unit time T and if holding cost essentially consists of
purchasesinvestment. Then C
1
= i * P where P is price per unit quantity.
So, q
o
= \2C
2
R / (iPT)
By substituting the value of qo of equation (4) in equation (3) we get minimum total
associated cost during scheduled period T
TAC (q
o
) = [\C
1
C
2
RT/2] + [\C
1
C
2
RT/2] = \2C
1
C
2
RT
Minimum total associated cost per unit time TAC (q) = \2C
1
C
2
r
This shows that for optimum value of qo when TAC minimum the costcomponent due to
holding of inventory is equal to the cost component due to reordering. Thus in for point P
lies on the ordinate from the point of intersection of the graphs of individual cost
components.
Now, t
cc
optimum cycle times is given by,
Tt
cc
= Tq
o
/R = T/R * [\2C
2
R/C
1
T] = \2C
2
T/(C
1
R) = \2C
2
/(C
1
r)
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 49
MODEL 4
Having considered deterministic demand models earlier, we may now consider the cases
where demand is probabilistic, which is quite realistic in every day world.
CHARACTERISTIC:
 Demand is in discrete quantities; demand is probabilistic and is met instantaneously.
 Setup cost is zero. Therefore, relevant cost elements are holding cost and shortage
cost.
Let r = number of discrete units required with probability, p
r
,
l = stock level in discrete units,
C
1
= holding cost per unit quantity per unit time,
C
3
= shortage cost per unit quantity per unit time,
When l > r the inventory cost is due to holding.
When r > l the inventory cost is due to shortage.
1. As demand is probabilistic the probable units in inventory is (l  r) pr and its cost is
C
1
(l  r) pr. The total inventory cost due to holding is,
r=1
E C
1
(l  r)p
r
r=0
As the demand exceeds and becomes l + 1 and above the cost is due to shortage. The
probable units in shortage is (r  l) pr and its cost is C
3
(r  l) pr. Thus, if the demand is to
vary from zero to infinity,
r=1 ∞
TAC (l) = E C
1
(l  r)p
r
+ E C
3
(r  l)p
r
.....(5)
r=0 r=l+1
Now, optimum stock level l
o
can be considered as one in which one more unit well as one
less unit will increase the cost. That is at l
o
.
TAC (l
o
+ 1) > TAC (l
o
) and
TAC (l
o
 1) > TAC (l
o
)
We may rewrite the equation (5) for l+1 stock level
l+2 ∞
TAC (l + 1) = E C
1
(l + 1  r)p
r
+ E C
3
(r  l  1)p
r
.....(6)
0 l+2
l+1 1
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 50
Now E C
1
(l + 1  r)p
r
= E C
1
(l + 1  r)p
r
+ C
1
(l + 1  l  1)p
l+1
0 0
1
= E C
1
(l + 1  r)p
r
0
l+1 1 1
So, E C
1
(l + 1  r)p
r
= E C
1
(l  r)p
r
+ E C
1
(l  r)p
r
.....(7)
0 0 0
∞
Similarly for E C
3
(r  l  1) p
r
l+2
∞
= E C
3
(r  l  1) p
r
 C
3
(l + 1  l  1) p
l+1
r=l+1
∞
= E C
3
(r  l – 1 )p
r
l+1
∞ ∞ ∞
So, E C
3
(r  l  1)p
r
= E C
3
(r  l)p
r
 E C
3
p
r
.....(8)
l+2 l+1 l+1
Substituting equations (7) and (8) in equation (6),
l l ∞ ∞
TAC (l + 1) = E C
1
(l  r)p
r
+ E C
1
p
r
+ E C
3
(r  l)p
r
 E C
3
p
r
0 0 l+1 l+1
So,
l ∞ l ∞
TAC (l + 1) = E C
1
(l  r)p
r
+ E C
3
(r  l)p
r
+ E C
1
(l  r)p
r
 E C
3
p
r
.....(9)
0 l+1 0 l+1
First two terms in equation (9) make TAC (lo) as in equation (5), therefore,
1 ∞
TAC (l + 1) = TAC (l) + E C
1
p
r
 E C
3
p
r
0 l+1
∞ l ∞
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 51
Now, E p
r
= 1 = E p
r
+ E p
r
0 0 l+1
l l
Therefore, TAC (l + 1) = TAC (l) + E C
1
p
r
 C
3
[1  E pr]
0 0
l
So, TAC (l + 1) = TAC (l) + (C
1
+ C
3
) E pr  C
3
0
For l to be optimum, TAC (l
o
+ 1) > TAC (l
o
)
l
o
i.e., (C
1
+ C
3
) E P
r
 C
3
> 0
0
l
o
So, E P
r
> C
3
0 C
1
+ C
3
So, P (r / l
o
) > C
3
.....(10)
C
1
+ C
3
P (r / l
o
) means summation of P (r) for the value of r from zero to l
o
. If stock level is l  1,
l1 ∞
TAC (l  1) = E C
1
(l  1  r) p
r
+ E C
3
(r  l + 1) p
r
.....(11)
0 l
l1
The first term E C
1
(l  1  r) p
r
can be rewritten as
0
l
E C
1
(l  1  r) p
r
+ C
1
p
r
0
l l
E C
1
(l  r) p
r
 E C
1
p
r
+ C
1
p
r
.....(12)
0 0
∞
The second term E C
3
(r  l + 1) p
r
can be rewritten as
l
∞ ∞
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 52
E C
3
(r  l + 1) p
r
= E C
3
(r  l + 1) p
r
+ C
3
p
l
l l+1
∞ ∞
= E C
3
(r  l) p
r
+ E C
3
p
r
+ C
3
p
l
l+1 l+1
∞ l
= E C
3
(r  l) p
r
+ C
3
(1  E p
r
) + C
3
p
l
.....(13)
l+1 0
Substituting equation (12) and (13) in equation (11),
l l ∞ l
TAC (l  1) = E C
1
(l  r) p
r
 E C
1
pr + E C
3
(r  l)p
r
+ C
3
(1  E p
r
) + C
3
p
l
0 0 l+1 0
l ∞ l
TAC (l  1) = E C
1
(l  r) p
r
+ E C
3
(r  l)p
r
 (C1 + C3) E pr + C
3
+ C
1
p
l
+ C
3
p
l
0 l+1 0
So,
l1
TAC (l  1) = TAC(l)  (C
1
+ C
3
) E p
r
+ C
3
0
For l to be optimum,
TAC (l
o
 1) > TAC(l
o
)
l1
So,  (C1 + C3) E pr + C3 > 0
0
l1
So, C
3
> E pr
C
1
+ C
3
0
So, C
3
> p(r / l
o
 1) .....(14)
C
1
+ C
3
Combining equations (10) and (14), for minimum TAC (l
o
) so the optimum stock level
should be such that,
p (r / l
o
 1) < C
3
< p (r / l
o
)
C
1
+ C
3
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 53
p (r / l
o
 1) means summation of p (r) for the value of r from zero to (l
o
 1).
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 54
EXPERIMENT.NO: 9
QUEUING THEORY
INTRODUTION
Most of the basic work on queues or waiting line theory was carried out by A. K. Erlang in
early nineteenth century. He experienced congestion of telephone traffic during his work at
Copenhagen Telephone Company and came out with most of the work for the mathematical
models in queuing theory.
Serviceseekers and serves make a system, which needs consideration. Service may be
anything that occupies a man, facility or location for some period of time so that potential
serviceseekers known as arrivals at any time and require the use of server. If at any time, in
a queue may drive out a potential customer. On the other hand, if more serves are provided
that reasonably necessary so that no queue or very short queue is formed, it costs more
money. Therefore, a tradeoff becomes necessary in designing service facility system. This
needs an insight into the growth of queue in a given system.
INPUT PROCESS:
In order to describe the input processes, a random variable such as time between successive
arrival events is designated. It is assumed that random variable is independent and
identically distributed. Thus knowing the probability distribution, the input process can be
described. Often the arrival of customers is completely random in nature and it is assumed
that the number of customers arriving in any time interval „t‟ follows through Poisson
distribution, with parameter ìt. Here, ì is the average number of arrivals per unit time.
There are some situations where customers arrive at fairly regular time intervals.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 55
One variation of input process is the effect of status of queue on arrival. If a queue is
specified to be finite, a customer arriving at the queue after this is not permitted to join the
queue. Further, even if queue is not finite, a customer, seeing a long queue ahead of him
might not join the queue. This is called balking. It may also happen that customer having
joined the queue earlier might leave it with or without any intention of returning. This is
called reneging. It is pertinent to queue discipline.
So, probability of one arrival = 1  (1ìAT) = ìAT
Let us note that if the density of inter arrival is exponential, the density g(y) of the total
arrival time y for ant n consecutive arrivals is given by,
g(y) = ì(ìy)
n1
e
ny
/ (n1)!; y > 0
and probability of n arrivals in time interval, AT = (ìT)
n
e  ìT/n!
Thus, assumption of exponential distribution of inter arrival times is equivalent to
assuming Poisson distribution for probability distribution of n arrivals in interval T.
Expectation, E (n/T) = ìT
Variance, o
2
(n/T) = ìT
Instead of exponential distribution of inter arrival time, Erlang distribution of order n can be
used. This is,
f(t) = (ì
n
) (ì
nt
) (n1) e  ìnt/(n1)!; t > 0
Expectation, E(t) = 1 / ì
Variance, o
2
(t) = 1/nì
2
(Erlang)
SERVICE TIMES
Assuming that one customer is served by one server at a time, and service time for one
customer is independent of that for other.
Let g(t) = probability density function of time t required to serve one customer. Then mean
time of service =
0
í
∞
e
ut
g(t) dt = 1 / u
where u is mean service rate or number of customers served per unit time. Let us assume
exponential (negative) distribution fro service time. Then,
g (t) = u e
ut
; t > 0
probability that service is not completed in time interval T = e
ut
for small values of T,
it is approximately.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 56
Probability of the event that service is complete in time interval T = 1  uT and the
probability of the event that service is completed in time interval T = uT
QUENUE DISCIPLINE
Queue discipline describes the policy of serving the serviceseekers of a queue. Normally
firstinfirstout popularly known ad FIFO (firstcomefirstserved) is a common queue
discipline. It is generally followed in practice. Some manufacturing concerns consider the
problem of machine breakdowns and repairs as a queuing model. In a group of machines,
breakdown takes place at random time.
ITERARRIVAL TIMES:
It is necessary to specify the probability distribution of inter arrival times. Considering f(t)
as the density function for the time interval t between any two consecutive arrivals, when t
> 0.
Meantime between two consecutive arrivals =
0
í
∞
f(t) t dt
So, 1/ì =
0
í
∞
f(t) t dt
It arrival is assumed completely random; the density function of inter arrival time can be
described by negative exponential distribution.
i.e., f(t) = ìe
ì/t
, t>0.
Now, let A be the event that there is no arrival in the interval (O, T). The probability that
there is no arrival in the interval (O, T) and no arrival in (O, T+T) is given by
P(AB) =
T+AT
í
∞
ì e
ì/t
dt = e
ì(T+AT)
Now, P(B/A) = P(AB)/P(A) = e
ì(T+AT)
/ e
ì/t
= e
ìAt
Probability of no arrival in interval AT = e
ìAt
This depends only on AT.
Expanding e
ìAt
by Taylor‟s series;
Probability of no arrival in interval AT = 1 + (ìAT)/1! + (ìAT)
2
/2! + (ìAT)
3
!
= 1ì(AT) + ì
2
(AT)
2
/2  ì
3
(AT)
3
/6
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 57
Neglecting higher terms of (AT) for very small values of AT, approximately probability of
no arrival AT = 1÷ìAT.
In any small interval of time AT, as it is presumed to be sufficiently small to permit only
one arrival at the most, only two mutually exclusive and collectively exhaustive events can
occur i.e. either there is one arrival or there is no arrival.
P [no arrival in any interval AT ] + P [only one arrival in interval AT] =1
KENDALL‟S NOTATIONS
Kendall‟s notations are to specify queuing models in standardized symbols, which mainly
relate to inter arrival time and service time distribution and number of serves in the system
as
Nature of input Nature of service Number of
Process time distribution servers
For symbols,
M designates Exponential distribution
En designates Erlang distribution of order n.
D designates Deterministic distribution
Thus, the queue model M/M/2 symbolizes Poisson input, exponential service time and two
servers.
It is believed that in order to specify a queuing model correctly one must specify the queue
discipline above. To cover this additional variations Kendall‟s rotations were extended by
three more terms in 1971 in a conference on Standardization of Notation in Queuing
Theory. Now it reads :
Input Service Number of Limit of number Number in Queue
Process process servers in system the source discipline
Last three terms are omitted is they are
(///∞/∞/ FIFO)
QUEUE M/M/1: BALANCE EQUATION METHOD
Let us consider a service system where there is only one server, input is Poisson and
se5vice time distribution is exponential. We want to obtain operating characteristics of the
queue. These characteristics will answer some of the following questions:
Will a queue be formed?
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 58
If so, will it be what is its probability of having n customers?
What will be the mean number of customers in the system?
What will be the queue length?
What will be the waiting time for a customer?
Let inter arrival time density function be ì e
ìAt
and service time density function be ue
ut
.
Let at any time n be the number of customers in the system, including those waiting and
those being served. Let Pn(T) be the probability that there are n customers in system at time
T.
In that case, the probability of having n customer in the system at time T+AT is the union of
four mutually exclusion and exhaustive events as:
 One arrival and no departure in time interval T, T+AT with n1 customers at time T.
 Neither arrival nor departure during time interval T, T+AT with n customers at time T.
 One departure and no arrival during time interval T, T+AT with n+1 customers at time
T.
 One arrival and one departure during time interval T, T+AT with n u
Therefore, the balance equation can be set out as under:
Pn(T+AT) = (Probability of 1 Arrive in AT)(Probability to no departure inAT)P
n1
(T) +
(Probability of no arrival in AT) (Probability of no departure in AT) P
n
(T) +
(Probability of no arrival in AT)(Probability to one departure in AT)P
n+1
(T) +
(Probability of 1 Arrive in AT)(Probability to one departure in AT) P
n
(T)
Making use of relations,
P
n
(T+AT) = (ìAT)(1uAT) P
n1
(T) + (1ìAT)(1uAT) P
n
(T) + (1ìAT) uAT P
n+1
(T)
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 59
Simplifying the terms on the right hand side,
[P
n
(T+AT)  P
n
(T)] / AT = ìP
n1
(T)  ìuATP
n1
(T)  ìP
n
(T)  uP
n
(T) + 2ìuAT P
n
(T) +
uP
n+1
(T)  ìuAT P
n+1
(T)
lim
dP
n
/ dT = AT 0 {P
n
(T+AT)  P
n
(T)} / AT
dP
n
/ dT = ìP
n1
(T) – (ì+u)P
n
(T) + uP
n+1
(T)
This expression is valid for n>0
For n=0, n1 is meaningless, and departure has no significance. Therefore uP
0
(T) is also
meaningless. Let this limiting value of probability distribution be n for equilibrium dP
n
/
dT must be zero, P
n
becomes independent of T. Then equation becomes:
ìP
n1
(T)  (ì+u)P
n
+ uP
n+1
= 0, for n = 1,2,3……..
subscript (T) is dropped for brevity and in equation ,
ì P
0
+ u P
1
= 0, for n = 0
P
1
= {ì/u}P
0
Let ì/u = p. If p>1, means rate of arrival is more than mean rate of service. As a
consequence, the queue continues to build up indefinitely and for this the stationary
probabilities do not exist. Only if ps1, there exist stationary probabilities Pn.
Equation can be rewritten as P
n+1
= {(ì+u)P
n
– P
n1
} / u
Putting n=1, P
2
= (ì+u)P1 / u
Putting p = ì/u, P
0
= pP0, P
2
= p
2
P
0
Likewise P
3
= p
3
P
0
and it can be shown that Pn = p
n
P
0
∞
Now, E P
n
= 1
n=0
P
0
+ P
1
+ P
2
+ P
3
+…+ P
n
= 1
P
0
+ pP
0
+ p
2
P
0
+ p
3
P
0
+ … + p
n
P
0
= 1.
P
0
[1 + p + p
2
+p
3
+ … +p
n
] = 1.
P
0
= 1 / [1 + p + p
2
+p
3
+ … +p
n
] = (1p).
And P
n
= (1p)p
n
,
for n = 0,1,2,…etc.
P
1
/ P
0
= P
2
/ P
1
= P
3
/ P
2
=…= P
n+1
/ P
n
= p
Hence, P
n
is in geometric progression for n = 0,1,2, …
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 60
The mean number of customer in the system, ň, is the summation of products of
probabilities with corresponding values of n, n = 0, 1, 2, 3, …
n=∞
ň = E n P
n
= P
1 + 2P
2
+ 3P
3
+… = pP
0
+ 2p
2
P
0
+…
n=0
ň = P
0
p (1 + 2p +3p
2
+ ...) = P
0
p / (1p)
2
Putting P
o
= (1p);
ň = p / (1p) = ì / (uì)
P
0
is the probability of having no customer in the system. Two events that the server is busy
and that he is not busy are collectively exhaustive events.
P
0
[server is not busy] + P [server is busy] = 1.
P
0
[server is busy] = 1 P
0
.
Thus, the traffic intensity is also the probability of server being busy.
The queue length of mean number of customer in the waiting line n
q
will be given by, the
summation of the product, pos probabilities and (n  1) for n = 1 to infinite as one customer
is being served.
n=∞ n=∞ n=∞
n
q
= E (n1)P
n
= E nP
n
 E P
n
n=1 n=1 n=1
n=∞ n=∞
n
q
= E nP
n
 E P
n
 P
0
n=1 n=1
n
q
= ň  [1  P
0
]
But, ň = p / (1  p) and P
0
= 1  p
n
q
= {p / (1p)}  p = p
2
n
q
= ì / u (uì)
For a customer joining a queue two matters are important. Firstly, how long he will have to
be in the system? Secondly, how long he will to wait till server becomes available to serve
him?
In steady state queue, a customer who joins at average n customer in the system and after
moving along the queue eventually gets served leaves behind the same average n customer.
This means that his duration stay in the system T
s
is such that provides for n arrivals at the
mean rate of ì. Therefore ň = ìT
s
and it is called Little”s formula.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 61
Here, T
s
= Expected number of customer in the system / Mean rate of arrival
= ň / ì = p / ì(1p) = 1 / (uì)
Similarly,
T
q
= Expected waiting time in the system  mean service time
= T
s
– (1/u) = {1/(uì)} – {1/} = ì / u(uì).
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 62
EXPERIMENT.NO: 10
REPLACEMENT THEORY
INTRODUCTION
Any engineering equipment has to give production service. While doing that, it
essentially needs operation cost. Its operational economy is rated through amount of
production or service per unit cost. With usage and years, equipment undergoes wear
and tear, needs progressively more maintenance, goes down in production or service
and requires frequent repairs. This involves additional cost, which in general increases
with years. Therefore, at some stage one intuitively feels, the existing equipment has to
be replaced by new similar or technologically better equipment. This is “Replacement
Problem” and it seeks to answer „when to replace‟ and „with what?‟ The answers to
these questions can be had with “Replacement Theory”.
COST OF “KEEPING IT ON” AND “REPLACING”
If we do not replace the existing equipment, we incur „operation, maintenance and
repairs cost‟ which increases every year.
Let C
m
(n) = annual cost „operation maintenance and repairs‟ in nth year
C = capital investment on the equipment which once incurred is charged as
88permanent cost
If existing equipment is replaced it is discarded either as scrap of as still productively
usable in some application fetching some price on desired either as scrap or as still
productively usable in some application fetching some price on disposal. This is called
salvage value.
The salvage value of the equipment naturally decreases with year the decrease being
very slow in initial years and then very steep, as it tends to be a scrap.
Let S
n
= salvage value of the equipment at the end of nth year
Let us presume for simplicity that money value does not change with year. That is
interest and inflation is insignificant.
The cumulative total cost of „keeping it on‟ upto nth years
= Capital investment value at the end on nth year + cumulative cost of
88operation, maintenance and repairs for nth year
n
TCK(n) = C + EC
m
0
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 63
If replacement is considered at the end of the nth year,
CCR(n) = cumulative cost if replacement at the end of nth year = TCK(n)  S
n
Average annual cost if replaced at the end of n
th
years
CCR(n) = [(C – S
n
) + EC
m
] / n
Obviously when average annual cost begins to rise, the replacement is due. This is the
year of replacement n is given by when,
CCR(n – 1) > CCR(n) and
CCR(n + 1) > CCR(n)
Implicit in the above is an assumption that level of performance of the equipment as
been maintained the same over the years.
REPLACEMENT BY ALTERNATIVE EQUIPMENT
In the previous article, we decided the due age of replacement on the basis of average
annual related cost of keeping the equipment. It may happen that while existing
equipment is being used, similar but improved version of equipment with different
investment cost, operation cost pattern and resale value pattern is available in the
market. In such a situation, decision is to be made whether the existing equipment needs
to be replaced by new version and if so when such a replacement should be made.
Let A be the existing equipment. Its own minimum average annual cost of keeping is
given by,
n
(CCR)
A
= [(C
a
S
an
) + EC
ma
] / n
0
where, C
a
= capital investment of A
S
an
= resale value of A at the end of nth year
C
ma
= operation maintenance cost of A from year to yea
n = number of years after which the replacement of A is due.
Let B be the new equipment. Its own minimum average annual cost of keeping is given
by,
(CCR)
B
= [(C
b
S
bn1
) + EC
mb
] / n
1
where, C
b
= capital investment of B
S
bn1
= resale value of B at the end of nth year
C
mb
= operation maintenance cost of B from year to yea
n
1
= number of years after which the replacement of B is due.
Obviously, if (CCR)
B
< (CCR)
A
replacement of A by B is apparently advisable. Next
question is when this replacement should be made? It is often suggested that in the year
K when,
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 64
[C
a
– S
a(k+1)
+ SC
ma
] [C
a
+ S
a(k)
+ SC
ma
] > (CCR)
B
when incremental cost of A next year over previous year is more than minimum of
annual average of B the replacement should be made. This is not correct because when
equipment B replaces equipment A there is (C – S) of A carried over as also cumulative
cost of maintenance of A till that year thus in composite arrangement of A replacing B
other cumulative costs of A till that year and that of B thereafter should be worked, and
on that basis the average annular cost be worked out, reckoning from the year of
installation of A.
CAPACITY CORRECTION IN ALTERNATIVE EQUIPMENT
When exiting equipment is replaced by new equipment the latter may not have in
general the same capacity specification. For example, existing lathe may have the
capacity for producing 10 units per hour, whereas new equipment may have capacity
for producing 12 units per hour. In such a situation comparison of costs has to be on the
basis of the same performance and the correction factor should be applied by an
appropriate multiplier.
MONEY VALUE CHANGING WITH TIME REPLACEMENT
The decisions regarding replacement of equipment are taking into account the money
values of related costs new purchase, salvage value, operation cost, maintenance cost,
repair cost from the year. So we have to work out the discounted value of money the
present worth of money against the future cost figures, and find out the numbers of
years, n at the end of which the equipment should be replaced.
If i is the net incremental rate per annum, after one year it becomes (1+i), after two
years (1+i)
2
, and after n years (1+i)
n
. thus, the present valve of 1 Rupee of nth year is 1 /
(1+i)
n
.
If we define r = 1 / (1+i) as discount range, present worth of 1 Rupee of n
th
year is r
n
. If
C is the capital investment of today, m
1
, m
2
,… are the maintained repair cost in 1, 2,
3,… year respectively.
Equipment is used to provide service or give production. It is presumed that repairs and
maintenance cost are carried out from year to year to retain the same level of service or
production.
SALVAGE VALUE CONSIDERED
Previously change of worth of money was considered without salvage value of the
equipment to be replaced. If S
i
is the salvage value of the equipment, its discounted worth at
present is S
i
/ (1 + k)
i
= S
i
r
i
as the salvage value is reckoned at the end of the concerned
year.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 65
Thus equation is modified as
P
n
= C + (m
1
+ rm
2
+ r
2
m
3
+ … + r
n1
m
n
)  S
n
r
n
Its weighted annual average can be shown as
n n
A = (C + E m
i
r
i
1 – S
n
r
n
) / E r
i
i=1 i=1
The replacement is due for that value of n for which A is minimum.
GROUP OF REPLACEMENT POLICY:
It may appear advantageous to replace and item as and when it fails, particularly that item
which does not deteriorate in its performance with lapse of time. However, the inherent risk
in this policy of replacement is that replacement, when needed may take some time may be
for producing item, putting a man to replace the item and time for replacing the item. This
will entail down time or nonworking period of the system. If it is a production system, this
means loss of production. This implies cost. Hence, it is worth considering replacing an
item before it has failed. The concept is similar to „preventive maintenance‟ or „inventory
system‟ in which cost of shortage is very high.
Let us consider a system having N number of identical items. This constitutes a group. If all
items of group are replaced simultaneously at one time, the cost per item is C
g
. Having
installed all N items as a group, the policy is to replace individual items as and when they
are in group. The cost of individual replacement on failure is C
i
. Obviously C
i
> C
g
. The
decision variable is period „t‟. The information needed is failure data over a period of time.
This provides probability of failure. Let us consider following situation.
A large hospital complex has several operation theatres. Each operation table has special
light bulge attachments. The bulge is prone to failure. There are 200 bulbs installed in all.
Considering 500 hours as „period‟ the failure of similar bulb has been as under,
Out of 100 bulbs
9 failed by the end of first period,
20 failed by the end of second period,
33 failed by the end of third period,
51 failed by the end of fourth period,
77 failed by the end of fifth period,
90 failed by the end of sixth period, and
100 failed by the end of seventh period.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 66
The management considers to make it a practice to replace all in a group at one time, then
replace individual bulb as and when it fails, and after „fixed interval‟ of time. It is given that
each bulb, when replaced in a group, costs Rs. 5, but when replaced individually on failure
costs Rs. 20.
Obviously, the criterion for optimization if the cost of the period. The period at the end of
which minimum „average cost per period‟ is obtained, that will be „the fixed interval‟ or
time for group management.
Agewise failure Probability is as under:
01 period 0.09
12 period 0.20 – 0.09 = 0.11
23 period 0.33 – 0.20 = 0.13
34 period 0.51 – 0.33 = 0.18
45 period 0.77 – 0.51 = 0.26
56 period 0.90 – 0.77 = 0.13
67 period 1.00 – 0.90 = 0.10
Let n
k
be the number of replacement at the end of k
th
period.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 67
PROBLEM
Two functionally identical machine tools P and Q are available in the market. There is no
market for machines even oneyear old and when eventually disposed off as scrap it fetches
insignificant amount.
Price of machine tool P is Rs. 12,000. Its operation, maintenance and repair cost is Rs. 400
in the first year, progressively increasing by Rs. 100 in next two years, then by Rs. 200 in
next two years and finally by Rs. 300, 400, 600, 800 in subsequent years.
Price of machine tool Q is Rs. 12,000. Its operation, maintenance and repair cost is Rs. 200,
350, 550, 750, 1000, 1300, 1800, 2400, 3000 from year to year.
(a) If worth of money remains constant, which is a better choice, P or Q? And in that case,
what should be the replacement policy?
(b) If worth of money increases by 10% every year, which is a better choice, P or Q? And
in that case, what should be the replacement policy?
(c) If Q is offered for purchase, through interestsubsidized scheme by State Finance
Corporation, to the extent of 5%, what should be the replacement policy? State the
assumptions you make.
SOLUTION
It is assumed that running costs – operation, maintenance and repairs including downtime
costs – are increased at the beginning of each concerned year.
Table – 1
Machine Tool – P
C = 12000, K = 0, r = 1 / (1+K) = 1
Year
i
m
i
r
i1
m
i
r
i1
E m
i
r
i1
Er
i1
C+Em
i
r
i1
E r
i1
1 400 1 400 400 1 12400
2 500 1 500 900 2 6450
3 600 1 600 1500 3 4500
4 800 1 800 2300 4 3575
5 1000 1 1000 3300 5 3060
6 1300 1 1300 4600 6 2767
7 1700 1 1700 6300 7 2615
8 2300 1 2300 8600 8 2575
9 3100 1 3100 11700 9 2634
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 68
(a) Table – 1 gives the weighted average of annual cost
i
C + E m
i
r
i1
A = i=1
i
E r
i1
i=1
It can be seen with m
8
= 2300, m
9
= 3100, A8 = 2575
i.e., m
8
< A
8
< m
9
Therefore, the minimum present worth of annual value for P is Rs. 2575
Corresponding value for Machine Tool Q, as given in Table – 2 is Rs. 2669.
Thus, machine tool P is a better choice and with this, the replacement policy is to affect the
replacement at the end of 8
th
year.
Table – 2
Machine Tool – Q
C = 13000, K = 0, r= 1 / (1+K) = 1
Year
i
m
i
r
i1
m
i
r
i1
E m
i
r
i1
Er
i1
C+Em
i
r
i1
E r
i1
1 200 1 200 200 1 13200
2 350 1 350 550 2 6775
3 550 1 550 1100 3 4700
4 750 1 750 1850 4 3713
5 1000 1 1000 2850 5 3170
6 1300 1 1300 4150 6 2859
7 1800 1 1800 5950 7 2708
8 2400 1 2400 8350 8 2669
9 3000 1 3000 11350 9 2706
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 69
(b) When change in money worth is taken into account, r = 1 / (1.10). The weighted average
of annual cost, A is tabulated in Table – 3 for machine tool P, and in Table – 4 for
machine tool Q.
Table – 3
Machine Tool – P
C = 12000, K = 0.10, r= 1 / 1.10
Year
i
m
i
r
i1
m
i
r
i1
E m
i
r
i1
Er
i1
C+Em
i
r
i1
E r
i1
1 400 1.000 400 400 1.000 13200
2 500 0.909 455 855 1.909 7081
3 600 0.826 496 1351 2.735 5109
4 800 0.751 601 1925 3.486 4168
5 1000 0.683 683 2635 4.169 3650
6 1300 0.621 807 3442 4.79 3568
7 1700 0.565 961 4403 5.355 3172
8 2300 0.513 1180 5583 5.868 3104
9 3100 0.466 1445 7028 6.334 3096
Table – 4
Machine Tool – Q
C = 13000, K = 0.10, r= 1 / (1.10)
Year
i
m
i
r
i1
m
i
r
i1
E m
i
r
i1
Er
i1
C+Em
i
r
i1
E r
i1
1 200 1.000 200 200 1.000 13200
2 350 0.909 318 518 1.909 7081
3 550 0.826 454 972 2.735 5109
4 750 0.751 63 1535 3.486 4168
5 1000 0.683 683 2218 4.169 3650
6 1300 0.621 807 3025 4.790 3568
7 1800 0.565 961 3986 5.355 3172
8 2400 0.513 1231 5217 5.868 3104
9 3000 0.466 1398 6615 6.664 3096
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 70
For machine too P, least value of A is Rs. 2996, whereas for machine tool Q this value
does not appear in the range of data given. However, the least value for Q as available is
Rs. 3096. Therefore, P is better choice with replacement at the end of 8th year.
(c) When interest for machine tool Q is subsidized, r = 1 / (1 + 0.05) and the value of A are
tabulated in Table – 5. Comparing the least value of Rs. 2873 with the least value of D
for P in Table – 3 i.e. Rs. 296, it can be concluded that the machine tool Q is better
choice, with replacement at the end of 8th year.
Table – 5
Machine Tool – Q
C = 13000, K = 0.05, r= 1 / (1.05)
Year
i
m
i
r
i1
m
i
r
i1
E m
i
r
i1
Er
i1
C+Em
i
r
i1
E r
i1
1 200 1.000 200 200 1.000 13200
2 350 0.952 333 533 1.925 6933
3 550 0.907 499 1032 2.859 4911
4 750 0.864 648 1680 3.723 3945
5 1000 0.823 823 2503 4.546 3411
6 1300 0.784 1019 3522 5.330 3101
7 1800 0.746 1268 4790 6.076 2929
8 2400 0.711 1706 6496 6.787 2873
9 3000 0.677 2031 8527 7.464 2884
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 71
EXPERIMENT.NO: 11
STUDY ABOUT SEQUENCING PROBLEM.
 HEUSTERIC PROBLEM SOLVING.
Problem solving methods discuss earlier posses well defined models. Analytical
techniques and algorithms to obtain optimum solution. As was stated in model formulation
phases in chapter of real life problems. In general defi. Clear cut crystallization and
trimming problem can be brought within the relation of models. The techniques to the tool
kit of O.R. will take care of the solution. However in many cases this may not be possible.
In such cases heustric approach in problem solving comes hardly.
It is now obvious that the illstructured problems are those problems which are not
amenable to solution by analytical or algorithms solution techniques. The concept of ill
structured problems is understood in contrast to well structured problem ( WSP), the
character being characterized by :
1. formulable as an acceptable model.
2. capable of attaining feasible solution.
3. defined criteria for feasibility and optimum solution.
4. availability of techniques for solving models.
 SEQUENCING PROBLEM :
The problem of sequencing falls within the realm of heuristic methods. Sequencing
problems encompasses those scheduling problems in which objective function or
sequence for processing set of jobs or activity to be undertaken on given facility in a
defined technological order. Sequence is a very important function in production
planning and control. This chapter will concentrate machine to perform operation on the
job which have unique design futures and have resulted from a specific customer order.
In order to describe a specific case of job shop scheduling problem we want to study for
4 factors which are
1. the job arrival . if „n‟ jobs arrive at a time in a shop where „m‟ facilities on
which they are to be processed are idle and available . the problem is called
as static problem. The problem will be known as dynamic problem if jobs
arrive at random with some stock tic process.
2. number of machines „m‟ that are in the job shops.
3. the flow process for „n‟ jobs through „m, machines.
4. the measure of performance to be optimized which plays a decisive roll in
job shop schedule problem.
Gantt chart is a most simple and widely expected representation of scheduling are
shown along vertical axes and operations time along horizontal axis. Each operation
on various resources (facility over machines) is represented by a horizontal bar.
Since the chart breaks down resources allocation by time. It gives a display of max
span. Job waiting chart breaks down resources allocation by time it gives a display
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 72
of observation of this chart. And improvement to the schedule depend on analyst
critical observation.
The general problem of sequencing is „n‟ jobs are to be performed on 1,2,or
several or all machine. Each job having its own technological order on machine. The
sequence be so determine as to minimize total elapsed time for „n‟ jobs to be
completed.
Rules of heuristic methods in sequencing problems have so far been
establish for a few specific cases viz.
1. n jobs on 1 machine.
2. n jobs on 2 machine x & y in the same order.
3. n jobs on 3 machine x,y & z and subject to special restriction of data.
4. 2 jobs on m machines each jobs may have its technological order and
5. n jobs on m machines having in the same technological order.
 „N‟ JOBS AND „1‟ MACHINE :
The simplest case of scheduling is to decide sequence of n jobs on a single resource or
machine. In a basic single machine problem n jobs are available for processing at time
„0‟ (static problem). Their facility setup times are independent on their sequence and
can be added to their individual processing time for the power of scheduling . this
aggregate of setup times and processing times for each job is constant. And known with
certainty . the facility or machine is continuously available without interruption for
processing the jobs and will not be idle. So long a job is waiting . once processing of
jobs started will be completed without any break . the total number of distinct schedules
will be n! . since all the schedules will results in the same next pan different
performance are used to evaluate the schedules.
 „N‟ JOBS AND „2‟ MACHINES:
For further detail description let us consider the illustrative problem –1 given in tutorial.
 „N‟ JOBS AND „3‟ MACHINES :
let us consider n jobs on 3 machines . F
1
,F
2
& F
3
is same technological order. There is
no general optimum or satisfying solution obtain so far.
If a problem satisfied for n jobs and 2 machines as desired in art. The condition is
minimum time element on F
1
>max. time element on F
2
. and minimum time element on
F
3
> max. time element on F
2
.
In that case , 3 columns (. F
1
,F
2
& F
3
) time elements are redesigneted for 2 machines
A & B elements as ,
A
i
= ( F
1
)
I
+ ( F
2
)
I
and B
i
= ( F
2
)
I
+ ( F
3
)
I
 „2‟ JOBS AND „M‟ MACHINES :
jobs J
1
& J
2
are to be made on machine F
1,
F
2,
F
3,
….. F
m
not necessarily in the same
technological order. Graphical methods provides a very, simple for obtaining solution.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 73
ASSIGNMENTS
LP FORMULATION – ASSIGNMENT
1. A small manufacturer employs five skilled men and ten semi skilled men for making a
product in two qualities: a deluxe model and an ordinary model. The production of a
deluxe model requires 2hours work by a skilled man and a 2hour work by a semi
skilled man. The ordinary model requires 1hour work by a skilled man and a 3hours
work by a semi skilled man. According to worker union‟s rules, no man can work more
than 8hours per day. The profit of the deluxe model is Rs.1000 per unit and that of the
ordinary model is Rs.800 per unit. Formulate a linear programming model for this
manufacturing situation to determine the production volume of each model such that the
total profit is maximized.
2. A firm manufactures three products A, B & C. Their profits per unit are Rs.300, Rs.200
and Rs.400, respectively. The firm has two machines and the required processing time
in minutes on each machine for each product is given in the following table:
Product
A B C
Machine 1 4 3 5
Machine 2 2 2 4
Machines 1 and 2 have 2000 and 2500 machine minute, respectively. The upper
limits for the production volumes of the product A, B & C are 100 units, 200 units and
50 units, respectively. But, the firm must produce a minimum of 50 units of the product
A. Develop a LP model for this manufacturing situation to determine the production
volume of each product such that the total profit is maximized.
3. The manager of an oil refinery has to decide on the optimal mix of two possible blending processes. The inputs and the outputs per
production run of the blending process are as follows:
Input Output
Process Crude A Crude B Gasoline G
1
Gasoline G
2
1 5 3 5 8
2 4 5 4 4
The maximum amounts of availability of crude A and crude B are 200 units and 150
units, respectively. Market requirements show that at least 100 units of Gasoline G
1
and 80 units of Gasoline G
2
must be produced. The profits per production run from
process 1 and process 2 are Rs.3 lacs and Rs. 4 lacs respectively. Formulate this
problem as a LP model to determine the number of production runs of each process
such that the total profit is maximized.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 74
SOLVE EXAMPLES USING SIMPLEX METHOD
1. Minimize Z=2x+9y4z
Subject to 2x+3y+4z s 16
x+6y4z>16
x,y, z >0 [S
1
=8, y=8/3,Z=24]
2. Maximize z=4x+5y
Subject to 2x+3ys8
3x+y>4
x,y>0
Use big M method. [S
2
=8, x=4, z=16]
3. Maximize Z=6x+10y+8z
Subject to 2x+3ys8
2y+5zs10
3x+2y+4zs15
x,y,z>0 [y=750/41, z=930/41, x=89/41, Z=212.78]
4. Maximize Z=40x+30y
Subject to 3x+2ys300
x+ys80
2x+ys200
3x+4ys300
xs60
ys60
x,y>0 [S
1
=80, y=20, S
3
=60, S
4
=40, x=60, S
6
=40, Z=3000]
5. Minimize Z=20x
1
+10x
2
Subject to x
1
+2x
2
s40
3x
1
+x
2
=30
4x
1
+3x
2
>60
x
1,
x
2
>0
Use (i) Big „M‟ Method.
(ii) Two – Phase Method [x
1
=6, x
2
=12, S
1
=10, Z=240]
6. Solve the following problem:
Maximize Z= x
1
4x
2
3x
3
Subject to 2x
1
+x
2
+3x
3
>4
x
1
+2x
2
+2x
3
>3SS
x
1
, x
2
>0 and x
3
is unrestricted in sign. [Problem is infeasible]
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 75
FORMULATE PROBLEM AND SOLVE USING SIMPLEX METHOD
1. Shri Rajendra Desai has booked orders for product A and product B at the sales
price of Rs. 6 and Rs. 8 per unit respectively, for M/S supreme Engineers with the
condition that materials cost will be borne by the company i.e. M/S Supreme
Engineers. Shri Desai negotiates the terms with the Kartik Engineering, a job order
shop, which can offer machines of type x for 60 hours per week and machines of
type y for 100 hours per week at hourly rates. Each machine can produce product A
and B in any combination.
Machinehours of type x are available with the operators of the Kartik Engineering only
and the rates are Rs. 4 per machinecumoperatorhour. Machines of type y are
available with two alternative offers viz. Desai can either hire the machines alone at the
rate of Rs. 1/machine hour as well as operator at the rate of Rs. 3 per hour or Desai can
hire his own operator for machines y at the rate of Rs. 3 per hour. The production rates
are given in the Exhibit A.
Production rate, units per hour Exhibit – A
Product Machines, Type x Machines, Type y
With operator of
Kartik Engg.
With Desai‟s
operator
A 10 6 8
B 20 12 18
Formulate L.P.P. to maximize profit of Shri Desai.
[ Maximize
Z = (6 4/10) X
1
+ (84/20) X
2
+ (64/6) X
3
+ (84/12)X
4
+(64/8)X
5
+ (84/18)X
6
Subject to X
1
/10 + X
2
/20 s 60
X
3
/6 + X
4
/12 + X
5
/8 + X
6
/18 s 10
X
1,
X
2,
X
3,
X
4,
X
5,
X
6,
> 0]
2. The State Industries Finance Corporation is to decide its investment of Rupees 20
lacs on two proposals of Private sector Industries X and Y> X undertakes to
guarantee annual return of 10 per cent, and Y of 15 per cent. As X is in the line of
basic consumer products and that too cooperative venture, the Government has laid
down that at least Rs.5 lacs should be invested in X. The corporation would like to
have investments so made as to procure the minimum of 12 per cent annual return.
For the reason best known to one of the Directors, following his rigid attitude in this
regard SIFC has decided not to have investment in Y and X more than in the ratio
1.6: 1.
Formulate this L.P. problem for maximum annual return. Solve it be either (1)
Graphical method or (2) Simplex method.
[maximize
Z = 0.1 X
1
+
0.15 X
2
Subject to X
1
+ X
2
= 20
0.1 X
1
+ 0.15 X
2
> 0.12 (X1
+ X
2
)
X
1
>
5
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 76
X
2
/ X
1
s 1.6 / 1
X
1,
X
2
> 0
X
1
= 765/91, X
2
= 1055/91, S
2
= 107/260, Z = 2.5797]
3. The following data is available for two products.
Product Production Alternatives Production Rate in units/hr.
Centre 1 Centre 2
1 1
2
0.4


0.5
2 1 0.5 1.0
Production hrs. available/week 500 400
Production cost Rs./hr. 6 8
Material cost is same for both the products and is Rs. 5 per unit. Selling price Rs.
Per unit for product one is 25 and that of product two is 31. Solve the problem for
optimal product mix.
[X1
= 1250, X
2
= 800, X
3
=0, Z = 9450]
4. A plant can produce three products, A, B and C, Product A needs 4 hours of
department 2 and one hour of department 3. Product B needs two hours of
department 1 and two hours of department 2, while product C needs two hours of
department 1 and two hours of department 3. Their respective profit coefficients in
Rs. Per unit are 50, 40, and 55. Hours available per month with department 1, 2, and
3 are 1000, 1000, and 800 respectively.
A purchased part is used in assembly of product A and C. Only 4000 parts are
available for the month. Two such parts are used in each piece of product A and
three parts in each piece of product C. There is a sales commitment of 200 units of
product A. Find the optimal product mix.
DUALITY & SENSITIVITY ANALYSIS
Problem 1:
Construct the dual of the problem:
Maximize Z = 3x+5y
Subject to X 2y > 3
X + 3y >9
X – y s 5
X > 0, is unrestricted in sign
Solve the dual.
[DUAL IS MAXIMIZE Z = 3A9B+5C
SUBJECT TO –AB+C >3
2A+3B+C >5
A,B,C >0.
C=11,A=8,S3=0,Z=31]
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 77
Problem: 2
Use dual simplex method to solve the following
Minimize Z = 2x + 3y
Subject to 3x + 4y > 5
4x + 5y > 7
x + 2y s 4
x , y > 0
[S1 = ¼, X= 7/4, S3 =9/4, Z = 7/2]
Problem: 3
The following tableau gives an optional solution to a standard linear programme:
Maximize Z = CX.
Subject to AX = B, X >0
Tableau 5 7 4 0 0
X1 X2 X3 S1 S2
7 x2 x3 ¾ 1 1/4 ¼ 0
0 s2 ½ ¼ 0 3/4 5/4 1
(z
j
– c
j
) 7/2 ¼ 0 9/4 7/4 0
S
1
and S
2
are the slack variables.
(a) How much can c
1
be increased before the current basis is no longer optimal?
(b) How much can c
2
be varied sp that the given basis ( x
2
, S
2
) is still optimal?
(c) What are the shadow prices of two resources?
[(a) c
1
=21/4, x
1
=2/3, S
2
=1/3,Z=4 (b) 20/3 c
2
16 (c) 7/4 and 0]
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 78
QUEING THEORY
1. Customer arrivals at a teller counter in a Commercial Bank are considered to be
following Poisson probability distribution with an average time of 10 minutes
between one arrival and the next. The time length of service that is rendered is
assumed to be distributed exponentially with mean three minutes
i) What is the probability that a person arriving at the teller will have to
wait?
ii) What is the average length of queue that forms from time to time?
iii) The Commercial Bank will install a second teller man when convince
that the arrival rate increase in order to justify the second booth. When
does this happen?
2. There is congestion on the platform of Ahmed Railway station. The trains arrive at
the rate of 30 trains per day. The waiting time for any train to flagoff is
exponentially distributed with an average of 36 minutes. Calculate the following:
i) The mean queue size.
ii) The probability that the queue size exceeds 10.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 79
INVENTORY CONTROL ASSIGNMENT
1. The Bombay Shoe Company has found that it purchases a large amount of industrial
tapes for production of its shoes. Currently, it purchases Rs.10,00,000 a year of the
various sized tapes from the local manufactures. A proposal was made by its
supplier, was offer consist of 1.25% discount if BSC places an order quarterly BSC
has calculated the cost to purchase at Rs.22.5 per order. Rs.20 per tape & inventory
carrying costs at 18%. Should BSC accept discount offer from its supplier? If not,
what alternate offer should be made in term of a discount?
2. The Himavan manufacturing company wishes to determine the most economy
quantity for one of its products. Manufacturing cost amount to Rs.15 per unit. The
production is 5000 units per annum. Each lot now requires a setup cost of Rs.25
and the inventory cost of 25% of the average inventory value. What is the most
economic lot size to manufacturers? What is the corresponding total yearly cost?
3. Demand for a certain part order by Shah brothers tends to be constant at the monthly
rate of 1000 units. The per unit carrying cost of this item is Rs.25 per year, & cost of
placing an order is Rs.75.
i) What is the optimal order size? How often should an order be placed?
ii) Show that the annual holding cost & ordering cost are equal when an
optimal order size is used. What is the total relevant cost?
iii) If the company wants to order only one every other week, by what
percentage would this increase the total relevant cost?
4. The Calcutta Tool Company can manufacture a certain type of tool at the rate of
1480 per run, the demand for this tool is quite steady at annual rate of 9000 units.
Unit cost of the tool is Rs.30, an set up cost per production run is Rs.500. The
annual cost per unit sort is Rs.20. The company has determined the optimal
production lot size to be 3000 units.
i) What is the annual inventory holding cost, expressed as percentage of
unit cost?
ii) What is the total relevant cost & the max. inventory level?
5. Let C1=50(Rs.), Cs=20C1, & the demand distribution is as under:
r 0 1 2 3 4 5 6
P(r) 0.9 0.05 0.02 0.01 0.01 0.01 0
What is the optimum quantity to be ordered?
5. Assume that demand during a certain time interval T is random with P( r ) being the
probability that the total demand is r during interval T. The demand rate is constant
during the interval T.
C1=10(Rs.), C2=20C1 and the demand distribution is as under:
r 0 1 2 3 4 5 6
P(r) 0.1 0.2 0.2 0.3 0.1 0.1 0
What should be the initial inventory level?
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 80
PROBLEMS ON GAMES THEORY
1. Linear programming method:
(i) (ii)
B B
1 2 3 1 2 3
1 1 7 2 A 1 5 7 9
A 2 6 2 7 2 4 8 11
3 6 1 6
(iii) (iv)
B B
1 2 3 1 2 3
A 1 1 6 4 1 4 7 9
2 5 3 6 A 2 11 6 10
6 3 5 6 9
(v) (vi)
B B
1 2 3 1 2 3 4
1 5 10 20 1 3 1 1 2
A 2 5 10 10 A 2 2 3 2 3
3 5 20 20 3 2 2 1 1
(vii) (viii)
B B
1 2 3 4 1 2 3 4
1 4 10 11 14 1 1 3 5 4
2 15 5 13 18 A 2 3 4 3 2
A 3 4 9 6 10 3 3 2 4 3
4 17 12 10 7
5 14 18 17 8
6 11 14 15 9
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 81
TRANSPORTATION
Problem: Find the initial basic feasible solution of the following transportation problem by
Vogel‟s approximation method:
Warehouses
W
1
W
2
W
3
W
4
Capacity
F
1
10 30 50 10 7
F
2
70 30 40 60 9
F
3
40 8 70 20 18
Requirement 5 8 7 14 34
Problem: A company has received a contract to supply gravel for three new construction
projects located in towns A, B and C. Construction engineers have estimated the required
amounts of gravel which will be needed at these construction projects as shown below:
Project location Weekly requirement
(Truck loads)
A 72
B 102
C 41
The company has three gravel plants X, Y and Z located in three different towns.
The gravel required by the construction projects can be supplied by these three plants. The
amount of gravel which can be supplied by each plant is as follows:
Plant Amount available/week
(Truck loads)
X 76
Y 62
Z 77
The company has computed the delivery cost from each plant to each project site.
These costs (in rupees) are shown in the following table:
Cost per load
A B C
X 4 8 8
Plant Y 16 24 16
Z 8 16 35
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 82
(a) Schedule the shipment from each plant to each project in such a manner so as to
minimize the total transportation cost within the constraints imposed by plant
capacities and project requirements.
(b) Find the minimum cost.
(c) Is the solution unique? If it is not, find alternative schedule with the same minimum
cost.
ASSIGNMENT
Problem: Solve the following assignment problem using Hungerian method. The matrix
entries are processing times in hours.
Operator
1 2 3 4 5
1 20 22 35 22 18
2 4 26 24 24 7
Job 3 23 14 17 19 19
4 17 15 16 18 15
5 16 19 21 19 25
Problem: Consider the problem of assigning four sales persons to four different sales
regions as shown below such that the total sales is maximized.
Sales region
1 2 3 4
1 5 11 8 9
2 5 7 9 7
Salesman 3 7 8 9 9
4 6 8 11 12
The cell entries represent annual sales figures in crores of rupees. Find the optimal
allocation of the sales persons to different regions.
Problem: The flight timings between two cities, X and Y are as given in the following two
tables. The minimum layover time of any crew in either of the cities is 3 hours. Determine
the base city for each crew so that the sum of the layover times of all the crews in nonbase
cities is minimized.
Timing of Flights form City X to City Y
Flight number Departure time Arrival time
(from City X) (to City Y)
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 83
101 6 a.m. 8.00 a.m.
102 10 a.m. 12.00 noon
103 3 p.m. 5.00 p.m.
104 8 p.m. 10.00 p.m.
Timing of Flights from City Y to City X
Flight number Departure time Arrival time
(from City Y) (to City X)
201 5.30 a.m. 7.00 a.m.
202 9.00 a.m. 10.30 a.m.
203 4.00 p.m. 5.30 p.m.
204 10.00 p.m. 11.30 p.m.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 84
NETWORK PROBLEMS:
Problem: Consider the details of a distance network as shown below:
Arc Distance Arc Distance
12 8 36 6
13 5 45 8
14 7 46 12
15 16 58 7
23 15 68 9
26 3 69 15
27 4 79 12
34 5 89 6
(a) Construct the distance network.
(b) Find the shortest path from Node 1 to Node 9 using the systematic method.
(c) Find the shortest path from Node 1 to Node 9 using Dijkstra‟s algorithm.
Problem: Consider the details of a distance network as shown below:
Arc Distance
12 3
13 8
14 10
23 4
24 7
34 2
35 8
36 6
(a) Construct the detail network.
(b) Apply Floyd‟s algorithm and obtain the final matrices, D
5
and P
5
.
(c) Find the shortest path and the corresponding distance for each of the following:
(i) from Node 1 to Node 5
(ii) from Node 2 to Node 5.
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 85
Problem: For the network shown in the Figure activity, resource requirements are given in
the table. Solve the problem for resource allocation. There are two cranes and 8 welders
available for the project.
Figure
Table
Activity A B C D E F
Duration
Days
10 10 8 14 3 9
Resources
Required

2
Cranes
8
Welders
6
Welders

1
Crane
1 2 4 6
3
5
A D E
A
C B
F
OPERATIONS RESEARCH
PRODUCTION ENGINEERING DEPT. 86
Problem: The critical path network analysis is given in Figure. The activity times are also
given for each activity. Determine the critical path and critical path time. Also determine
the floats for each activity.
Figure
1 2
4
3
6
8
5 7
2
A
12
12
8
9
3
4
5
7
10
This action might not be possible to undo. Are you sure you want to continue?
Use one of your book credits to continue reading from where you left off, or restart the preview.