0% found this document useful (0 votes)
8 views130 pages

Or All Units Lecture Notes

Operations Research (OR) is a systematic study aimed at optimizing decision-making within organizations through quantitative methods. The development of OR spans pre-World War II, during the war, and post-war applications, with techniques utilized across various sectors including industry, defense, agriculture, and public utilities. Key components of OR include problem identification, data collection, model formulation, and the use of mathematical models, while also addressing limitations such as resistance to change and the challenge of quantifying qualitative factors.

Uploaded by

reddyjayanth7675
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
8 views130 pages

Or All Units Lecture Notes

Operations Research (OR) is a systematic study aimed at optimizing decision-making within organizations through quantitative methods. The development of OR spans pre-World War II, during the war, and post-war applications, with techniques utilized across various sectors including industry, defense, agriculture, and public utilities. Key components of OR include problem identification, data collection, model formulation, and the use of mathematical models, while also addressing limitations such as resistance to change and the challenge of quantifying qualitative factors.

Uploaded by

reddyjayanth7675
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 130

UNIT-1

OPERATIONS RESEARCH

Introduction
It is the systematic method oriented study of the basic structure, characteristics, functions
and relationships of an organization to provide quantitative basis for decision making. OR is
concerned with scientifically deciding how to best design and operate man machine systems
usually requiring the allocation of Resources.

Development of Operations Research


The development of OR can be studied under the following classification.

i) Pre World War-II Developments

Some of the techniques of today's OR have been actually developed and used even
before the term OR was coined, some of the techniques are Inventory, Queuing Theory,
Statistical quality control etc.

ii) During World War-II Developments

Scientists from different fields were jointly directed to do research on military


operations for improving its effectiveness with the limited resources. Later on this scientific and
inter disciplinary approach became an important problem solving aspect of OR methodology.

iii) Post World War-II Developments

After the World war-II the industries in America & Britain concentrated in applying
the Operations research methodologies to industrial problems for maximizing the profit with
limited Resources.

Definitions of Operations Research


• OR is the Research of operations. An operation may be called a set of acts required for
the achievement of desired outcome.

• According to H. M. Wagner, OR is a scientific approach to problem solving for executive


management.

• According to OR society of America, OR is an experimental & applied science devoted


to observing, understanding& predicting the behavior of purposeful, man machine
systems.
Scope of Operations Research
Whenever there is a problem of optimization, there is a scope of application of OR. Its
techniques have been used in a wide range of situations.

1) In Industry

OR techniques are useful in the field of production, inventory control, demand& forecast,
sales & purchase, transportation, repair & maintenance, scheduling & sequencing, planning,
scheduling & control of the projects.

2) In Defense

OR has a wide scope of application in defense operation. The activities performed by air
force, army, navy can be further divided into sub activities. i. e operations, intelligence,
administrates, and training etc. There is a need to co-ordinate various activities involved in order
to arrive optimum strategies and to achieve consistent Goal.

3) Planning

In both developing & developed economic OR approach is equally applicable. The basic
problem in more of the countries from Asia & Africa is remove poverty. Therefore there is a
great scope of Economics, Statistics, Administration, Technician &Agriculture experts working
together to solve the problem with an OR approach.

4) Agriculture

With population expansion & shortage of food every country is facing the problem of
optimum allocation of land with various crops in according to climate conditions and available
facilities.

5) Public utilities

OR Techniques can be applied in the area of transformation to regulate train arrivals and
their running times. OR methods can also applied to big hospitals to reduce waiting time of
patients. It has extensive uses in petroleum, chemical, paper, metal processing, air craft, rubber
mining and textile industries.

Phases of OR Study
The scientific method of OR study involves the following phases.

1) Judgment Phase: This phase includes


a) Identification of real life problem.

b) Selection of an appropriate objective

c) Application of the appropriate measurement

d) Formulation of an appropriate of model of the problem.

2) Research phase: This phase is the largest & longest among all the phases. This phase utilizes

a) Observations and data collection of better understanding of the problem.

b) Formulation of hypothesis and model.

c) Observation & experimentation to test the hypothesis on the basis of data.

d) Analysis of the available information & verification of the hypothesis.

e) Prediction of various results from the hypothesis.

f) Generalization of result & consideration of alternate methods.

3) Action phase:

This phase consists of making recommendations for implementing the decision. This
decision is implemented by an individual who is in a position to implement result. This
individual must aware of the environment in which the problem is occurred.

Objectives of OR
The objective of OR is to minimize the total cost and maximize the profit. Every
organization has a number of functional units or departments each perform a part of whole
job. With Economic growth, uncertainty is also growing; this makes each decision
complicate and time consuming.

In the competitive world today one has to take quick decision, because any delay may
help the competitors. The decision has to be quick as well as sound & requires a scientific
approach to the problem. The application of OR methods to helps in making the decisions in
such complex situations. OR combines the knowledge of various disciplines such as
mathematics, statistics, economics & engineering.
Objective of OR is to provide a scientific basis to the managers of an organization for
solving problems involving components of the system by employing a team of scientists
drawn from different disciplines.

Models of OR
There is no set of unique problems which can be solve by using OR models or techniques.
Several OR models or techniques can be grouped into some basic categories as given below.

1) Allocation model

2) Inventory model

3) Competitive model

4) Network model

5) Waiting line model

6) Sequencing model

7) Replacement model

8) Dynamic programming model

9) Markov chain model

10) Decision Analysis model

Uses of OR
• In all industries & other situations, resources are limited. OR studies help in effectively
selecting and distributing these resources.

• By using OR techniques the quality of decisions are better and in most of the cases
optimal.

• As it is a scientific way of making decisions, it gives a lot of scopes for today's manager
to develop skills or knowledge to arrive at optimal decisions.

• Many of OR techniques do not give a single decision.

• They give a number of decisions yielding the same results; hence depending upon the
situation most appropriate decision can be selected or used.
Limitations of OR
• Mathematical models which essence of OR, do not take into account qualitative factors or
emotions factor which are quite real. All influencing factors which cannot be quantified
find no place in mathematical models.

• Mathematical models are applicable to only specific category problems.

• OR tries to find optimal solution taking all factors of the problem into account. Present
day problems involve numerous such factors, expressing them in quantity and
establishing relations among them requires huge calculations.

• Being a new field generally there is a resistance from the employees to the new
proposals.

• Management who has to implement the proposals may itself offer a lot of resistance due
to conventional thinking.

• Young enthusiastic persons over taken by its advantages and exactness generally forget
that OR mean for men and not that men are meant for it.

Linear Programming Problem (LPP)


It is an optimization (Maximization or Minimization) technique which optimizes a linear
function called objective function by using some inequalities called as constraints and the
variables are called the non-negative restriction.

Formulation of LPP
Given a situation describing the availability of resources and conditions and the object to be
optimized. we have to convert in to a mathematical model , this process is called formulation.

It consists of the following steps

Step-1: Defining the Decision variables and writing the objective function.

Step-2: Expressing the conditions of availability of the resources in the form of linear
inequalities or equalities using the decision variables.

Usually LPP is of the form

𝑴𝒂𝒙( 𝒐𝒓 )𝑴𝒊𝒏𝒁 = 𝒄𝟏 𝒙𝟏 + 𝒄𝟐 𝒙𝟐 + − − − − − − + 𝒄𝒏 𝒙𝒏 (𝑶𝒃𝒋𝒆𝒄𝒕𝒊𝒗𝒆 𝒇𝒖𝒏𝒄𝒕𝒊𝒐𝒏)

Subject to constraints:
𝒂𝟏𝟏 𝒙𝟏 + 𝒂𝟏𝟐 𝒙𝟐 + − − − − − − + 𝒂𝟏𝒏 𝒙𝒏 (≤, =, ≥) 𝒃𝟏

𝒂𝟐𝟏 𝒙𝟏 + 𝒂𝟐𝟐 𝒙𝟐 + − − − − − − + 𝒂𝟐𝒏 𝒙𝒏 (≤, =, ≥) 𝒃𝟐

−−−−−−−−−−−−−−−−−−−−−−−−−

𝒂𝒎𝟏 𝒙𝟏 + 𝒂𝒎𝟐 𝒙𝟐 + − − − − − − + 𝒂𝒎𝒏 𝒙𝒏 (≤, =, ≥) 𝒃𝒎

𝒂𝒏𝒅 𝒙𝟏 , 𝒙𝟐 , − − − − −𝒙𝒏 ≥ 𝟎 (𝒏𝒐𝒏 − 𝒏𝒆𝒈𝒂𝒕𝒊𝒗𝒆 𝒓𝒆𝒔𝒕𝒓𝒊𝒄𝒕𝒊𝒐𝒏𝒔)

Solved Problems on formation of LPP:


1. A company produces 2 products A &B. To get a finished product of A&B, the raw
material has to process on different processes P1 and P2. One unit of product A takes 2
hours of processing time on P1 and 1 hour processing time on P2. One unit of B takes 4
hours processing time on P1 and 3 hours processing time on P2. On any working day P1 and
P2 are available 20 hours and 23 hours respectively. By selling one unit of A the company
gets Rs. 20 as profit. By selling one unit of B the company gets Rs.12 as profit. Formulate
given problem as LPP.
Soutionl. Let x1, x2 be the
number of units of A & B to be produced. We can write the following table from given data

Processing Time
Product Profit
P1 P2
A(x1) 2 hrs. 1 hrs 20
B(x2) 4 hrs 3 hrs 12
Availability 20 hrs 23 hrs

𝑴𝒂𝒙 𝒁 = 𝟐𝟎𝒙𝟏 + 𝟏𝟐𝒙𝟐 [𝑶𝒃𝒋𝒆𝒄𝒕𝒊𝒗𝒆 𝒇𝒖𝒏𝒄𝒕𝒊𝒐𝒏]


Subject to constraints

𝟐𝒙𝟏 + 𝟒𝒙𝟐 ≤ 𝟐𝟎,

𝒙𝟏 + 𝟑𝒙𝟐 ≤ 𝟐𝟑
𝒙𝟏 , 𝒙𝟐 ≥ 𝟎 [𝒏𝒐𝒏 − 𝒏𝒆𝒈𝒂𝒕𝒊𝒗𝒆 𝒓𝒆𝒔𝒕𝒓𝒊𝒄𝒕𝒊𝒐𝒏𝒔]
2. National paints produce both interior and exterior paints from 2 raw materials M1 and
M2. The following table provides basic data of the problem.

Product Maximum daily


Exterior Interior availability
Raw M1 6 Tons 4 Tons 24
material M2 1 Tons 2 Tons 6
Profit per Ton 5 4

A market survey indicates that the daily demand for interior paint cannot exceed that of
exterior paint by more than one ton, also maximum daily demand for interior paint is 2
tons. Formulate as LPP.

Solution. Let x1, x2 be the exterior and interior paints produce daily in tons. Assume Z be the
daily profit.

𝑴𝒂𝒙 𝒁 = 𝟓𝒙𝟏 + 𝟒𝒙𝟐

Subject to

𝟔𝒙𝟏 + 𝟒𝒙𝟐 ≤ 𝟐𝟒,

𝒙𝟏 + 𝟐𝒙𝟐 ≤ 𝟔,

𝒙𝟐 − 𝒙𝟏 ≤ 𝟏,

𝒙𝟐 ≤ 𝟐

𝒙𝟏 , 𝒙𝟐 ≥ 𝟎

Exercise:

1. A firm manufactures 2 types of products A &B and sells them at a profit of Rs. 2 on type
A and Rs.3 on type B. Each product is processed on 2 machines G and H. Type A requires
1 minute processing time on G and 2 minutes on H. Type B requires 1 minute on G an 1
minute on H. The machine G is available for not more than 6hours 40 minutes, while H
available for 10 hours on any working day. Formulate this problem as LPP.
Solution of LPP
To solve LPP we use the following methods.

1) Graphical Method: If LPP consists of only 2 decision variables and whatever be the form of
constraints (≤, =, ≥) and inequalities we use graphical method.

2) Simplex Method: If LPP having 2 or more decision variables and if all the inequalities are in
the form of ≤, we use simplex method.

3) Artificial variable technique: If LPP having 2 or more variables and at least one constraint is
in the form of ≥, = we use artificial variable technique.

The following 2 methods are there in artificial variable technique.

a) Two phase Simplex method

b) Big-M method

1) Graphical Method:

Procedure

i) Consider each inequality constraint as equation.

ii) Plot each equation on the graph geometrically each one represents a straight line.

iii) Every point on the line will satisfies the equation of the line.

iv) If the inequality constraints corresponding to the line is ‘≤’. The region below the line in the
1st quadrant is shaded, for any inequality constraint with ‘≥’ the region above the line in 1st
quadrant is shaded, the points lying in the common region will satisfy all the constraints
simultaneously. The common region is called as “Feasible Region”.

v) Find the co-ordinates of corner points of feasible region.

vi) Locate the corner points of optimal solution by calculating the value of Z for each corner
point.

Solved problems on Graphical Method


1. Solve the following LPP by Graphical method

𝑴𝒂𝒙 𝒁 = 𝟐𝒙𝟏 + 𝟑𝒙𝟐

Subject to
𝒙𝟏 + 𝒙𝟐 ≤ 𝟒𝟎𝟎,

𝟐𝒙𝟏 + 𝒙𝟐 ≤ 𝟔𝟎𝟎,

𝒙𝟏 , 𝒙𝟐 ≥ 𝟎

Solution.

Let 𝑥1 + 𝑥2 = 400 -------------(1)

2𝑥1 + 𝑥2 = 600 -------------(2)

Put 𝑥2 = 0 in equation (1)

 𝑥1 = 400
At (400, 0) equation (1) meets 𝑥1 -axis

Put 𝑥1 = 0 in equation (1)

 𝑥2 = 400
At (0, 400) equation (1) meets 𝑥2 -axis

Put 𝑥2 = 0 in equation (2)

 𝑥1 = 300
At (300, 0) equation (2) meets 𝑥1 -axis

Put 𝑥1 = 0 in equation (2)

 𝑥2 = 600
At (0, 600) equation (2) meets 𝑥2 -axis
Here OABC is the feasible region where O(0,0), A(300,0), B(200,200),C(0,400)

Max Z at O(0,0) = 0

Max Z at A(300,0)=600

Max Z at B(200,200)=1000

Max Z at C (0,400)=1200

∴ Max Z =1200

And 𝑥1 = 0, 𝑥2 = 400

2. Solve the following LPP by Graphical method


𝑴𝒊𝒏 𝒁 = 𝟐𝟎𝒙𝟏 + 𝟏𝟎𝒙𝟐

Subject to

𝒙𝟏 + 𝟐𝒙𝟐 ≤ 𝟒𝟎,Type equation here.

𝟑𝒙𝟏 + 𝒙𝟐 ≥ 𝟑𝟎,

𝟒𝒙𝟏 + 𝟑𝒙𝟐 ≥ 𝟔𝟎,

𝒙𝟏 , 𝒙𝟐 ≥ 𝟎

Solution.

Let 𝑥1 + 2𝑥2 = 40 -------------(1)

3𝑥1 + 𝑥2 = 30 -------------(2)

4x1 + 3x2 = 60, -------------(3)

Put 𝑥2 = 0 in equation (1)

 𝑥1 = 40
At (40, 0) equation (1) meets 𝑥1 -axis

Put 𝑥1 = 0 in equation (1)

 𝑥2 = 20
At (0, 20) equation (1) meets 𝑥2 -axis

Put 𝑥2 = 0 in equation (2)

 𝑥1 = 10
At (10, 0) equation (2) meets 𝑥1 -axis

Put 𝑥1 = 0 in equation (2)

 𝑥2 = 30
At (0, 30) equation (2) meets 𝑥2 -axis

Put 𝑥2 = 0 in equation (3)


 𝑥1 = 15
At (15, 0) equation (2) meets 𝑥1 -axis

Put 𝑥1 = 0 in equation (3)

 𝑥2 = 20
At (0, 20) equation (3) meets 𝑥2 -axis
Here ABCD is the feasible region where A(15,0), B(40,0),C(4,18) ,D(6,12)

Here we get point C by solving (1) and (2) and by solving (2) and (3) we get D.

Min Z at A(15,0)=300

Min Z at B(40,0)=800

Min Z at C(4,18) =260

Min Z at D(6,12)=240

∴ Min Z =240

And 𝑥1 = 6, 𝑥2 = 12

General LPP

The problem of determining 𝑥1 , 𝑥2 , 𝑥3 , − − − − −𝑥𝑛 (maximize/minimize) and satisfies the


constraints and non-negative restrictions are called general LPP.

Let Z be the linear function defined by

𝒁 = 𝒄𝟏 𝒙𝟏 + 𝒄𝟐 𝒙𝟐 + − − − − − − + 𝒄𝒏 𝒙𝒏

Where 𝑐1 , 𝑐2 , 𝑐3 , − − − − −𝑐𝑛 are constants.

Let 𝑎𝑖𝑗 be an m × n matrix .

Let {𝑏1 , 𝑏2 , 𝑏3 , − − − − −𝑏𝑚 } be a set of constants such that

𝒂𝟏𝟏 𝒙𝟏 + 𝒂𝟏𝟐 𝒙𝟐 + − − − − − − + 𝒂𝟏𝒏 𝒙𝒏 (≤, =, ≥) 𝒃𝟏

𝒂𝟐𝟏 𝒙𝟏 + 𝒂𝟐𝟐 𝒙𝟐 + − − − − − − + 𝒂𝟐𝒏 𝒙𝒏 (≤, =, ≥) 𝒃𝟐

−−−−−−−−−−−−−−−−−−−−−−−−−

𝒂𝒎𝟏 𝒙𝟏 + 𝒂𝒎𝟐 𝒙𝟐 + − − − − − − + 𝒂𝒎𝒏 𝒙𝒏 (≤, =, ≥) 𝒃𝒎

𝒂𝒏𝒅 𝒙𝒊 ≥ 𝟎 , i = 1,2,3, − − − − −n

Solution to LPP

A set of {𝑥1 , 𝑥2 , 𝑥3 , − − − − −𝑥𝑛 } satisfying the constraints of general LPP is called solution
to LPP.
Feasible Solution

Any solution which is also satisfies non-negative restriction of the problem is called feasible
solution.

Basic Solution

If there are m equality constraints and m+n is the number of variables(m≤n), a start for the
optimal solution is made by putting n unknowns(variables) equal to zero and then solving for m
equations in remaining m unknowns. The ‘n’ zero variables are called non basic variables and
remaining m variables are called basic variables which form a basic solution.

Basic Feasible Solution (BFS)

A Feasible solution which also a basic solution to LPP is called Basic Feasible Solution .It is
of two types.

i) Degenerate BFS: If any of the basic variables are zero then that solution is degenerate BFS.

ii) Non-Degenerate BFS: If all m variables are positive then that solution is Non-Degenerate
BFS.

Optimum Basic Feasible Solution:

A Basic feasible solution which optimizes the objective function.

Simplex Method

In solving LPP we find the set of non-negative values which satisfies the constraints and
objective function. If we have a set of m equations with (m+n) unknowns then an infinite set of
solutions are possible. A hit & trail for optimal solution is not feasible, there is one efficient &
systematic procedure which gives solution in a finite number of trails known as simplex method
or simplex technique.

Slack variable:

Let constraints of LPP are in the form

∑𝑛𝑗=1 𝑎𝑖𝑗 𝑥𝑗 ≤ 𝑏𝑖 , 𝑖 = 1,2, − − − − 𝑛

𝑥𝑛+𝑖 which satisfies

∑𝑛𝑗=1 𝑎𝑖𝑗 𝑥𝑗 + 𝑥𝑛+𝑖 = 𝑏𝑖 is known as slack variable.

Surplus variable:
Let constraints of LPP are in the form

∑𝑛𝑗=1 𝑎𝑖𝑗 𝑥𝑗 ≥ 𝑏𝑖 , 𝑖 = 1,2, − − − − 𝑚

Then a non-negative restriction 𝑥𝑛+𝑖 which satisfies

∑𝑛𝑗=1 𝑎𝑖𝑗 𝑥𝑗 − 𝑥𝑛+𝑖 = 𝑏𝑖 is known as surplus variable.

Standard form of LPP

𝑀𝑎𝑥 𝑍 = 𝑐1 𝑥1 + 𝑐2 𝑥2 + − − − − − − + 𝑐𝑛 𝑥𝑛

Subject to Constraints : 𝑎𝑖1 𝑥1 + 𝑎𝑖2 𝑥2 + − − − − − − + 𝑎𝑖𝑛 𝑥𝑛 = 𝑏𝑖 , 𝑖 = 1,2, − − − − 𝑛

and 𝑥1 , 𝑥2 , − − − − −𝑥𝑛 ≥ 0

Characteristics of Simplex method

• The objective function should be maximization type ,if it is minimization convert by using the
formula

𝑀𝑎𝑥𝑧1 = −𝑚𝑖𝑛𝑧 where 𝑧1 = −𝑧

• All inequality constraints must be expressed as equalities by introducing slack or surplus


variables.
• The RHS constraints must be positive.
• All the variables must be non-negative. if any variable 𝑥𝑗 is unrestricted replace it by
(𝑥𝑗1 − 𝑥𝑗11 ) where 𝑥𝑗1 − 𝑥𝑗11 ≥ 0

Simplex Algorithm

The existence of initial basic feasible solution is always assumed the steps for competition of
an optimum solution given below.

Step-1: Check whether the objective function of the given LPP is to be maximized or minimized.
If it is to be minimized, we have to convert into maximization using the formula

𝑀𝑎𝑥𝑧1 = −𝑚𝑖𝑛𝑧 where 𝑧1 = −𝑧

Step-2: Check whether all 𝑏𝑖 (𝑖 =1,2,3,-------m) are positive ,if any 𝑏𝑖 is negative then multiply
the equation with -1 on both sides.

Step-3: Express the problem in standard form by introducing slack or surplus variables to convert
inequalities into equalities.
Step-4: Let 𝑋𝐵 be Initial Basic Feasible Solution to the LPP

𝑀𝑎𝑥 𝑍 = 𝐶𝑋

𝐴𝑋 = 𝑏

𝑋≥0

Such that it satisfies 𝑋𝐵 = 𝐵 −1 𝑏 , where b is the basis matrix formed by the basic variables.

Now we find initial simplex table given below.

𝑐1 𝑐2 𝑐3 𝑐𝑛 0

𝐶𝐵 𝐵𝑉 𝑋𝐵 𝑥1 𝑥2 𝑥3 𝑥𝑛 𝑥𝑛+1

Step-5: Compute the evaluation 𝑧𝑗 using the formula

𝑍𝑗 = ∑ 𝐶𝐵𝑖 𝑎𝑖𝑗 , 𝑗 = 1,2, − − − − 𝑛 + 𝑚


𝑖=1

Net evaluation (𝑍𝑗 − 𝐶𝑗 )

i) If all (𝑍𝑗 − 𝐶𝑗 ) ≥ 0 then the initial basic feasible solution 𝑋𝐵 is an optimum basic feasible
solution.

ii) If at least one (𝑍𝑗 − 𝐶𝑗 ) < 0 then proceed to next step.

Step-6: To find the entering variable

If there are more than one negative (𝑍𝑗 − 𝐶𝑗 ). Choose the most negative of them.

Let (𝑍𝑘 − 𝐶𝑘 ) be the most negative of for j=k, this gives the entering variable 𝑋𝑘 and indicate it
by arrow ↑ at the bottom of kth column.

If there are more than one variable having the same most negative (𝑍𝑗 − 𝐶𝑗 ) . Choose any one of
them arbitrary as entering variable.

Step-7: To find the leaving variable


𝑋𝐵𝑖
Find the ratio of the minimum among these will leave the basis. the element at the
𝑋𝑘
intersection of entering variable and leaving row is called pivot element (key element).

Step-8: For new basis by dropping the leaving variable and introducing the entering variable
along with associated value under 𝐶𝐵 column .convert the pivot element to 1 and the remaining
elements in that corresponding column to zero.

Step-9: Go to step-5 and repeat the process until either an optimum solution is obtained or there
is an indication of unbalanced solution.

Solved problems on finding standard form of LPP

1. Express following LPP in standard form


𝑴𝒂𝒙 𝒁 = 𝟒𝒙𝟏 + 𝟐𝒙𝟐

Subject to

𝟐𝒙𝟏 + 𝟑𝒙𝟐 ≥ 𝟔,

𝟑𝒙𝟏 + 𝟒𝒙𝟐 ≤ 𝟖,

𝒙𝟏 , 𝒙𝟐 ≥ 𝟎

Answer.

Standard form of given LPP is

If the inequality is ≥ type we in constraints we subtract Surplus variable (Let it be 𝒙𝟑 ). Therefore


the first constraint becomes

𝟐𝒙𝟏 + 𝟑𝒙𝟐 − 𝒙𝟑 = 𝟔

If the inequality is ≤ type we in constraints we add Slack variable (Let it be 𝒙𝟒 ). Therefore the
second constraint becomes

𝟑𝒙𝟏 + 𝟒𝒙𝟐 + 𝒙𝟒 = 𝟖

And non negative restrictions are 𝒙𝟏 , 𝒙𝟐 , 𝒙𝟑 , 𝒙𝟒 ≥ 𝟎

And new objective function is 𝑴𝒂𝒙 𝒁 = 𝟒𝒙𝟏 + 𝟐𝒙𝟐 + 𝟎. 𝒙𝟑 + 𝟎. 𝒙𝟒

Therefore the standard form of given LPP is

𝑴𝒂𝒙 𝒁 = 𝟒𝒙𝟏 + 𝟐𝒙𝟐 + 𝟎. 𝒙𝟑 + 𝟎. 𝒙𝟒


Subject to

𝟐𝒙𝟏 + 𝟑𝒙𝟐 − 𝒙𝟑 = 𝟔,

𝟑𝒙𝟏 + 𝟒𝒙𝟐 + 𝒙𝟒 = 𝟖

𝒙𝟏 , 𝒙𝟐 , 𝒙𝟑 , 𝒙𝟒 ≥ 𝟎

2. Express following LPP in standard form


𝑴𝒊𝒏 𝒁 = 𝟐𝒙𝟏 + 𝟑𝒙𝟐 + 𝒙𝟑

Subject to

𝟔𝒙𝟏 + 𝟒𝒙𝟐 − 𝒙𝟑 ≤ 𝟒,

𝟕𝒙𝟏 + 𝟑𝒙𝟐 + 𝟐𝒙𝟑 ≥ −𝟒

𝒙𝟏 + 𝒙𝟐 + 𝒙𝟑 ≤ 𝟓,

𝒙𝟏 , 𝒙𝟐 , 𝒙𝟑 ≥ 𝟎

Answer.

Standard form of given LPP is

If the objective function is Minimization case we convert it into Maximization by following


formula

𝑀𝑎𝑥 𝑧 / = −𝑀𝑖𝑛 𝑧

There fore

𝑀𝑎𝑥 𝑧 / = −(2𝑥1 + 3𝑥2 + 𝑥3 )

6𝑥1 + 4𝑥2 − 𝑥3 + 𝑥4 = 4,

7𝑥1 + 3𝑥2 + 2𝑥3 − 𝑥5 = −4  − 7𝑥1 − 3𝑥2 − 2𝑥3 + 𝑥5 = 4,

𝑥1 + 𝑥2 + 𝑥3 + 𝑥6 = 5,

𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 , 𝑥5 , 𝑥6 ≥ 0

New objective function is

𝑀𝑎𝑥 𝑧 / = −2𝑥1 − 3𝑥2 − 𝑥3 + 0. 𝑥4 + 0. 𝑥5 + 0. 𝑥6


Solved problems on Simplex method
1. Solve the following LPP by Simplex method

𝑴𝒂𝒙 𝒁 = 𝟔𝒙𝟏 + 𝟕𝒙𝟐

Subject to

𝟐𝒙𝟏 + 𝟑𝒙𝟐 ≤ 𝟏𝟐,

𝟐𝒙𝟏 + 𝒙𝟐 ≤ 𝟖,

𝒙𝟏 , 𝒙𝟐 ≥ 𝟎

Solution. Standard form of given LPP is

2𝑥1 + 3𝑥2 + 𝑥3 = 12

2𝑥1 + 𝑥2 + 𝑥4 = 8

𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 ≥ 0

New objective function is

Max Z = 6x1 + 7x2 + 0. x3 + 0. x4

The Initial Basic Feasible Solution (IBFS) is


𝑥3
𝑋𝐵 = [𝑥 ] = 𝐵𝑏 −1
4

1 0 12 12
=[ ] [ ]=[ ]
0 1 8 8

 𝑥3 = 12, 𝑥4 = 8
Simplex starting table is

𝑪𝒋 6 7 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 =
𝑿𝒌 𝑿𝟐
𝟏𝟐
0 𝒙𝟑 12 2 𝟑 1 0 𝟑
= 𝟒(Min)←
𝟖
0 𝒙𝟒 8 2 1 0 1 =𝟖
𝟏
Z=𝑪𝑩 𝑿𝑩 =0 0 0 0 0 𝒁𝒋
-6 -7 0 0 𝒁𝒋 − 𝑪𝒋

Here 𝐶𝐵 is cost of basic variable

𝑋𝐵 is Initial Basic Feasible Solution

The intersection of leaving variable and entering variable is known as Pivot element (Indicated in
Red color).

Here 𝑥2 is entering variable

To get first iteration table introduce 𝑥2 , leave 𝑥3


𝑅1
New 𝑅1 → 𝑂𝑙𝑑 3

New 𝑅2 → 𝑂𝑙𝑑 𝑅2 − New 𝑅1

First iteration table is

𝑪𝒋 6 7 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 =
𝑿𝒌 𝑿𝟏
𝟒
7 𝒙𝟐 4 2/3 𝟏 1/3 0 𝟐/𝟑
=𝟔
𝟒
0 𝒙𝟒 4 4/3 0 -1/3 1 𝟒/𝟑
=𝟑←

Z=𝑪𝑩 𝑿𝑩 =28 14/3 7 7/3 0 𝒁𝒋


-4/3 0 7/3 0 𝒁𝒋 − 𝑪𝒋

Second iteration table is

𝑪𝒋 6 7 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 =
𝑿𝒌 𝑿𝟏
7 𝒙𝟐 2 0 1 1/2 -1/2
6 𝒙𝟏 3 1 0 -1/4 3/4
Z=𝑪𝑩 𝑿𝑩 =32 6 7 2 1 𝒁𝒋
0 0 2 1 𝒁𝒋 − 𝑪𝒋

Here all 𝑍𝑗 − 𝐶𝑗 values are positive.

∴ Max z=32, 𝒙𝟏 = 𝟑, 𝒙𝟐 = 𝟐
2. Solve the following LPP by Simplex method

𝑴𝒊𝒏 𝒁 = 𝒙𝟏 − 𝟑𝒙𝟐 + 𝟐𝒙𝟑

Subject to

𝟑𝒙𝟏 − 𝒙𝟐 + 𝟐𝒙𝟑 ≤ 𝟕,

−𝟐𝒙𝟏 + 𝟒𝒙𝟐 ≤ 𝟏𝟐,

−𝟒𝒙𝟏 + 𝟑𝒙𝟐 + 𝟖𝒙𝟑 ≤ 𝟏𝟎

𝒙𝟏 , 𝒙𝟐 , 𝒙𝟑 ≥ 𝟎

Solution. Standard form of given LPP is

3𝑥1 − 𝑥2 + 2𝑥3 + 𝑥4 = 7

−2x1 + 4x2 + x5 = 12

−4x1 + 3x2 + 8x3 + x6 = 10

𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 , x5 , x6 ≥ 0

New objective function is

𝑀𝑎𝑥 𝑧 / = −𝑥1 + 3𝑥2 − 2𝑥3 + 0. 𝑥4 + 0. 𝑥5 + 0. 𝑥6

The Initial Basic Feasible Solution (IBFS) is


𝑥4
𝑋𝐵 = [𝑥5 ] = 𝐵𝑏 −1
𝑥6

1 0 0 7
12
=[0 1 0] [12] = [ ]
8
0 0 1 10

 𝑥4 = 7, 𝑥5 = 12, 𝑥6 = 10
Simplex starting table is
𝑪𝒋 -1 3 -2 0 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 𝑥5 𝑥6 =
𝑿𝒌 𝑿𝟐
0 𝒙𝟒 7 3 -1 2 1 0 0 ------
1 0 𝟏𝟐
0 𝑥5 12 -2 4 0 0 =𝟑
𝟒
0 1 𝟏𝟎
0 𝑥6 10 -4 3 8 0 = 𝟑. 𝟑𝟑
𝟑
𝑧/ 0 0
𝒁𝒋
0 0 0 0
=𝑪𝑩 𝑿𝑩 =0
1 -3 2 0 0 0 𝒁𝒋 − 𝑪𝒋

First iteration table is

𝑪𝒋 -1 3 -2 0 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 𝑥5 𝑥6 =
𝑿𝒌 𝑿𝟏
0 𝒙𝟒 0 5/2 0 2 1 1/4 0 4
3 𝒙𝟐 3 -1/2 1 0 0 1/4 0 −−

0 𝑥6 1 -5/2 0 8 0 -3/4 1 −−

𝑧/ 3/4 0
𝒁𝒋
-3/2 3 0 0
=𝑪𝑩 𝑿𝑩 =9
-1/2 0 2 0 3/4 0 𝒁𝒋 − 𝑪𝒋

Second iteration table is

𝑪𝒋 -1 3 -2 0 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 𝑥5 𝑥6 =
𝑿𝒌 𝑿𝟏
-1 𝒙𝟏 4 1 0 4/5 2/5 1/10 0

3 𝒙𝟐 5 0 1 2/5 1/5 3/10 0

0 𝑥6 11 0 0 10 1 -1/2 1

𝑧/ 4/5 0
𝒁𝒋
-1 3 10 1
=𝑪𝑩 𝑿𝑩 =11
0 0 12/5 1/5 4/5 0 𝒁𝒋 − 𝑪𝒋
Here all 𝑍𝑗 − 𝐶𝑗 values are positive.

𝑧 / = 11

∴ 𝑍 = −11

∴ Min z=-11

𝒙𝟏 = 𝟒, 𝒙𝟐 = 𝟓, 𝒙𝟑 = 𝟎

Exercise
1. Solve the following LPP by Simplex method

𝑴𝒂𝒙 𝒁 = 𝟑𝒙𝟏 + 𝟐𝒙𝟐

Subject to

𝒙𝟏 + 𝒙𝟐 ≤ 𝟒,

𝒙𝟏 − 𝒙𝟐 ≤ 𝟐,

𝒙𝟏 , 𝒙𝟐 ≥ 𝟎

2. Solve the following LPP by Simplex method

𝑴𝒂𝒙 𝒁 = 𝟕𝒙𝟏 + 𝟏𝟒𝒙𝟐

Subject to

𝟑𝒙𝟏 + 𝟐𝒙𝟐 ≤ 𝟑𝟔,

𝒙𝟏 + 𝟒𝒙𝟐 ≤ 𝟏𝟎

𝒙𝟏 , 𝒙𝟐 ≥ 𝟎

Artificial Variable Technique

LPP may have constraints of ≥ type .In such cases the starting simplex table will not contain
an identity matrix. We introduce a new type of variable called artificial variable. These variables
are imaginary and cannot have any physical meanings. These variables can be eliminated when
they left the basis.

a) Big-M method (Method of penalty)

b) Two phase Simplex method


a) Big-M method (Method of penalty):

Proceure:
Step-1: Express the LPP in standard form by introducing slack/surplus or artificial variables.
Assign zero co-efficient to surplus variables and large penalty to artificial variable in the
objective function.

Step-2: Solve modified LPP by simplex method at any iteration of any simplex method any one
of the following cases may arise.

I) If at least one artificial variable is present in the basis when all (𝑍𝑗 − 𝐶𝑗 ) ≥ 0 then the given
LPP does not have any solution.

II) If there is no artificial variable in the basis then the Simple method is continued until optimal
solution has been obtained.

1. Solve the following LPP by Big-M method

𝑴𝒂𝒙 𝒁 = 𝟔𝒙𝟏 + 𝟒𝒙𝟐

Subject to

𝟐𝒙𝟏 + 𝟑𝒙𝟐 ≤ 𝟑𝟎,

𝟑𝒙𝟏 + 𝟐𝒙𝟐 ≤ 𝟐𝟒,

𝒙𝟏 + 𝒙𝟐 ≥ 𝟑

𝒙𝟏 , 𝒙𝟐 ≥ 𝟎

Solution. Standard form of given LPP is

2𝑥1 + 3𝑥2 + 𝑥3 = 30

3x1 + 2x2 + 𝑥4 = 24

x1 + x2 − x5 + a1 = 3

𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 , x5 , a1 ≥ 0

New objective function is

Max Z = 6𝑥1 + 4𝑥2 + 0. 𝑥3 + 0. 𝑥4 + 0. 𝑥5 − 𝑀a1 .


The Initial Basic Feasible Solution (IBFS) is
𝑥3
𝑋𝐵 = [𝑥4 ] = 𝐵𝑏 −1
a1

1 0 0 30 30
=[0 ] [
1 0 24 ] = [ 24]
0 0 1 3 3
 𝑥3 = 30, 𝑥4 = 24, a1 = 3

Simplex starting table is

𝑪𝒋 6 4 0 0 0 -M
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 𝑥5 a1 =
𝑿𝒌 𝑿𝟏
0 𝑥3 30 2 3 1 0 0 0 15
0 0 8
0 𝑥4 24 3 2 0 1
-1 1 𝟑←
-M a1 3 1 1 0 0
Z M -M
=𝑪𝑩 𝑿𝑩 = -M -M 0 0 𝒁𝒋
-3M
-M- M 0
-M-4 0 0 𝒁𝒋 − 𝑪𝒋
6

First iteration table is

𝑪𝒋 6 4 0 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 𝑥5 =
𝑿𝒌 𝑿𝟏
0 𝑥3 24 0 1 1 0 2 12
0 𝑥4 15 0 -1 0 1 3 5←

6 𝒙𝟏 3 1 1 0 0 -1 ---

Z -6
6 6 0 0 𝒁𝒋
=𝑪𝑩 𝑿𝑩 =18
0 2 0 0 -6 𝒁𝒋 − 𝑪𝒋

Second iteration table is

𝑪𝒋 6 4 0 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 𝑥5 =
𝑿𝒌 𝑿𝟏
0 𝑥3 14 0 5/3 1 -2/3 0

0 𝑥5 5 0 -1/3 0 1/3 1

6 𝒙𝟏 8 1 2/3 0 1/3 0

Z=𝑪𝑩 𝑿𝑩 =48 6 4 0 2 0 𝒁𝒋
0 0 0 2 0 𝒁𝒋 − 𝑪𝒋

Here all 𝑍𝑗 − 𝐶𝑗 values are positive.

𝐙 = 48

∴ Max z=48

𝒙𝟏 = 𝟖, 𝒙𝟐 = 𝟎𝟓

Exercise:
1. Solve the following LPP by Big-M method(Method of Penalities)
𝑴𝒂𝒙 𝒁 = 𝟒𝒙𝟏 + 𝟔𝒙𝟐

Subject to

𝒙𝟏 + 𝟐𝒙𝟐 ≤ 𝟐,

𝟖𝒙𝟏 + 𝟔𝒙𝟐 ≥ 𝟐𝟒,

𝒙𝟏 , 𝒙𝟐 ≥ 𝟎

b) Two phase Simplex method


Phase-I: Formulate any objective function 𝑍 ∗ by assigning zero cost to all variables and -1 to
artificial variables.

i.e, 𝑍 ∗ = 0. 𝑥1 + 0. 𝑥2 + − − − − − − −𝑎1 − 𝑎2

Now apply simplex method.


If (𝑍𝑗 − 𝐶𝑗 ) ≥ 0 ,any one of the following cases arises.

i)𝑀𝑎𝑥𝑍 ∗ < 0 and artificial variable at the level in this case given LPP has no solution.

ii) 𝑀𝑎𝑥𝑍 ∗ = 0 and no artificial variable appears in the basis then proceed to phase-II.

iii)𝑀𝑎𝑥𝑍 ∗ = 0 and artificial variable appears in the basis at zero level then proceed to phase-II.

Phase-II: Assign original cost to variables in the objective function and solve by simplex method
for this phase. We use the table obtained at the end of the phase-I.

Solved problems

1. Solve the following LPP by Two phase Simplex method


𝑴𝒂𝒙 𝒁 = 𝟒𝒙𝟏 + 𝟔𝒙𝟐

Subject to

𝒙𝟏 + 𝟐𝒙𝟐 ≤ 𝟐,

𝟖𝒙𝟏 + 𝟔𝒙𝟐 ≥ 𝟐𝟒,

𝒙𝟏 , 𝒙𝟐 ≥ 𝟎

Solution. Standard form of given LPP is

𝑥1 + 2𝑥2 + 𝑥3 = 2

8x1 + 6x2 − 𝑥4 + a1 = 24

𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 , a1 ≥ 0

New objective function is

Max Z ∗ = 0. 𝑥1 + 0. 𝑥2 + 0. 𝑥3 + 0. 𝑥4 − a1 .

The Initial Basic Feasible Solution (IBFS) is

𝑥3 = 2, a1 = 24
Simplex starting table is

𝑪𝒋 0 0 0 0 -1
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 a1 =
𝑿𝒌 𝑿𝟏
0 𝑥3 2 1 2 1 0 0 2←
1 3
-1 a1 24 8 6 0 -1
Z
=𝑪𝑩 𝑿𝑩 = -8 -6 0 1 -1 𝒁𝒋
-24
-8 -6 0 1 0 𝒁𝒋 − 𝑪𝒋

First iteration table is

𝑪𝒋 0 0 0 0 -1
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 a1 =
𝑿𝒌 𝑿𝟏
0 𝑥1 2 1 2 1 0 0 2←
1 3
-1 a1 8 0 -10 -8 -1
Z -1
=𝑪𝑩 𝑿𝑩 = 0 10 8 1 𝒁𝒋
-8
0 10 8 1 0 𝒁𝒋 − 𝑪𝒋
Here all 𝑍𝑗 − 𝐶𝑗 greater than or equal to zero,but an artificial variable is present in the basis(i.e
a1 ).

∴ Given LPP has no solution.

2. Solve the following LPP by Two phase Simplex method

𝑴𝒊𝒏 𝒁 = (𝟏𝟓/𝟐)𝒙𝟏 − 𝟑𝒙𝟐

Subject to

𝟑𝒙𝟏 − 𝒙𝟐 − 𝒙𝟑 ≥ 𝟎,

𝒙𝟏 − 𝒙𝟐 + 𝒙𝟑 ≥ 𝟐,

𝒙𝟏 + 𝒙𝟐 ≥ 𝟑
𝒙𝟏 , 𝒙𝟐 , 𝒙𝟑 ≥ 𝟎

Solution. Standard form of given LPP is

Max Z ∗ = −(𝟏𝟓/𝟐)𝒙𝟏 + 𝟑𝒙𝟐

3𝑥1 − 𝑥2 − 𝑥3 − 𝑥4 + 𝑎1 = 3

x1 − x2 − x3 − 𝑥5 + a2 = 2

𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 , x5 , a1 , a2 ≥ 0

New objective function is

Max Z ∗ = 0. 𝑥1 + 0. 𝑥2 + 0. 𝑥3 + 0. 𝑥4 + 0. 𝑥5 − a1 − a2

The Initial Basic Feasible Solution (IBFS) is

a1 = 3, a2 = 2

Simplex starting table is

𝑪𝒋 0 0 0 0 0 -1 -1
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 𝑥5 a1 a2
=
𝑿𝒌 𝑿𝟏
-1 a1 3 3 -1 -1 -1 0 1 0 1←
-1 a2 2 1 -1 1 0 -1 0 1 2

Z
=𝑪𝑩 𝑿𝑩 =
-4 2 0 1 1 -1 -1 𝒁𝒋
-5

-4 2 0 1 1 0 0 𝒁𝒋 − 𝑪𝒋

First iteration table is

𝑪𝒋 0 0 0 0 0 -1
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 𝑥5 a2
=
𝑿𝒌 𝑿𝟑
0 𝒙𝟏 1 1 -1/3 -1/3 -1/3 0 0 ---
-1 a2 1 0 -2/3 4/3 1/3 -1 1 𝟑/𝟒 ←

Z
=𝑪𝑩 𝑿𝑩 =.
0 2/3 -4/3 -1/3 1 -1 𝒁𝒋
-1

0 2/3 -4/3 -1/3 1 0 𝒁𝒋 − 𝑪𝒋


Second iteration table is

𝑪𝒋 0 0 0 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 𝑥5 =
𝑿𝒌 𝑿𝟑
0 𝒙𝟏 5/4 1 -1/2 0 -1/4 -1/4

0 𝒙𝟑 3/4 0 -1/2 1 1/4 -3/4

Z
=𝑪𝑩 𝑿𝑩 =.
0 0 0 0 0 𝒁𝒋
-1

0 0 0 0 0 𝒁𝒋 − 𝑪𝒋

Here all 𝑍𝑗 − 𝐶𝑗 greater than or equal to zero and no artificial variable in the basis.

Max Z ∗ = 0

So we have to proceed for Phase-II

Phase-II
New objective function is
15
Max Z ∗ = −( 2 ). 𝑥1 + 3. 𝑥2 + 0. 𝑥3 + 0. 𝑥4 + 0. 𝑥5
𝑪𝒋 -(15/2) 3 0 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 𝑥5 =
𝑿𝒌 𝑿𝟑
-15/2 𝒙𝟏 5/4 1 -1/2 0 -1/4 -1/4

0 𝒙𝟑 3/4 0 -1/2 1 1/4 -3/4

Z
=𝑪𝑩 𝑿𝑩 =.
-(15/2) 15/4 0 15/8 15/8 𝒁𝒋
-(75/8)

0 3/4 0 15/8 15/8 𝒁𝒋 − 𝑪𝒋

Here all 𝑍𝑗 − 𝐶𝑗 ≥ 0

75
Max Z ∗ = −( )
8
75
Min Z= 8

𝟐
𝒙𝟏 = 𝟒 , 𝒙𝟐 = 𝟎

Exercise:

1. Solve the following LPP by Two phase Simplex method


𝑴𝒂𝒙 𝒁 = 𝟐𝒙𝟏 + 𝒙𝟐 + (𝟏/𝟒) 𝒙𝟑

Subject to

𝟒 + 𝟔𝒙𝟐 + 𝟑𝒙𝟑 ≤ 𝟎,

𝟑𝒙𝟏 − 𝟔𝒙𝟐 − 𝟒𝒙𝟑 ≤ 𝟏,

𝟐𝒙𝟏 + 𝟑𝒙𝟐 − 𝟓𝒙𝟑 ≥ 𝟒

𝒙𝟏 , 𝒙𝟐 , 𝒙𝟑 ≥ 𝟎
Special cases of Simplex method
Degeneracy:

At the stage of improving the solution during simplex procedure minimum ratio is
determined in the last column of simplex table to find the leaving variable. If the minimum ratio
is same then the problem is said to be degenerate. The following procedure is to resolve
degeneracy.

i) Pick up the rows for which the minimum ratio is same.

ii) Re arrange the columns forming identity matrix comes in first order.
𝐸𝑙𝑒𝑚𝑒𝑛𝑡𝑠 𝑜𝑓 1𝑠𝑡 𝑐𝑜𝑙𝑢𝑚𝑛 𝑜𝑓 𝑢𝑛𝑖𝑡 𝑚𝑎𝑡𝑟𝑖𝑥
iii) Find the minimum ratio {𝑐𝑜𝑟𝑒𝑠𝑝𝑜𝑛𝑑𝑖𝑛𝑔 𝑒𝑙𝑒𝑚𝑒𝑛𝑡𝑠 𝑜𝑓 𝑒𝑛𝑡𝑒𝑟𝑖𝑛𝑔 𝑐𝑜𝑙𝑢𝑚𝑛 }

only for rows for which the minimum ratio is same.

iv) If the minimum ratio is also same then find the ratio of
𝐸𝑙𝑒𝑚𝑒𝑛𝑡𝑠 𝑜𝑓 2𝑛𝑑 𝑐𝑜𝑙𝑢𝑚𝑛 𝑜𝑓 𝑢𝑛𝑖𝑡 𝑚𝑎𝑡𝑟𝑖𝑥
{𝑐𝑜𝑟𝑒𝑠𝑝𝑜𝑛𝑑𝑖𝑛𝑔 𝑒𝑙𝑒𝑚𝑒𝑛𝑡𝑠 𝑜𝑓 𝑒𝑛𝑡𝑒𝑟𝑖𝑛𝑔 𝑐𝑜𝑙𝑢𝑚𝑛 } , if the ratio is again same, repeat the above procedure
till we get unique minimum ratio.

Note: There is no need to apply the procedure when a tie for artificial variable and other
variables occurred.

1. Solve the following Simplex method


𝑴𝒂𝒙 𝒁 = 𝟐𝒙𝟏 + 𝒙𝟐

Subject to

𝟒𝒙𝟏 + 𝟑𝒙𝟐 ≤ 𝟏𝟐,

𝟒𝒙𝟏 + 𝟑𝒙𝟐 ≤ 𝟖,

𝟒𝒙𝟏 − 𝒙𝟐 ≤ 𝟖,

𝒙𝟏 , 𝒙𝟐 ≥ 𝟎

Solution. Standard form of given LPP is

4𝑥1 + 3𝑥2 + 𝑥3 = 12

4x1 + x2 + 𝑥4 = 8

𝟒𝒙𝟏 − 𝒙𝟐 + 𝑥5 = 𝟖
𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 , x5 ≥ 0

New objective function is

Max z = 2𝑥1 + 𝑥2 + 0. 𝑥3 + 0. 𝑥4 − 0. x5 .

The Initial Basic Feasible Solution (IBFS) is

𝑥3 = 12, x4 = 8, 𝑥5 = 8

Simplex starting table is

𝑪𝒋 2 1 0 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 x5 =
𝑿𝒌 𝑿𝟏
0 𝑥3 12 4 3 1 0 0 𝟑
0 2
0 x4 8 4 1 0 1
1 2
0 𝑥5 8 4 -1 0 0
Z 0
0 =𝑪𝑩 𝑿𝑩 = 0 0 0 0 𝒁𝒋
0
-2 -1 0 0 0 𝒁𝒋 − 𝑪𝒋

First iteration table is

𝑪𝒋 2 1 0 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟑 𝒙𝟒 x5 𝒙𝟏 𝒙𝟐 =
𝑿𝒌 𝑿𝟏
0 𝑥3 12 1 0 0 4 3 --
1/4
0 x4 8 0 1 0 4 1
0←
0 𝑥5 8 0 0 1 4 -1
Z 0
=𝑪𝑩 𝑿𝑩 = 0 0 0 0 𝒁𝒋
0
0 0 0 -2 -1 𝒁𝒋 − 𝑪𝒋

Second iteration table is

𝑪𝒋 2 1 0 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟑 𝒙𝟒 x5 𝒙𝟏 𝒙𝟐 𝒙𝒌
=
𝒙𝟓
0 𝑥3 4 1 0 -1 0 4 1
0←
1 x2 0 0 1 -1 0 2
---
2 𝑥1 2 0 0 1/4 1 -1/4
Z -1/2
=𝑪𝑩 𝑿𝑩 = 0 0 1/2 2 𝒁𝒋
4
0 0 1/2 0 -3/2 𝒁𝒋 − 𝑪𝒋

Third iteration table is

𝑪𝒋 2 1 0 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟑 𝒙𝟒 x5 𝒙𝟏 𝒙𝟐 =
𝑿𝒌 𝑿𝟏
4
0 𝑥3 4 1 -2 1 0 0

---
1 x2 0 0 1/2 -1/2 0 1
16
2 𝑥1 2 0 1/8 1/8 1 0
Z 1
=𝑪𝑩 𝑿𝑩 = 0 3/4 -1/4 2 𝒁𝒋
4
0 3/4 -1/4 0 0 𝒁𝒋 − 𝑪𝒋

Fourth iteration table is

𝑪𝒋 2 1 0 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟑 𝒙𝟒 x5 𝒙𝟏 𝒙𝟐 =
𝑿𝒌 𝑿𝟏
0 𝑥5 4 1 -2 1 0 0

1 x2 2 1/2 -1/2 0 0 1

2 𝑥1 3/2 -1/8 3/8 0 1 0


Z 1
=𝑪𝑩 𝑿𝑩 = 1/4 1/4 0 2 𝒁𝒋
5
1/4 1/4 0 0 0 𝒁𝒋 − 𝑪𝒋

Here all 𝑍𝑗 − 𝐶𝑗 ≥ 0

Max Z = 5
𝟑
𝒙𝟏 = 𝟐 , 𝒙𝟐 = 𝟐
Unbounded Solution

1. Solve the following LPP by Simplex method


𝑴𝒂𝒙 𝒁 = 𝟐𝒙𝟏 + 𝟑𝒙𝟐 + 𝟒 𝒙𝟑

Subject to

−𝒙𝟏 − 𝟓𝒙𝟐 − 𝟗𝒙𝟑 ≤ 𝟐,

𝟑𝒙𝟏 − 𝒙𝟐 + 𝒙𝟑 ≤ 𝟏𝟎,

𝟐𝒙𝟏 + 𝟑𝒙𝟐 − 𝟕𝒙𝟑 ≤ 𝟐

𝒙𝟏 , 𝒙𝟐 , 𝒙𝟑 ≥ 𝟎

Solution. Standard form of given LPP is

−𝑥1 − 5𝑥2 − 9𝑥3 + 𝑥4 = 2

3x1 − x2 + x3 + x5 = 12

2x1 + 3x2 − 7x3 + x6 = 2

𝑥1 , 𝑥2 , 𝑥3 , 𝑥4 , x5 , x6 ≥ 0

New objective function is

𝑀𝑎𝑥 𝑍 = 2𝑥1 + 3𝑥2 + 4𝑥3 + 0. 𝑥4 + 0. 𝑥5 + 0. 𝑥6


The Initial Basic Feasible Solution (IBFS) is

 𝑥4 = 2, 𝑥5 = 10, 𝑥6 = 0
Simplex starting table is

𝑪𝒋 2 3 4 0 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 𝑥5 𝑥6 =
𝑿𝒌 𝑿𝟑
0 𝒙𝟒 2 -1 -5 -9 1 0 0 ------
1 0 10
0 𝑥5 10 3 -1 1 0
0 1 -------
0 𝑥6 2 2 3 -7 0
𝑍 0 0
0 0 0 0 𝒁𝒋
=𝑪𝑩 𝑿𝑩 =0
-2 -3 -4 0 0 0 𝒁𝒋 − 𝑪𝒋

First iteration table is

𝑪𝒋 2 3 4 0 0 0
𝑿𝑩 𝑿𝑩
𝑪𝑩 𝑩𝑽 𝑿𝑩 𝒙𝟏 𝒙𝟐 𝒙𝟑 𝒙𝟒 𝑥5 𝑥6 =
𝑿𝒌 𝑿𝟑
0 𝒙𝟒 92 26 -14 0 1 9 0 ------
4 𝑥3 10 3 -1 1 0
1 0 ------

7 1 -------
0 𝑥6 72 23 -4 0 0
𝑍 4 0
12 -4 4 0 𝒁𝒋
=𝑪𝑩 𝑿𝑩 =40
10 -7 0 0 4 0 𝒁𝒋 − 𝑪𝒋

Since all elements in column of pivot element are negative. Is is unbounded solution.
UNIT-II

Transportation Problem

The transportation problem is a particular case of LPP. In which the objective is to transport
various quantities of single homogeneous commodity that are initially stored at various origins in
such a way that the total transportation cost is minimum.

In a Transportation Problem the total availability should be equal to total demand. Otherwise
it is said to be unbalanced transportation problem.

Ex: A Soft drink manufacturing company has m plants situated in different cities. The total
manufacture product is observed by n retail shops located in different cities. We have to
determine the transportation schedule that minimizes the total cost of soft drink from various
plants to various retail shops.

Mathematical Formulation of Transportation Problem

Let 01 , 𝑂2 , 𝑂3 , − − −−, 𝑂𝑚 be m plants locations (origins) and 𝐷1 , 𝐷2 , 𝐷3 , − − −−, 𝐷𝑛

Are n retail shops (destinations) respectively.


Let 𝑎𝑖 be the amount of soft drink available at plant 𝑂𝑖 and 𝑏𝑗 be the amount required at jth
retail shop 𝐷𝑗.

Let 𝐶𝑖𝑗 represents unit transportation cost for transporting the unit from origin I to destination j.

The objective is to determine the number of units to be transported from origin i to destination j
so that total transportation cost is minimum.

If 𝑥𝑖𝑗 is the number of units transported from origin i to destination j , then equivalent LPP
will be

Find 𝑥𝑖𝑗 (𝑖 = 1,2, − − −𝑚, 𝑗 = 1,2, − − −𝑛) so as to be minimize

𝑍 = ∑𝑚 𝑛
𝑖=1 ∑𝑗=1 𝑥𝑖𝑗 𝑐𝑖𝑗

S to C: ∑𝑛𝑗=1 𝑥𝑖𝑗 = 𝑎𝑖

∑𝑚
𝑖=1 𝑥𝑖𝑗 = 𝑏𝑗

And 𝑥𝑖𝑗 ≥ 0

General Transportation Table

Source 𝑫𝟏 𝑫𝟐 ---------- 𝑫𝒋 ---------- 𝑫𝒏 Availability


𝑶𝟏 𝑪𝟏𝟏 𝑪𝟏𝟐 ---------- 𝑪𝟏𝒋 ---------- 𝑪𝟏𝒏 𝒂𝟏
𝑶𝟐 𝑪𝟐𝟏 𝑪𝟐𝟐 ---------- 𝑪𝟐𝒋 ---------- 𝑪𝟐𝒏 𝒂𝟐
---------- --------- --------- ---------- ---------- ---------
𝑶𝒊 𝑪𝒊𝟏 𝑪𝒊𝟐 ---------- 𝑪𝒊𝒋 ---------- 𝑪𝒊𝒏 𝒂𝒊
--------- --------- --------- ---------- ---------- ---------
𝑶𝒎 𝑪𝒎𝟏 𝑪𝒎𝟐 ---------- 𝑪𝒎𝒋 ---------- 𝑪𝒎𝒏 𝒂𝒎
Requirement 𝒃𝟏 𝒃𝟐 ---------- 𝒃𝒋 ----------- 𝒃𝒎

Note : 1) If ∑𝑚 𝑛
𝑖=1 𝑎𝑖 = ∑𝑗=1 𝑏𝑗 then the transportation problem is said to balance
Transportation Problem.

2) If ∑𝑚 𝑛
𝑖=1 𝑎𝑖 ≠ ∑𝑗=1 𝑏𝑗 then the transportation problem is said to unbalanced Transportation
Problem.

Finding IBFS

To find IBFS we have the following three methods.

1) North West Corner Method (NWC)


2) Matrix Minima Method (Least Cost Method)

3) Vogel’s Approximation Method (VAM)

1) North West Corner Method (NWC):

Procedure:

Step-1: The first assignment made in the cell occupying the upper left hand (North-West Corner)
in the table. The minimum possible amount is to be allocated as

𝑥11 = 𝑚𝑖𝑛(𝑎1 , 𝑏1 ).

This value of 𝑥11 is then entered in the cell (1, 1) of the Transportation table.

Step-2: If 𝑏1 > 𝑎1 move vertically down towards the second row and make the second
allocation of amount 𝑥12 = 𝑚𝑖𝑛(𝑎1 − 𝑥11 , 𝑏2 ) in the cell (1, 2).

If 𝑏1 < 𝑎1 move horizontally right side to the Second column and make the second
allocation of amount 𝑥21 = 𝑚𝑖𝑛(𝑎2 , 𝑏1 − 𝑥11 ) in the cell (2, 1).

If 𝑏1 = 𝑎1 then there is a tie for second element, we can make the allocation randomly.

Step-3: Start from the North-West Corner after Transportation table and repeat step-1 and step-2
until all requirements are satisfied.

Solved problems
1. Find IBFS by North-West Corner Rule

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability
𝑶𝟏 6 4 1 5 14
𝑶𝟐 8 9 2 7 16
𝑶𝟑 4 3 6 2 5
Requirement 6 10 15 4

Solution.

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability
𝑶𝟏 (6) 14 8
6 4 1 5
𝑶𝟐 16
8 9 2 7
𝑶𝟑 5
4 3 6 2
Requirement 6 10 15 4
0

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability
𝑶𝟏 (6) (8) 14 8 0
1 5
6 4
𝑶𝟐 8 9 2 7 16
𝑶𝟑 4 3 6 2 5
Requirement 6 10
15 4
0 2

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability

𝑶𝟏 (6) (8) 14 8 0
6 4 1 5

𝑶𝟐 (2) 16 14
8 2 7
9
𝑶𝟑 4 3 6 2 5
Requirement 10
6
2 15 4
0
0

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability

𝑶𝟏 (6) (8) 14 8 0
1 5
6 4
𝑶𝟐 (2) (14) 16 14 0
8 7
9 2
𝑶𝟑 (1) (4) 5 4 0
4 3
6 2
Requirement 10 15
6 4
2 1
0 0
0 0

Total cost: 6×6+8×4+2×9+14×2+6×1+4× 𝟐=128

2) Matrix Minima Method (Least Cost Method):

Procedure:

Step-1: Determine the smallest cost of Transportation Table .Let it be 𝐶𝑖𝑗 .

Allocate 𝑥𝑖𝑗 = 𝑚𝑖𝑛(𝑎𝑖 , 𝑏𝑗 ) in the cell (𝑖 , 𝑗).

Step-2: If 𝑥𝑖𝑗 = 𝑏𝑗 then cross out j th column of the Transportation table and decrease 𝑎𝑖 𝑎𝑛𝑑 𝑏𝑗
then go to step-3.

Step-3: If 𝑥𝑖𝑗 = 𝑎𝑖 then cross out i th row of the Transportation table and decrease 𝑏𝑗 𝑎𝑛𝑑 𝑎𝑖
then go to step-4.

Step-4: Repeat step 2 and Step 3 for resulting reduced Transportation table until all the
requirements are satisfied. Whenever the minimum cost is not unique make an arbitrary choice
(select randomly).

Solved problems
1. Find IBFS by Matrix Minima Method

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability
𝑶𝟏 6 4 1 5 14
𝑶𝟐 8 9 2 7 16
𝑶𝟑 4 3 6 2 5
Requirement 6 10 15 4

Solution.

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability
𝑶𝟏 (14) 14 0
4 5
6 1
𝑶𝟐 8 9 2 7 16
𝑶𝟑 4 3 6 2 5
Requirement 6 10 15 4
1

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability
𝑶𝟏 (14) 14 0
4 5
6 1
𝑶𝟐 (1) 16 15
8 9 7
2
𝑶𝟑 4 3 6 2 5
Requirement 15
6
10 1 4
0

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability
𝑶𝟏 (14) 14 0
4 5
6 1
𝑶𝟐 (1) 16 15
8 9
2 7
𝑶𝟑 (4) 5 1
4 3 6
2
Requirement 15
6 4
10 1
0
0

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability
𝑶𝟏 (14) 14 0
4 5
6 1
𝑶𝟐 (6) (1) 16 15 9
9
8 2 7
𝑶𝟑 (1) (4) 5 1 0
4 6
3 2
Requirement 6 15
10 4
0 1
9 0
0

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability

𝑶𝟏 (14) 14 0
6 4 1 5
𝑶𝟐 (6) (9) (1) 16 15 9 0
8 9 2 7
𝑶𝟑 (1) (4) 5 1 0
4 6
3 2
Requirement 6 10 15
4
0 9 1
0
0 0

Total cost: 14× 𝟏+6× 𝟖+9×9+1×2+1× 𝟑+4× 𝟐=156

3) Vogel’s Approximation Method (VAM):

Step-1: For each Transportation table identify smallest and next to smallest cost. Determine the
difference between them for each row and column. These are called penalties. Put them along the
side of the Transportation table by enclosing them in the parenthesis against the rows and
columns.

Step-2: Identify rows or column with largest penalty among all rows and column. If a tie occurs
use any arbitrary choice. Let the Largest penalty corresponding to i th row then 𝐶𝑖𝑗 is the
smallest cost in the i th row. Allocate the amount 𝑥𝑖𝑗 = 𝑚𝑖𝑛(𝑎𝑖 , 𝑏𝑗 ) in the cell (𝑖 , 𝑗) and cross
out the ith row and jth column in usual manner.
Step-3:Again compute the column and row penalties for the reduced Transportation table and
then go to step-2.Repeate the procedure until all the requirements are satisfied.VAM is also
known as penalty method or method of penalty.

1. Find IBFS by Vogel’s Approximation Method

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability
𝑶𝟏 6 4 1 5 14
𝑶𝟐 8 9 2 7 16
𝑶𝟑 4 3 6 2 5
Requirement 6 10 15 4

Solution.

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability
𝑶𝟏 6 4 1 5 14 (3)
𝑶𝟐 (15) 16 (5) ←
8 9 7
2 1
𝑶𝟑 4 3 6 2 5 (1)
Requirement 6 10 15 0 4
(2) (1) (1) (3)

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability
𝑶𝟏 6 4 1 5 14 (3)(1)
𝑶𝟐 (15) 16 (5) ←(1)
8 9 2 7 1
𝑶𝟑 (4) 5 (1) (1)
4 3 6
2 1
Requirement 6 10 15 0 4 0
(2) (1) (1) (3)
(2) (1) (3)

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability
𝑶𝟏 6 4 1 5 14 (3)(1)(2)
𝑶𝟐 (15) 16 (5) ←(1)(1)
8 9 2 7 1
𝑶𝟑 (1) (4) 5 (1) (1)(1)
3 6
4 2 1 0
Requirement 6 5 10 15 0 4 0
(2) (1) (1) (3)
(2) (1) (3)
(2) (1)

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability
𝑶𝟏 (10) 1 5 14 (3)(1)(2)(2)
6
4 4
𝑶𝟐 (15) 16
8 9 2 7 (5) ←(1)(1)(1)
1
𝑶𝟑 (1) (4) 5 (1) (1)(1)
3 6
4 2 1 0
Requirement 6 5 10 0 15 0 4 0
(2) (1) (1) (3)
(2) (1) (3)
(2) (1)
↑ (5)
(2) ↑

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability
𝑶𝟏 (4) (10) 1 5 14 (3)(1)(2)(2)
6 4 4 0
𝑶𝟐 (15) 16
(1)
9 2 7 (5) ←(1)(1)(1)
8
1 0
𝑶𝟑 (1) (4) 5 (1) (1)(1)
3 6
4 2 1 0
Requirement 6 5 0 10 0 15 0 4 0
(2) (1) (1) (3)
(2) (1) (3)
(2) (1)
↑ (5)
2) ↑
Total cost: 24+40+8+30+4+8=114

u- v Method (or) MODI Method of Transportation problem (Modified Distribution


Method)

Step-1: Formulate the given problem as Transportation problem.

Step-2: Test whether given problem is balanced problem or not.

Step-3: If it is unbalanced make it balanced by adding dummy rows or dummy columns by


taking corresponding values zero.

Step-4: Find IBFS by Vogel’s approximation method or by any of the given methods.

Step-5: For m regions and n destinations transportation problem if (m+n-1) = number of


allocations in independent positions. These types of Problems are called non-degenerate
Transportation Problem.

Step-6: For a non-degenerate transportation problem for all basic cells (the cells which have
allocations) Solve the system of Equations 𝑢𝑖 + 𝑣𝑗 = 𝐶𝑖𝑗 for which the cell ( i , j) is in the basis
starting initially with 𝑢𝑖 = 0 or 𝑣𝑗 = 0.(Generally we will take for the row or column which
contains max number of allocations. enter the values of 𝑢𝑖 & 𝑣𝑗 on the transportation table.

Step-7: To test whether the IBFS obtained in step-4 is optimum or not. Compute the difference
𝑑𝑖𝑗 = 𝐶𝑖𝑗 − (𝑢𝑖 + 𝑣𝑗 ) for all non basic cells.(not allocated cell)

Step-8: If all 𝑑𝑖𝑗 > 0 , then current IBFS is optimum solution.

Step-9: If at least one 𝑑𝑖𝑗 < 0 , select variable 𝑥𝑟𝑠 (having the most negative 𝑑𝑟𝑠 ) to enter the
basis.

Step-10: Let the variable 𝑥𝑟𝑠 enter the basis, allocate an unknown quantity 𝜃 to the cell ( r , s),
then construct a loop that starts and ends at the cell (r, s) and connects some of the basic cells.
The amount 𝜃 is added and subtracted from the transition cells of the loop in such a manner that
availabilities and requirements are satisfied.

Step-11: Assign a largest possible value 𝜃 in such a way that the value of at least one basic
variable becomes zero and other basic variables remain non-negative.

The basic cell whose allocation has been made zero will leave the basis.

Step-12: Now return to step-6 until optimum solution has been obtained.
Solved problems:
1. Find optimum solution for following transportation problem

𝑫𝟏 𝑫𝟐 𝑫𝟑 𝑫𝟒 Availability
𝑺𝟏 19 30 50 10 7
𝑺𝟐 70 30 40 60 9
𝑺𝟑 40 8 70 20 18
Requirement 5 8 7 14

Solution. First we find Initial Basic Feasible Solution by Vogel’s Approximation Method

(5) (2)
30 50
19 10
(7) (2)
70 30
40 60
(8) (10)
40 70
8 20

Total cost= 5×19+7×40+2×60+8×8+10×20=779

Now we solve the system of equations 𝑐𝑖𝑗 = 𝑢𝑖 + 𝑣𝑗 and then find 𝑑𝑖𝑗 = 𝑐𝑖𝑗 − (𝑢𝑖 + 𝑣𝑗 ) for non
basic cells.

19 -2 -10 10 𝑢1 = 10
69 48 40 60 𝑢2 = 60
29 8 0 20 𝑢3 = 20
𝑣1 = 9 𝑣2 = −12 𝑣3 = −20 𝑣4 = 0

Now we find 𝑑𝑖𝑗 = 𝑐𝑖𝑗 − (𝑢𝑖 + 𝑣𝑗 ) values

. 30 50 . . -2 -10 .
70 30 . . - 69 48 . .
40 . 70 . 29 . 0 . -

. 32 60 .
= 1 -18 . .
11 . 70 .
Here one element of 𝑑𝑖𝑗 is less than zero. Form a loop starting at the cell (2,2) and end at the
same cell by connecting some of the basic cells.

The amount ‘θ’ to be added and subtracted from the transition cells of the loops in such a way
that the corners (availabilities) and requirements satisfied.

. 32 60 .

+θ 2- θ
-18 .
1 .
.
. 10+ θ
11 8- θ 70

Θ=Min{ 2- θ,8- θ}

Assign θ values and form new allocation table as

(5) (2)
30 50
19 10
(2) (7) (0)
70
30 40 60
(6) (12)
40 70
8 20
19 -2 8 10 𝑢1 = −32
51 30 40 42 𝑢2 = 0
29 8 18 20 𝑢3 = −22
𝑣1 = 51 𝑣2= 30 𝑣3 = 40 𝑣4 = 42

Now we find 𝑑𝑖𝑗 = 𝑐𝑖𝑗 − (𝑢𝑖 + 𝑣𝑗 ) values

. 30 50 . . -2 8 .
70 . . 60 - 51 . . 42
40 . 70 . 29 . 18 . -

. 32 42 .
19 . . 18
11 . 52 .
=

Here elements of 𝑑𝑖𝑗 are greater than or equals to zero. Therefore the optimal solution has been
obtained.

∴ Total cost=743

Degeneracy in Transportation Problem


If (m+n-1) is not equal to number of basic cells, then the basic solution is said to be degeneracy
solution. Degeneracy solution may occur in two ways.

Case (i): While finding the IBFS allocate a small quantity to one or more basic non basic cells,
so that number of basic cells becomes (m+n-1). This quantity is denoted by ∆ . This quantity is
introduced to least cost independent cell which does not form a loop.

Case (ii): During the testing of optimal solution to resolve degeneracy which occurs during
finding optimal solution. The quantity may be allocated to one or more basic cell which has
become non basic recently to have (m+n-1) in the new solution.

1. Find optimum solution for following transportation problem

𝑫𝟏 𝑫𝟐 𝑫𝟑 Availability
𝑶𝟏 8 7 3 60
𝑶𝟐 3 8 9 70
𝑶𝟑 11 3 5 80
Requirement 50 80 80
Solution.

We find Initial Basic Feasible Solution by Vogel’s Approximation Method

(60)
7
8 3
(20)
3 8
9
(50) (80) 5
11 3

Here m+n-1=3+3+-1=5

No. of basic cells=4


∴m+n-1>No. of basic cells

∴ This is degenerate problem.

To resolve degeneracy we can assign ∆ to cells (1,2), (2,2), (3,1), (3,3) which does not form
loop. But we assign to (3,3) because it is having least cost i.e 5.

(60)
7
8 3
(20)
3 8
9
(50) (80) (∆ )
11 3 5

Now we find 𝑑𝑖𝑗 = 𝑐𝑖𝑗 − (𝑢𝑖 + 𝑣𝑗 ) values

-3 1 3 𝑢1 = 3
8 7 . 3 7 9 𝑢2 = 9
. 8 . 𝑢3 = 5
- -1 3 5
11 . . 𝑣1 = −6 𝑣2= − 2 𝑣3 = 0

11 6 .
= . 1 .
12 . .

Here elements of 𝑑𝑖𝑗 are greater than or equals to zero. Therefore the optimal solution has been
obtained.

𝑶𝟏 → 𝑫𝟑

𝑶𝟐 → 𝑫𝟏

𝑶𝟐 → 𝑫𝟑

𝑶𝟑 → 𝑫𝟐

∴ Total cost=180+150+180+240=750

2. Find optimum solution for following transportation problem


𝑫𝟏 𝑫𝟐 𝑫𝟑 Availability
𝑶𝟏 25 17 25 300
𝑶𝟐 15 10 18 500
Requirement 300 300 500
Solution. Given

𝑫𝟏 𝑫𝟐 𝑫𝟑 Availability
𝑶𝟏 25 17 25 300
𝑶𝟐 15 10 18 500
Requirement 300 300 500

Given problem is unbalanced Transportation problem, make it balanced by introducing a


new row

𝑫𝟏 𝑫𝟐 𝑫𝟑 Availability
𝑶𝟏 25 17 25 300
𝑶𝟐 15 10 18 500
𝑶𝟑 0 0 0 300
Requirement 300 300 500

We find Initial Basic Feasible Solution by Vogel’s Approximation Method

(100) (200)
25 17 25
(300) (200)
15 10 18
0 (300)
0
0

Here m+n-1=3+3+-1=5

No. of basic cells=5

Now we find 𝑑𝑖𝑗 = 𝑐𝑖𝑗 − (𝑢𝑖 + 𝑣𝑗 ) values

22 17 25 𝑢1 = −7
25 . . 15 10 18 𝑢2 = 0
. . 18 𝑢3 = −18
- -3 -8 0
0 0 . 𝑣1 = 15 𝑣2 = 10 𝑣3 = 18
3 . .
= . . 0
3 8 .

Here elements of 𝑑𝑖𝑗 are greater than or equals to zero. Therefore the optimal solution has been
obtained.

𝑶𝟏 → 𝑫𝟐 =100× 𝟏𝟕 = 𝟏𝟕𝟎𝟎

𝑶𝟏 → 𝑫𝟑 =200× 𝟐𝟓 = 𝟓𝟎𝟎𝟎

𝑶𝟐 → 𝑫𝟏 =300× 𝟏𝟓 = 𝟒𝟓𝟎𝟎

𝑶𝟐 → 𝑫𝟐 =2000

𝑶𝟑 → 𝑫𝟑 =0

∴ Total cost=13200

Assignment Problem

It is an optimum technique whose aim is to assign number of jobs equal to number of persons
at a minimum cost or a maximum profit. Suppose there are n jobs and n persons are available for
doing this job. Assume that each person can do each job with varying degree of efficiency. If 𝐶𝑖𝑗
be the cost and if i th person is assigned to j th job, the problem is to find an assignment so that
the total cost for performing all jobs is minimum. The general assignment problem as follows.

Jobs
1 2 ---------- J ---------- 𝑛

1 𝐶11 𝐶12 ---------- 𝐶1𝑗 ---------- 𝐶1𝑛


2 𝐶21 𝐶22 ---------- 𝐶2𝑗 ---------- 𝐶2𝑛
---------- --------- --------- ---------- ---------- ---------
Persons 𝑖 𝐶𝑖1 𝐶𝑖2 ---------- 𝐶𝑖𝑗 ---------- 𝐶𝑖𝑛
--------- --------- --------- ---------- ---------- ---------
𝑛 𝐶𝑛1 𝐶𝑛2 ---------- 𝐶𝑛𝑗 ---------- 𝐶𝑛𝑛
𝑏1 𝑏2 𝑏𝑗 𝑏𝑚

This method is used to assign men to offices, machines to jobs, drivers to trucks, trucks to
delivery routes etc.

In an Assignment problem the number of persons should be equal to number of jobs.


Otherwise it is said to be an unbalanced Assignment Problem. To make it balanced add a dummy
job or a dummy person.

Hungarian Method to find Optimum Solution of an Assignment Problem

To solve an Assignment Problem using this method follow the following steps.

Step-1: Formulate the given problem as Assignment problem.

Step-2: Test whether it is balanced or not, if it unbalanced make it balanced by adding dummy
rows or dummy columns.

Step-3: Test whether the problem is minimization or maximization. If it is minimization locate


the largest element from the cost matrix and subtract all the elements from the largest element.

Step-4: Locate the smallest element in each row and subtract it from all the elements of
corresponding row.

Step-5: Locate the smallest element in each column of the matrix obtained in above step and
subtract it from all the elements of corresponding column.

Step-6: Draw the minimum number of horizontal and vertical lines to cover all zeros.

Step-7: Test whether the number of lines drawn be equal to number of rows.

Step-8: If number of lines drawn equal to number of rows, it is possible to assign, otherwise go
to next step.

Step-9: Locate the smallest uncovered elements by the lines and subtract it from all uncovered
elements and add it at the intersection of lines.

Step-10: Repeat the steps 6 to 9 until number of lines drawn is equal to number of rows.

Step-11: Locate the rows and columns which is having a single zero and assign that element and
delete other zeros in the corresponding row or column. If no row or column having a single zero
then assign any zero and delete other zeros in the corresponding row or column. Repeat this
procedure until all elements are satisfied.

1. Following table gives man hours to do tasks by four persons. Find the optimum
assignment to minimize the total man hours.

I II III IV
A 9 27 20 12
B 14 29 8 27
C 39 20 19 16
D 20 28 26 10
Solution.

I II III IV Column
Minimum
A 9 27 20 12 9
B 14 29 8 27 8
C 39 20 19 16 16
D 20 28 26 10 10

I II III IV
A 0 18 11 3
B 6 21 0 19
C 23 4 3 0
D 10 18 16 0
Row 0 4 0 0
minimum

I II III IV
A 0 14 11 3
B 6 17 0 19
C 23 0 3 0
D 10 14 16 0
4
3
2

1 Here number of lines=Number of rows


1
It is possible to assign
1
I II III IV
A 0 14 11 3
B 6 17 0 19
C 23 0 3 0
D 10 14 16 0
The assignment is A→I=9

B→III=8

C→II=20

D→IV=10

∴ Total man hours=9+8+20+10=47

2. In a textile emporium there are 4 sales men available to handle any counter and service time are
given in below table. How should the sales men assigned to the counters so as to minimize the total
service time.

I II III IV
A 41 72 39 52
B 22 29 49 65
C 27 39 60 51
D 45 50 48 52
Solution.

I II III IV Column
Minimum
A 41 72 39 52 39
B 22 29 49 65 22
C 27 39 60 51 27
D 45 50 48 52 45

I II III IV
A 2 33 0 13
B 0 7 27 43
C 0 12 33 24
D 0 5 3 7
Row 0 5 0 7
minimum
I II III IV
A 2 28 0 6
B 0 2 27 36
C 0 7 33 17
D 0 0 3 0 2

1
1 3
Here number of lines=3 1

Number of rows=4

It is not possible to assign

I II III IV
A 2 26 0 4 4
B 0 0 27 34 3
C 0 5 33 15
D 2 0 5 0 2

1
1
1 of rows=4
Here number of lines=Number

It is possible to assign

I II III IV
A 2 26 0 4
B 0 0 27 34
C 0 5 33 15
D 2 0 5 0
The assignment is A→ 𝐼𝐼𝐼=39

B→ 𝐼𝐼=29

C→I=27

D→IV=52

∴ Total service time=147

Assignment Problems in case of Maximization


Sometimes the Assignment Problem with Maximization. In Maximization problem we have to change
to minimization by subtracting all elements from largest element of matrix or by multiplying all elements
with (-1).

Solved Problem.

1. A company has a team of 4 sales men. And there are 4 districts; the company wants to
start its business. The capabilities of salesmen and nature of districts are vary. The company
estimates the profit per day in rupees for each sales man in each district as follows. Find the
optimum assignment and total profit.

I II III IV
A 41 72 39 52
B 22 29 49 65
C 27 39 60 51
D 45 50 48 52

Solution.

I II III IV
A 6 10 14 11
B 14 11 15 15
C 15 15 13 12
D 13 12 14 15
Maximum is 15, we subtract all the elements from this.

I II III IV Column
minimum
A 9 5 1 4 1
B 1 4 0 0 0
C 0 0 2 3 0
D 2 3 1 0 0

I II III IV
A 8 4 0 3
B 1 4 0 0
C 0 0 2 3
D 2 3 1 0 3

1
2 1
Here number of lines=3 1

Number of rows=4

It is not possible to assign

Locate smallest element in uncovered elements i. e ‘1’ and subtract it from all the elements
and add at the intersection of lines.

I II III IV
A 7 3 0 3 4
B 0 3 0 0 2
C 0 0 3 4
D 1 2 1 0 3

1
1
1
Here number of lines=Number of rows=4

It is possible to assign

I II III IV
A 7 3 0 3
B 0 3 0 0
C 0 0 3 4
D 1 2 1 0

The assignment is A→ 𝐼𝐼𝐼=14

B→ 𝐼=14

C→ 𝐼I=15

D→IV=15

∴ Total service time=58

Travelling Salesmen Problem

This is a special case of Assignment Problem. The problem is meant for allotment of salesmen to
different regions through different routes such that the variables involved such as distance, time, Cost
are optimized.

Suppose a Salesman wants to visit a certain number of cities allotted to him. He knows the distance
or cost or time of journey between every pair of cities usually denoted by 𝐶𝑖𝑗 . The Problem is to select a
route such that it starts from his home city and passes through each city one and only once and returns
to his home city in the shortest possible distance or cost or time.

Mathematical Formulation of Travelling Salesman Problem

If 𝐶𝑖𝑗 is the distance or cost or time from city I to city j and 𝑥𝑖𝑗 is equal to 1 if a salesman goes
directly from city I to city j and zero otherwise. In this case the problem is of minimization ∑𝑖 ∑𝑗 𝑥𝑖𝑗 𝐶𝑖𝑗
with an additional restriction that 𝑥𝑖𝑗 must be Chosen that no city is visited twice and it cannot go from
city i to i itself. The general Transportation Problem as follows.

𝐴1 𝐴2 −−− −−− 𝐴𝑛
𝐴1 ∞ 𝐶12 −−− −−− 𝐶1𝑛
𝐴2 𝐶21 ∞ −−− −−− 𝐶2𝑛
−−− −−− −−− −−− −−− −−−
−−− −−− −−− −−− −−− −−−
𝐴𝑛 𝐶𝑛1 𝐶𝑛2 −−− −−− ∞

Solved Problem.
1. A sales man want to visit 4 cities 1,2,3,4. He does not want to visit any city twice. The cost of going
from one city to another city in rupees is given below. Find the least cost route.

1 2 3 4
1 ∞ 30 80 50
2 40 ∞ 140 30
3 40 50 ∞ 20
4 70 80 130 ∞

Solution.

We find the optimum route for the given travelling sales man problem by Hungarian method

1 2 3 4 Column
minimum
1 ∞ 30 80 50 30
2 40 ∞ 140 30 30
3 40 50 ∞ 20 20
4 70 80 130 ∞ 70

1 2 3 4
1 ∞ 0 50 20
2 10 ∞ 110 0
3 20 30 ∞ 0
4 0 10 60 ∞
Row 0 0 50 0
minimum
1 2 3 4
1 ∞ 0 0 20 3
2 10 ∞ 60 0
3 20 30 ∞ 0
4 0 10 10 ∞

2
3
1
Here number of lines=3 1

Number of rows=4

It is not possible to assign

Locate smallest element in uncovered elements i. e ‘10’ and subtract it from all the
elements and add at the intersection of lines.

1 2 3 4
1 ∞ 0 0 30 2
2 10 ∞ 50 0
3 20 20 ∞ 0
4 0 0 0 ∞ 1

3
1
Here number of lines=3 1

Number of rows=4

It is not possible to assign.

Locate smallest element in uncovered elements i. e ‘10’ and subtract it from all the
elements and add at the intersection of lines.
1 2 3 4
1 ∞ 0 0 40 2
2 0 ∞ 40 0 4
3 10 10 ∞ 0 1
4 0 0 0 ∞ 1 1

Here number of lines=4

Number of rows=4

It is possible to assign.

1 2 3 4
1 ∞ 0 0 40
2 0 ∞ 40 0
3 10 10 ∞ 0

4 0 0 0 ∞

∴ The least cost route is 1-3-4-2-1

Total cost is 80+40+20+80=220.


UNIT-3
(SEQUENCING AND REPLACEMENT)

SEQUENCING

The selection of appropriate order in which jobs may be done (or) waiting customers may
be served is called Sequencing.

Suppose there are n jobs (n = 1,2,3, ------- ) each of which has to be processed at a time at each
machine ( M = A,B,C,---------- ) .The order of processing each job through M machines must
required the time on each machine is known. The problem is to find a sequence of possible
sequences for processing the jobs, so that the total allotted time for all jobs will be minimum.

Mathematically

Let 𝐴𝑖 = 𝑡𝑖𝑚𝑒 𝑓𝑜𝑟 𝑗𝑜𝑏 𝑖 𝑜𝑛 𝑚𝑎𝑐ℎ𝑖𝑛𝑒 𝐴

𝐵𝑖 = 𝑡𝑖𝑚𝑒 𝑓𝑜𝑟 𝑗𝑜𝑏 𝑖 𝑜𝑛 𝑚𝑎𝑐ℎ𝑖𝑛𝑒 𝐵

T = time from start of ith job to completion of last job

then the problem is determine for each machine a sequence of jobs 𝑖1 , 𝑖2 , 𝑖3 , − − − − − − 𝑖𝑛 is


the permutation of integers which will minimize T.

Solution of Sequencing Problems


There are three ways.

1) n jobs 2 Machines.
2) n jobs 3 Machines.
3) n jobs m Machines.

Processing n jobs through 2 machines

This problem can be described as follows.

Only two machines A and B are involved, each job is processed in the order AB and
expected processing times 𝐴1 , 𝐴2 , 𝐴3 , − − − − − − −𝐴𝑛 , 𝐵1 , 𝐵2 , 𝐵3 , − − − − − − 𝐵𝑛 are
known.

Processing Times Jobs (i = 1,2,3,----------)


𝐴𝑖 𝐴1 , 𝐴2 , 𝐴3 , − − − − − − −𝐴𝑛
𝐵𝑖 𝐵1 , 𝐵2 , 𝐵3 , − − − − − − 𝐵𝑛

The problem is to find sequence of jobs so as to minimize the total elapsed time T.

Procedure to find Sequence of jobs in case of n jobs two machines (Johnsons Algorithm)

Step-1: Select the least processing time appear in the list 𝐴1 , 𝐴2 , 𝐴3 , − − − − − − 𝐴𝑛 .

Step-2: If there is a time any one of the smallest processing time should be selected.

i)If the minimum is 𝐴𝑘 for some k, process the k th job first of all.

ii) If the minimum is 𝐵𝑟 for some r, process the rth job last of all.

Step-3: i) If the tie for minimum of 𝐴𝑘 and 𝐵𝑟 (i.e. 𝐴𝑘 = 𝐵𝑟 ) occurs then process the k th job
first of all and r th job in the last.

ii)If the tie for minimum occurs among 𝐴𝑖 select the job corresponding to the minimum of 𝐵𝑖 s
and process it first of all.

iii) If the tie for minimum occurs among 𝐵𝑖 select the job corresponding to the minimum of 𝐴𝑖 s
and process it in the last.

Step-4: Cross out the jobs already assigned and repeat the steps 1,2,3 arranging the jobs next to
first (or) next to last until all jobs have been assigned.

Note: - IDLE TIME


1) Idle time for machine A = Total time – Time when last job is finished on A.

2) Idle time for machine B = Time at which 1st job is finished on A +


∑[(𝑡𝑖𝑚𝑒 𝑜𝑓 𝑠𝑡𝑎𝑟𝑡𝑖𝑛𝑔 𝑜𝑓 𝑗 𝑡ℎ 𝑗𝑜𝑏) − (𝑡𝑖𝑚𝑒 𝑤ℎ𝑒𝑛 (𝑗 − 1)𝑡ℎ 𝑗𝑜𝑏 𝑖𝑠 𝑓𝑖𝑛𝑖𝑠ℎ𝑒𝑑)]

1. Find the sequence of jobs that minimizes total elapsed time to compute the following jobs on
two machines.

Jobs 1 2 3 4 5 6
A 3 12 5 2 9 11
B 8 10 9 6 3 1

Solution. Given

Jobs Machine A Machine B


1 3 8
2 12 10
3 5 9
4 2 6
5 9 3
6 11 1

Jobs Machine A Machine B


1 3 8
2 12 10
3 5 9
4 2 6
5 9 3

4 6

Jobs Machine A Machine B


1 3 8
2 12 10
3 5 9
5 9 3

4 1 5 6
Jobs Machine A Machine B
2 12 10
3 5 9

∴ The job sequence is

4 1 3 2 5 6

To calculate idle time and minimum elapsed time corresponding to the optimal sequence we use
the following table

Job Machine A Machine B


Time in Time out Time in Time out
4 0 2 2 8
1 2 5 8 16
3 5 10 16 25
2 10 22 25 35
5 22 31 35 38
6 31 42 42 43

∴ Total elapsed time=43

Idle time for machine A=43-42=1 hour

Idle time for machine B=2+4=6 hour

2. Calculate the total idle time for machine A and B and total elapse time

Jobs 1 2 3 4 5
A 5 1 9 3 10
B 2 6 7 8 4
Solution. Given

Jobs Machine A Machine B


1 5 2
2 1 6
3 9 7
4 3 8
5 10 4
2

Jobs Machine A Machine B


1 5 2
3 9 7
4 3 8
5 10 4

2 1

Jobs Machine A Machine B


3 9 7
4 3 8
5 10 4

2 4 1

Jobs Machine A Machine B


3 9 7
5 10 4

∴ The job sequence is

2 4 3 5 1

To calculate idle time and minimum elapsed time corresponding to the optimal sequence we use
the following table

Job Machine A Machine B


Time in Time out Time in Time out
2 0 1 1 7
4 1 4 7 15
3 4 13 15 22
5 13 23 23 27
1 23 28 28 30

∴ Total elapsed time=30 hours


Idle time for machine A=30-28=2 hours

Idle time for machine B=1+1+1=3 hours

Processing of n jobs through 3 Machines


This problem can be defined as follows.

Three Machines A, B, C are involved. Each job is processed in the order A, B, C.

Transfer of jobs is not permitted. Exact or Expected Processing times are given as

Job Machine-A Machine-B Machine-C


1 𝐴1 𝐵1 𝐶1
2 𝐴2 𝐵2 𝐶2
3 𝐴3 𝐵3 𝐶3
4 𝐴4 𝐵4 𝐶4
--- ----- ---- -----
N 𝐴𝑛 𝐵𝑛 𝐶𝑛

The following conditions are to be considered where either one or both holds.

a) Minimum time on Machine A ≥ Maximum time on Machine B.

b) Minimum time on Machine C ≥ Maximum time on Machine B.

Procedure

We convert n jobs 3 machines problem to n jobs 2 machines problem. For this we take 2
fictitious Machines denoted by G and H and we consider

𝐺𝑖 = 𝐴𝑖 + 𝐵𝑖 And 𝐻𝑖 = 𝐵𝑖 + 𝐶𝑖

And the order of new problem will become GH.

The resulting optimal sequence for original problem.

Solved problems
1. There are 5 jobs each of which must go through the machines A,B, C in order ABC and
processing times are given as

Jobs Ai Bi Ci
1 8 5 4
2 10 6 9
3 6 2 8
4 7 3 6
5 11 4 5

Determine the sequence for these 5 jobs that will minimize the total elapse time. Also find the
idle time for machines A,B,C.

Solution. Min(Ai)=6, Max ( Bi)=6, Min(Ci)=4

Here Min(Ai)≥ Max ( Bi)

Therefore we can adopt the procedure followed in ‘n’ jobs 2 machines case. We take two
fictitious machines G&H and we find processing times Gi and Hi using the formula

Gi= Ai+ Bi

Hi= Bi + Ci

Jobs Gi Hi
1 13 9
2 16 15
3 8 10
4 10 9
5 15 9

Minimum processing time is 8 on job 3 on Gi

Remaining table

Jobs Gi Hi
1 13 9
2 16 15
4 10 9
5 15 9

3 4
Jobs Gi Hi
1 13 9
2 16 15
5 15 9

3 1 4

Jobs Gi Hi
2 16 15
5 15 9

3 2 5 1 4

Machine Ai Machine Bi Machine Ci


Job
Time in Time out Time in Time out Time in Time out
3 0 6 6 8 8 16
2 6 16 18 22 22 31
5 16 27 27 31 31 36
1 27 35 35 40 40 44
4 35 42 42 45 45 51

Total elapsed time is 51 hours.

Idle time for A=51-42

Idle time for B= 6+8+5+4+1+6=31

Idle time for C=8+6+4+1=19

Exercise

1. The processing time for 6 jobs and 3 machines are given below
Job 1 2 3 4 5 6
Ai 3 12 5 2 9 11
Machine Bi 8 6 4 6 3 1
Ci 13 14 9 12 8 13
Determine the optimal sequence of jobs.

Processing of n jobs through m Machines


Let there are n jobs𝐴1 , 𝐴2 , 𝐴3 , − − − − −𝐴𝑛 . To be processed through M
machines 𝑀1 , 𝑀2 , 𝑀3 , − − − − −𝑀𝑛 .

The optimal solution to this problem can be found if either one or both of the condition
satisfied.

𝑖) 𝑀𝑖𝑛𝐴𝑖 ≥ 𝑀𝑎𝑥𝐴𝑖𝑗 , { i =1,2,3,-------n : j = 1,2,3,-------M-1}

ii) 𝑀𝑖𝑛𝐴𝑖𝑀 ≥ 𝑀𝑎𝑥𝐴𝑖𝑗 , { i =1,2,3,-------n : j = 1,2,3,-------M-1}

If the condition is satisfied, following procedure is done.

Convert M Machine problem into 2 Machine problems by introducing G and H.

𝐺𝑖 = 𝐴𝑖1 + 𝐴𝑖2 + 𝐴𝑖3 + − − − − − − + 𝐴𝑖𝑀−1 (𝑖 = 1,2,3, − − − − 𝑛)

𝐻𝑖 = 𝐴𝑖2 + 𝐴𝑖3 + − − − − − − + 𝐴𝑖𝑀 (𝑖 = 1,2,3, − − − − 𝑛)

After obtaining the new processing times 𝐺𝑖 𝑎𝑛𝑑 𝐻𝑖 we use the procedure followed in case of
n jobs 2 machines.

Solved problems
1. Solve the following sequencing problem of 4 jobs on 5 machines

Machines
M1 M2 M3 M4 M5
7 5 2 3 9
Jobs 6 6 4 5 10
5 4 5 6 8
8 3 3 2 6

Min Ai1=5, Min Aim=6, Max Aij=6

Condition is satisfied.

Gi=Ai1+ Ai2+ Ai3+ Ai4


Hi=Ai2+ Ai3+ Ai4+ Ai5

Jobs Gi Hi
A 17 19
B 21 25
C 20 23
D 16 14

A C B D

Job M1 M2 M3 M4 M5
Time Time Time Time Time Time Time Time Time Time
in out in out in out in out in out
A 0 7 7 12 12 14 14 17 17 26
C 7 12 12 16 16 21 21 27 27 35
B 12 18 18 24 24 28 28 33 35 45
D 18 26 26 29 29 32 33 35 45 51

Total elapsed time=51 hours

Ideal time for M1=51-26=25 hours

Ideal time for M2=(51-29)+2+2=26+7=33

Ideal time for M3=(51-32)+2+3+1+12=37

Ideal time for M4=(51-35)+14+4+1=35

Ideal time for M5=17+1=18

Processing of 2 jobs through M Machines


If there are m jobs 𝐴1 , 𝐴2 , 𝐴3 , − − − − −𝐴𝑚 and one has to perform 2 jobs (1) & (2) and
the order of jobs with their processing time known and order may not be same for both jobs and
exact on expected processing times are given as
(1) (1) (1) (1)
𝐴1 , 𝐴2 , 𝐴3 , − − − − − − − − 𝐴𝑚
(2) (2) (2) (2)
𝐴1 , 𝐴2 , 𝐴3 , − − − − − − − − 𝐴𝑚

Here the problem is minimizing the total Elapsed time T. We use Graphical Method to
solve 2 jobs M Machine Problem.
Solved problem.
1. Using graphical method determine the optimal sequence needed to process jobs 1 &2 on
5 machines A,B,C,D,E; for each machine find the job which should be done first. Also
calculate the total time needed to complete both the jobs.

Sequence A B C D E
Job 1
Time(hours) 1 2 3 5 1
Sequence A B C D E
Job 1
Time(hours) 3 4 2 1 5
Solution.

1) Draw two axis at right angles each other represents processing time on job 1 along horizontal
axis, processing time on job 2 along vertical axis, scale used must be same for both the
jobs.2)Layout the machine times for 2 jobs on the corresponding axis in the given technological
order.3) Machine A requires 1 hour for job 1 and 4 hours for job 2. A rectangle LMNP thus
constructed for machine A. Similarly for the machine BCDE.

4) Make a program by starting from origin ‘O’ and moving through the various stages of
completion (points) till the point finish is reached. Choose path consisting only of horizontal
vertical and 450 lines, a horizontal line represents work on job 1 while job 2 remains idle, A
vertical line represents work on job 2 while job 1 remains idle and a 450 line to the base
represents simultaneous work on both jobs.

5) Find the optimal path and an optimal path is one that minimizes the idle time for job 1(vertical
movement) likewise an optimum path is one that minimizes the idle time for job 2 (horizontal
line) , obviously the optimal path is one which coincides with 450 line to the maximum extent.
Both jobs cannot processed simultaneously on one machine graphically this means that diagonal
movement through shaded areas is not allowed, a good path accordingly is chooses and drawn on
graph.
6) Find the elapsed time it is obtained by adding the idle time for either job to the processing
time for job, idle time for the chosen path is formed to be 3 hours for job 1.
Similarly idle time for job 2 is 0.
Total elapsed time=12+3=15 hours
Idle time for job 1 =3 hrs

Idle time for job 2=0 hrs.


Machine A: Job 1 precedes Job 2

Machine B: Job 1 precedes Job 2

Machine C: Job 2 precedes Job 1

Machine D: Job 2 precedes Job 1

Machine E: Job 2 precedes Job 1

REPLACEMENT
Introduction

The study of Replacement is concerned with situations that arise when some items such as Machines,
Men, Electric light bulbs etc need replacement due to their deteriorating efficiency, failure or break down.

The deteriorating efficiency or complete breakdown may be either gradual or all of sudden.

Eg : A Machine becomes more and more expensive due to maintenance after a number of years, An
electric bulbs fails all of a sudden, pipeline may be blocked or an employee may loses his job and so on.

In all such situations there is a need to formulate a most economic replacement policy for replacing the
deteriorating units.

Following are the situations when the replacement of certain item needed to be done.

1) An old item has failed and does not work at all (or) old item is expected to fail shortly.

2) The old item has deteriorated and work badly or requires expensive maintenance.

3) A better Design of Equipment has been developed.

4) Replacement Problems can be broadly classified as the following two categories.

a) When Equipment or asset deteriorates with time and value of money

i) Doesn’t change with Time

ii) Change with Time

b) When items or units fail completely all of sudden

Replacement of Equipment or asset that deteriorates gradually (Deteriorate with Time)

Generally the cost maintenance and repair of certain items (equipments) increases with time
and stage may come when this cost becomes so high that it is more economical to replace the
item by a new one.

At this point a optimal Replacement policy is required. It has following cases.

Case (i): Value of Money doesn’t change with Time

To determine the optimum Replacement age of an item whose maintenance cost increases
with time and value of money remains static during that period.

Let C: Capital (value) Cost of equipment

S: Scrap value of equipment

n: number of years that equipment would be use


f(t): maintenance cost function

A(n): Average Annual Total Cost

1) When t is a continuous variable

If the Equipment is used for n years, then total cost included during this period.

𝑇𝐶 = 𝐶𝑎𝑝𝑖𝑡𝑎𝑙 𝐶𝑜𝑠𝑡 − 𝑆𝑐𝑟𝑎𝑝 𝑉𝑎𝑙𝑢𝑒 + 𝑀𝑎𝑖𝑛𝑡𝑎𝑛𝑎𝑛𝑐𝑒 𝐶𝑜𝑠𝑡


𝑛
= 𝐶 − 𝑆 + ∫0 𝑓(𝑡)𝑑𝑡

𝑇𝐶
Average Annual Total Cost 𝐴(𝑛) = 𝑛

𝐶−𝑆 1 𝑛
= 𝑛
+ 𝑛 ∫0 𝑓(𝑡)𝑑𝑡

The equipment should be replaced when maintenance cost is equal to the average annual
total cost.

2) When t is a discrete variable

Here the period of time is considered as fixed and n, t takes the values 1, 2, 3, -------- then
𝐶−𝑆 1
𝐴(𝑛) == + 𝑛 ∑𝑛𝑖=1 𝑓(𝑡)
𝑛

Note: A (n) will be minimum for n which

𝐴(𝑛 + 1) ≥ 𝐴(𝑛) And

𝐴(𝑛 − 1) ≥ 𝐴(𝑛)

In this case replace the equipment at the end of n years, if the maintenance cost (n+1) th year
is more than the total average cost in the nth year and n th year maintenance cost < the previous
year’s average total cost. .

Problems

1. A firm is considering replacement of a machine whose cost is rupees 12, 200 and the
scrap value is rupees 200, the running cost (maintenance cost) in rupees are found from
experience to be as follows

Year 1 2 3 4 5 6 7 8
Running 200 500 800 1200 1800 2500 3200 4000
cost
When the machine should be replaced.
Sol. Given

Cost of machine ( C ) =12,200

Scrap value (S) 200

Depreciation cost=12,200-200=12000

Running cost=f(n)

In order to determine the optimal time ‘n’ when the machine should be replaced we calculate an
average total cost per year during the life of machine as shown below.

Year of Running Cumulative Depreciation Total cost Average cost


service cost(f(n)) running cost cost(C-S) (A(n))
⅀𝑓(𝑛)
(1) (2) (3) (4) (5)=(3)+(4) (6)= (5)/ (1)
1 200 200 12000 12200 12200
2 500 700 12000 12700 6350
3 800 1500 12000 13500 4500
4 1200 2700 12000 14700 3675
5 1800 4500 12000 16500 3300
6 2500 7000 12000 19000 3166.66
7 3200 10200 12000 22500 3171.4
8 4000 14200 12000 26200 3275

From the above table we observed that the average total cost per year A(n) is minimum in the 6th
year i. e Rs. 3167.

Also average cost in 7th year is more than the cost in 7th year. Therefore the machine should
replaced after 6 years.

2) Machine A costs Rs. 9000, annual operating costs are Rs. 200 for 1st year and then
increased by Rs. 2000 every year. Determine the best age at which the machine should
replaced. If the optimum replacement policy is followed, what will be the average yearly
cost of owning and operating the machine.

Sol. Let the machine have no resale value when replaced. For this machine from the given data
the running cost was given in the following table.

Year 1 2 3 4 5 6
Running 200 2200 4200 6200 8200 10200
cost
Cost of machine ( C ) =Rs. 9000
Scrap value(S)=0

Depreciation cost=C-S=9000

Running cost=f(n)

Year of Running Cumulative Depreciation Total cost Average cost


service cost(f(n)) running cost cost(C-S) (A(n))
⅀𝑓(𝑛)
(1) (2) (3) (4) (5)=(3)+(4) (6)= (5)/ (1)
1 200 200 9000 9200 9200
2 2200 2400 9000 11400 5700
3 4200 6600 9000 15600 5200
4 6200 12800 9000 21800 5450
5 8200 21000 9000 30000 6000
6 10200 31200 9000 40200 6700

Case (ii): Value of Money change with Time

When value of money is taken into consideration

1) The equipment has no salvage value.

2) The maintenance cost is included in the beginning of different time periods. Since it is
assumed that the maintenance cost increased with time and each cost is to be paid in the start of
the period, let the money carry a rate of interest or r per year. Thus a rupee invested now will be
worth (1+r) after a year, thus (1 + 𝑟)2 after two years, and so on.

In other words (1 + 𝑟)−𝑛 is the present worth of one rupee spent after n years. The quantity
(1 + 𝑟)−𝑛 is called present worth factor (p.w.f).The expression (1 + 𝑟)𝑛 is known as payment
compound amount factor (c.a.f) of one rupee spent in n years. The procedure for determining the
weighted average cost (annualized cost) summarized in the following steps.

Step-1: Find the value of the maintenance cost for each of the year’s

i.e. ∑ 𝑅𝑛 𝑣 𝑛−1 (𝑛 = 1,2,3, − − − − − − −−)


1
Where 𝑣 = (1 + 𝑟)−1 = (1+𝑟)

𝑅𝑛 = 𝑅𝑢𝑛𝑛𝑖𝑛𝑔 𝑐𝑜𝑠𝑡
Step-2: Calculate the cost plus value of the cumulative present values obtained in step-1

i.e. 𝐶 + ∑ 𝑅𝑛 𝑣 𝑛−1

Step-3: Find the cumulative present value factor for all the years’ i.e. ∑ 𝑣 𝑛−1

Step-4: Determined the average annualized W (n) by dividing the entries obtained in step-2 by
corresponding entries obtained in step-3.
𝐶+∑ 𝑅𝑛 𝑣 𝑛−1
i.e. ∑ 𝑣 𝑛−1

1) The initial cost of an item is Rs. 15,000 and maintenance cost for different years is given
below

Year 1 2 3 4 5 6 7
Running 2500 3000 4000 5000 6500 8000 10000
cost(Rs.)
What is the replacement policy to be adopted if the rate of interest is 10% and there is no
solvage value.

Sol. Cost of item=15,000

Rate of interest ( r) =10%

Present value factor ( V)=1/(1+r) =m0.9091

Year Running Discount Discount C+∑ 𝑅𝑛 𝑉 𝑛−1 ∑ 𝑉 𝑛−1 Average


(n) cost(Rn) Factor(Vn-1) cost(Rn. Vn-1) annual total
cost
(1) (2) (3) (4)= (2)x(3) (5)=C+∑(4) (6)= ∑(3) (7)= (5)/ (6)
1 2500 1 2500 17500 1 17500
2 3000 0.9091 2727.3 20227 1.9091 10595
3 4000 0.8264 3306 23533 2.7355 8603
4 5000 0.7513 3756 27289 3.4868 7826
5 6500 0.6830 4440 31729 4.1698 7609
6 8000 0.6209 4967 36690 4.7909 7660
7 10000 0.5645 5645 42341 5.3552 7906

From above table we observe that the average annual total cost in 5th year is minimum i. e, Rs.
7609. The item should replace at end of 5th year.

Group Replacement
This policy is concerned with items that either work perfectly or partially or inefficiently or
fails completely. This situation is generally happens when the system consists of large number of
identical low cost items that are increasingly liable to failure with age. In such cases the
replacement of individual items would incurred a set of cost which is independent of number
replaced and it may be advantageous to replace all the items at a time at a fixed interval. This
policy is known as Group Replacement Policy and is very attractive particularly when

i)The value of any individual item is so small.

ii)The cost of keeping records of individual items is high.

iii)The purchase of such identical items in bulk is at discounted rate.

iv) Average individual replacement would be costlier than the average group replacement.

v) If sufficient number of stand by machines are available.

vi) New design of equipment considerably increases the production rate.

In all the above cases two types of replacement are considered.

1) Individual Replacement: Under this policy an item is replaced immediately after its failure.

2) Group Replacement: under this policy a decision will be taken so as to replace all the items
irrespective of the fact that the items have failed or not failed before the optimal time.

Algorithm for deciding Group Replacement Policy

Step-1: Find the probability of failure of each items at the end of each period.

Step-2: Find the number of replacements made at the end of each period with reference to
probability of failures at the end of each period considering its previous Replacement.

Step-3: Calculate the cost of individual replacement at the end of each period.

Step-4: Calculate the cost of group replacement at the end of each period.

Step-5: Calculate the total cost of group replacement including individual replacement by adding
step3 and step4.

Step-6: Calculate the average cost per period by dividing the result in step-5 with period number.

Step-7: Identify cost among the average cost per period and it will be the period of Group
Replacement Policy.

1) A factory has a large number of bulbs, all of which must be in working condition. The
mortality of bulbs is given below.
Week 1 2 3 4 5 6
Proportion of bulbs 0.1 0.15 0.25 0.35 0.12 0.03
failing during the
week

If a bulb fails in service it costs rupees 3.50 to replace but if all bulbs are replaced at a time
it costs rupees 1.20 each. Find the optimum group replacement policy by assuming
thousand bulbs are available in beginning.

Sol. Let Ni be the number of replacements made at end of ith week.

Given that initially 1000 bulbs are there.

∴ N0= 1000

N1=1000 X 0.1=100 bulbs

N2=100 X0.1+1000 X0.15

=160 bulbs

N3=281

N4=427

N5=280

N6=260

End of Total no. of Cumulative Cost of Cost of Total cost Average


week bulbs failed number of individual group cost per
bulbs failed replacement replacement week
1 100 100 350 1200 1550 1550
2 160 260 910 1200 2110 1055
3 281 541 1893.5 1200 3093.5 1031
4 427 968 3388 1200 4588 1147
5 280 1248 4368 1200 5568 1114
6 260 1508 5278 1200 6478 1080

From the above table we observe that the average cost in 3rd week is minimum i. e. Rs. 1031.
The bulbs should be replaced after every 3 weeks.
UNIT-4

GAME THEORY

Introduction

Game Theory is one of the mathematical theory’s having general characteristics of


competitive situations. It is also called as theory of Games or competitive strategies.
The Game theory helps the individual or organizations having conflictive objectives to make
effective decisions. This theory is suitable for the situations in which two players are trying to
win the Game.

Ex :- Chess, Candidates fighting for selections, Organizations introducing advertisements


campaigns by companies.

Marketing, Competing, Product to enemies planning for war tactics etc. In these cases the
decision taken by one decision makers influences the decision taken by other decision makes and
final results are mainly influenced by decisions of all the parties.

This theory mainly deals with competitive problems and is based on the minimax principle.
This principle implies that each competitor will act in order to minimize this maximum loss or
maximize its minimum gain.

GAME

A Game refers to a situation in which two or more players are competitive and are involved in
activities. Where the players must follow certain set of rules and at end of which each person
receives some benefit or satisfaction while other will have loss.

PLAYER

An individual or a group or an organization participating in the game referred to as a player.

PLAY

A Play is a set of rules and regulations which are followed by the players in a Game.

STRATEGY

A Strategy is a set of rules (programs) which specifies which of the available course of action a
player should make at each play. It is rule for decision making in advance for all the place by
which the players decide which activity to adopt.

TYPES OF STRATEGIES

1. Optimal Strategy;- An Optimal Strategy is one which provides the best situation in the Game
which involves Maximum payoff to the player. Any deviation from this strategy results in a
decreased payoff for the player.
2.Pure Strategy;- A Pure Strategy is a decision role always used by a player to choose a
particular course of action which means each player knows exactly what the other is going to do.

3. Mixed Strategy;- Mixed Strategy is a selection among pure strategies with fixed probabilities.

Payoff

A Quantitative measure of satisfaction which a player gets at the end of each play is called
Payoff.

Payoff Matrix

The Payoff Matrix is a decision Analysis tool which summarizes the Games and loss of a
decision in a tabular form. So the general payoff matrix is as follows.

𝐵1 𝐵2 ------ 𝐵𝑛
𝐴1 𝑎11 𝑎12 ------ 𝑎1𝑛
𝐴2 𝑎21 𝑎22 ------ 𝑎2𝑛
------ ------ ------ ------ ------
𝐴𝑚 𝑎𝑚1 𝑎𝑚2 ------ 𝑎𝑚𝑛
Value of the Game

It is the Expected payoff of the Game when all the players of the Game adopt their optimal
strategies.

If the value of the Game is zero, i.e. neither player wins or loses. Game is said to be fair game.

If one of the players wins (other loses) or the value of the game is non zero it is said to be
unfair game.

Saddle Point

It is the value of the game when the players adopt pure strategy. It is obtained by computing
maximin value for the row player and minimax value for the column player .If this two values
coincide the value. So obtained is called saddle point.

Limitations of Game Theory

The Game Theory was received with a great enthusiasm initially. Later it has been found with
lot of limitations. The most prominent limitations are

1. Unrealistic Assumption:- The assumption that the play known their own as well as others
payoffs in unrealistic these are only guesses.
2. Complexity with more number of players:- As the number of players increase the problem
becomes so complex that become too difficult to solve.

3. Minimax-Maximin in impractical:- The Assumption of Minimax-Maximin is through


knowledge of strategies which seems impractical.

4. Non-dynamic nature:- In practical markets generally changes and are very frequent and fast.
The game theory gives least scope to work out the strategy.

5. No-base for calculating payoff:- There is no correct formula and scientific base for
calculating payoffs of the game.

Characteristics of Game Theory

1) Change of strategy:- A game involves game of strategy as well as a chance. The activities are
evaluated by skill is known as a game of strategy and activities are determined by chance is
known as game of chance.
2) Number of activities;- it may be finite or infinite.
3) Number of persons;- if the n number of persons playing in a game is said to be an n-person game.
4) Number of alternatives available to each person: -In a particular course of action a finite number
of activities involving finite number of alternatives, otherwise the game said to be infinite.
5) Information to the players about the past activities of the other payer: -In this information is
completely available or not available.

Two Person - Zero Sum Game


Games having zero sum character where the algebraic sum of games and loses of all the players is
zero are called zero sum game. Zero sum game with two players is called rectangular games. In
this case the loss of one player is exactly equal to the gain of other player.

Rules for determining Saddle Point


Step1:- Select minimum element of each row of payoff matrix and circle it.
Step2:- Select the maximum element in each column of a payoff matrix and draw a square box around
it.
Step3:- In the payoff matrix if an element is find the circle and square it is known as saddle point.

Problems
1.Solve the following Game.

1 2 3 4 5
I -2 0 0 5 3
II 3 2 1 2 2
III -4 -3 0 -2 6
IV 5 3 -4 2 -6
Sol: - Given Payoff matrix is

Row Min
1 2 3 4 5

I -2 0 0 5 3 -2

II 3 2 1 2 2 1

III -4 -3 0 -2 6 0

IV 5 3 -4 2 -6 -6
Col max
5 3 1 5 6

Saddle point = 1
Max {Row Min}=Min{Col Max}=1
∴ Value of the game = 1
Best Strategy for player A is = II
Best Strategy for player B is = 3
2.Find the solution to the following Game using maximin and minimax principle.
𝐵1 𝐵2 𝐵3 𝐵4 Row min
𝐴1 2 5 8 3 2
𝐴2 -1 2 4 6 -1
𝐴30 4 3 2 0
𝐴4 1 2 3 4 1
Col max 2 5 8 6
Saddle point = 2
Max {Row Min} = Min{Col Max}=2
∴ Value of the game = 2
Best Strategy for player A is = 𝐴1
Best Strategy for player B is = 𝐵1

Games without saddle points (Mixed Strategies)


Consider the (m × n) game without saddle point.
Let 𝑃𝑖 be the probability associated with this strategies of player A and 𝑞𝑗 be the probabilities
associated with strategies of player B , such that
∑𝑚
𝑖 𝑝𝑖 = 1 , ∑𝑛𝑗 𝑞𝑗 = 1 , 𝑖 = 1 𝑡𝑜 𝑚, 𝑗 = 1 𝑡𝑜 𝑛
The Expected pay off in this case is given by
𝐸(𝑝, 𝑞) = ∑𝑖 ∑𝑗 𝑝𝑖 𝑎𝑖𝑗 𝑞𝑗

Where 𝑎𝑖𝑗 is the pay off corresponding to the i th strategies of player A, against j th strategy of
player B

For payer A Maximin is denoted by 𝑣 is given by

𝑣 = 𝑀𝑎𝑥𝑝 𝑀𝑖𝑛𝑞 𝐸(𝑝, 𝑞)

=𝑀𝑎𝑥𝑝 {𝑀𝑖𝑛𝑞 [∑𝑚


𝑖=1 𝑝𝑖 𝑎𝑖𝑗 ]}

Similarly for player B minimax is denoted by 𝑣̅ is given by

𝑣̅ = 𝑀𝑖𝑛𝑞 𝑀𝑎𝑥𝑝 𝐸(𝑝, 𝑞)

= 𝑀𝑖𝑛𝑞 {𝑀𝑎𝑥𝑝 [∑𝑛𝑗=1 𝑞𝑗 𝑎𝑖𝑗 ]}

If 𝑣 = 𝑣̅ = v then it is called fair game and v is the value of the game. The corresponding
strategy for player A and B are called the optimum strategies respectively. The fundamental
theorem of game theory says that

𝑣̅ ≥ 𝑣

Game without saddle points mixed strategies; we can solve using following methods.

1. Arithmetic Method

2. Graphical method

3. Linear Programming Method

1. Arithmetic Method

We apply this method for (2× 2) rectangular games. Let a pay off matrix be

𝐵1 𝐵2
𝐴1 𝑎11 𝑎12 𝑝1
𝐴2 𝑎21 𝑎22 𝑝2
𝑞1 𝑞2

Such that 𝑝1 + 𝑝2 = 1
𝑞1 + 𝑞2 = 1

Where 𝑝1 , 𝑝2 are probabilities associated with the strategies of player A and 𝑞1 , 𝑞2 are
probabilities associated with player B.
𝑎22 −𝑎21
We have 𝑝1 = (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )
and 𝑝2 = 1 − 𝑝1

𝑎22 −𝑎12
𝑞1 = (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )
And 𝑞2 = 1 − 𝑞1

𝑎11 𝑎22 −𝑎12 𝑎21


Value of the game 𝑣 = (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )

Problems

1. for the game with the following payoff matrix. Determine the optimum strategies and value of
the game.

𝐵1 𝐵2
𝐴1 5 1
𝐴2 3 4
Sol:-

𝐵1 𝐵2 Row
min
𝐴1 5 1 1
𝐴2 3 4 3
Col 5 4
max
Max {Row Min}≠Min{Col Max}

There is no saddle point.

Clearly the given problem has no saddle point.


By Arithmetic method
𝑎 −𝑎 4−3 1
𝑝1 = (𝑎 +𝑎 22)−(𝑎21 +𝑎 ) = (5+4)−(1+3) = 5
11 22 12 21

1 4
𝑝2 = 1 − 𝑝1 = 1 − 5 = 5
𝑎22 −𝑎12 4−1 3
𝑞1 = (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )
= (5+4)−(1+3)
= 5

3 2
𝑞2 = 1 − 𝑞1 = 1 − 5 = 5

𝑎11 𝑎22 −𝑎12 𝑎21 (5.4)−(1.3) 17


∴ Value of the Game 𝑣 = (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )
= (5+4)−(1+3) = 5
1 4
Optimum strategy for player A is 𝑝1 = 5 , 𝑝2 = 5
3 2
Optimum strategy for player B is 𝑞1 = , 𝑞2 =
5 5

2. There are two players in a game named as matching player and non matching player. The
matching player is paid Rs 8.00 if the two coins turn both heads and Rs 1.00 if the coins turn
both tails. The non matching player is paid Rs 3.00 when the two coins do not match. Given the
choice of being the matching or non matching player, which you choose and what would be your
strategy?

The payoff matrix for the matching player is given by

𝐻 𝑇 Row
min
𝐻 8 −3 −3
𝑇 −3 1 −3
Col 8 1
max
Clearly the given problem has no saddle point.
By Arithmetic method
𝑎 −𝑎 1−(−3) 4
𝑝1 = (𝑎 +𝑎 22)−(𝑎21 +𝑎 ) = (8+1)—(−3−3) =
11 22 12 21 15

4 11
𝑝2 = 1 − 𝑝1 = 1 − 15 = 15
𝑎22 −𝑎12 1−(−3) 4
𝑞1 = (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )
= (8+1)—(−3−3)
= 15

4 11
𝑞2 = 1 − 𝑞1 = 1 − 15 = 15

𝑎11 𝑎22 −𝑎12 𝑎21 (1.8)−(−3)(−3) −1


∴ Value of the Game 𝑣 = (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )
= (1+8)−(−3−3)
= 15
4 11
Optimum strategy for matching player is 𝑝1 = 15 , 𝑝2 = 15
4 11
Optimum strategy for non matching player is 𝑞1 = 15 , 𝑞2 = 15

2. Graphical method

We apply Graphical method for 2 × 𝑛 or 𝑚 × 2 games.consider the following 2 × 𝑛


payoff matrix without saddle points.

𝐵1 𝐵2 ----------- 𝐵𝑛
𝐴1 𝑎11 𝑎12 ----------- 𝑎1𝑛
𝐴2 𝑎21 𝑎22 ----------- 𝑎2𝑛
Player A has two strategies 𝐴1 , 𝐴2 with probabilities 𝑃1 , 𝑃2 . Such that 𝑃1 + 𝑃2 = 1

Expected payoff for A corresponding to the huge strategies of B are given by

B s Strategy A s Expected pay off


𝐵1 𝑎11 𝑃1 + 𝑎12 𝑃2
𝐵2 𝑎12 𝑃1 + 𝑎22 𝑃2
---------- --------------
𝐵𝑛 𝑎1𝑛 𝑃1 + 𝑎2𝑛 𝑃2
According to maximin criteria for mixed strategy for player A should select the value of 𝑃1
which maximizes its minimum expected payoff. This may be done by plotting the lines
representing player A s expected payoff values. The highest point of lower boundary of these
lines will give maximum expected payoff among minimum expected payoffs on the lower
boundary .The two optimal strategies for B are given by two lines which passes through
maximum point can be determined.

The 𝑚 × 2 games are also treated in the same way for these 𝑚 × 2 games we will get
minimax point which will be the lowest point on the upper boundary.

Problems

1. Find value of the game and optimal strategy for following games graphically.

I II III IV
I 1 0 4 -1
II 0 1 -2 5

Sol:-

I II III IV
I 1 0 4 -1 -1
II 0 1 -2 5 -2
1 1 4 5

Clearly the given game without saddle point. We can apply Graphical method.
Player A s expected payoff against B s strategies are given by

B s strategies A s expected payoff


BI 1. 𝑃1 + 0. 𝑃2
BII 0. 𝑃1 + 1. 𝑃2
BIII 4. 𝑃1 − 2. 𝑃2
BIV −1. 𝑃1 + 5. 𝑃2
These expected payoff equations are plotted as functions of 𝑃1 (𝑃2 = 1 − 𝑃1 ) in the following
figure. which shows the payoffs of each column represented as points on two vertical axes I and
II with unit distance.

Player A wishes to maximize his minimum expected payoff .The highest point of intersection
H on the lower boundary of A. This point H represents the maximin expected value of the game.

The lines 𝐵2 𝑎𝑛𝑑 𝐵1 passing through H. Therefore the reduced matrix is

I II
I 1 0
II 0 1
By Arithmetic method
𝑎 −𝑎 1−0 1
𝑝1 = (𝑎 +𝑎 22)−(𝑎21 +𝑎 = (1+1)—(0+0)
=
11 22 12 21 ) 2
1 1
𝑝2 = 1 − 𝑝1 = 1 − 2 = 2
𝑎22 −𝑎12 1−0 1
𝑞1 = (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )
= (1+1)—(0+0)
= 2

1 1
𝑞2 = 1 − 𝑞1 = 1 − 2 = 2

𝑎11 𝑎22 −𝑎12 𝑎21 (1.1)−(0.0) 1


∴ Value of the Game 𝑣 = (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )
= (1+1)−(0+0) = 2
1 1
Optimum strategy for player A is 𝑝1 = 2 , 𝑝2 = 2
1 1
Optimum strategy for player B is 𝑞1 = 2 , 𝑞2 = 2

2. Solve the following game graphically.

𝐵1 𝐵2
𝐴1 1 8
𝐴2 3 5
𝐴3 11 2

Sol: - Given payoff matrix is

𝐵1 𝐵2 Row min
𝐴1 1 8 1
𝐴2 3 5 3
𝐴3 11 2 2
Col max 11 8

Clearly the given game without saddle point. We can apply Graphical method.

Let 𝑞1 &𝑞2 be the strategies against player A where 𝑞1 + 𝑞2 = 1


Player B s expected payoff against A s strategies are given by

B s strategies A s expected payoff


𝐴1 𝐸1 (𝑞) = 𝑞1 + 8𝑞2
= 𝑞1 + 8(1 − 𝑞1 )
= 7 − 7𝑞1
𝐴2 𝐸2 (𝑞) = 3𝑞1 + 5𝑞2
= 3𝑞1 + 5(1 − 𝑞1 )
= 5 − 2𝑞1
𝐴3 𝐸3 (𝑞) = 11𝑞1 + 2𝑞2
= 11𝑞1 + 2(1 − 𝑞1 )
= 2 + 9𝑞1
These expected payoff equations are plotted as functions of 𝑞1 (𝑞2 = 1 − 𝑞1 ) in the following
figure. which shows the payoffs of each column represented as points on two vertical axes I and
II with unit distance.
Player B wishes to minimum his maximize expected payoff .The highest point of intersection
H on the upper boundary of B. This point H represents the minimax expected value of the game.

The lines 𝐴1 𝑎𝑛𝑑 𝐴3 passing through H. Therefore the reduced matrix is

𝐵1 𝐵2
𝐴1 1 8
𝐴3 11 2
By Arithmetic method
𝑎 −𝑎 2−11 9
𝑝1 = (𝑎 +𝑎 22)−(𝑎21 +𝑎 ) = (1+2)—(8+11)
=
11 22 12 21 16

9 7
𝑝2 = 1 − 𝑝1 = 1 − 16 = 16
𝑎22 −𝑎12 2−8 3
𝑞1 = (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )
= (1+2)—(8+11)
= 8

3 5
𝑞2 = 1 − 𝑞1 = 1 − 8 = 8

𝑎11 𝑎22 −𝑎12 𝑎21 1.2−8.11 43


∴ Value of the Game 𝑣 = (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )
= (1+2)—(8+11)
= 8
9 7
Optimum strategy for player A is 𝑝1 = 16 , 𝑝2 = 16
3 5
Optimum strategy for player B is 𝑞1 = 8 , 𝑞2 = 8

Dominance Property

Sometimes we may observe that one of the pure strategies of either player is always inferior
to at least one of the remaining strategy. The superior strategy is said to be dominated by inferior
strategy .By eliminating the inferior strategy one can reduce the dimension of payoff matrix in
order to solve the game.

General Rules for Dominance

a) If all the elements of a row say k th row are ≤ the corresponding elements of any other r th
row then we say k th row is dominated by r th row.
b) If all the elements of a column say k th column are ≥ the corresponding elements of any
other r th column then we say k th column is dominated by r th column.
c) Dominated rows or columns may be deleted to reduce the size of the matrix.

Note: - Dominance property is not always based on the superiority of pure strategy. A given
strategy can also be said to be dominate if it is inferior to average of two or more other pure
strategies.

Problems
1. Use dominance property to obtain the optimal strategies and value of the Game.

I II III IV
I 3 5 4 2
II 5 6 2 4
III 2 1 4 0
IV 3 3 5 2
Sol: - Given

I II III IV Row min


I 3 5 4 2 2
II 5 6 2 4 2
III 2 1 4 0 0
IV 3 3 5 2 2
Col max 5 6 5 4
Max {Row Min}≠Min{Col Max}

Clearly the given problem has no saddle point.

By using dominance property

All the elements of 3rd row ≤ corresponding elements of 4rth row.


Then delete 3rd row.
The reduced matrix is
I II III IV
I 3 5 4 2
II 5 6 2 4
IV 3 3 5 2
All the elements of column 2 is ≥ corresponding elements of column 1.Then delete column 1.
II III IV
I 5 4 2
II 6 2 4
IV 3 5 2
Column2 ≥ column4.Then delete column 2.

III IV
I 4 2
II 2 4
IV 5 2

Row1 ≤ Row4.Then delete Row1.


III IV
II 2 4
IV 5 2

By Arithmetic method
𝑎22 −𝑎21 2−5 3
𝑝1 = (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )
= (2+2)−(4+5)
= 5

3 2
𝑝2 = 1 − 𝑝1 = 1 − 5 = 5
𝑎22 −𝑎12 2−4 2
𝑞1 = (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )
= (2+2)−(4+5)
= 5

2 3
𝑞2 = 1 − 𝑞1 = 1 − 5 = 5

𝑎11 𝑎22 −𝑎12 𝑎21 (2.2)−(4.5) 16


∴ Value of the Game 𝑣 = (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )
= (2+2)−(4+5)
= 5
2. Use dominance property to solve the following game.

I II III IV
I 3 2 4 0
II 3 4 2 4
III 4 2 4 0
IV 0 4 0 8
Sol: - Given

I II III IV Row min


I 3 2 4 0 0
II 3 4 2 4 2
III 4 2 4 0 0
IV 0 4 0 8 0
Col max 4 4 4 8
Max {Row Min}≠Min{Col Max}

Clearly the given problem has no saddle point.

By using dominance property


Row1 ≤ Row3.Then deletes row1.

I II III IV
II 3 4 2 4
III 4 2 4 0
IV 0 4 0 8
Column1 ≥ Column3.then delete column1.

II III IV
II 4 2 4
III 2 4 0
IV 4 0 8
Dominance property is not always based on superiority of pure strategies. A given strategy can
also said to be dominated if it is inferior to average of two or more strategies.
Column2 ≥ Average of column 3 and column4.Then delete column2.

III IV
II 2 4
III 4 0
IV 0 8
Row2 ≤ Average of row 3 and row 4.Then delete row2.

III IV
III 4 0
IV 0 8
By Arithmetic method
𝑎 −𝑎 8−0 2
𝑝1 = (𝑎 +𝑎 22)−(𝑎21 +𝑎 = (4+8)−(0+0)
=
11 22 12 21 ) 3

2 1
𝑝2 = 1 − 𝑝1 = 1 − 3 = 3
𝑎22 −𝑎12 8−0 2
𝑞1 = (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )
= (4+8)−(0+0)
= 3

2 1
𝑞2 = 1 − 𝑞1 = 1 − 3 = 3

𝑎11 𝑎22 −𝑎12 𝑎21 (4.8)−(0.0) 8


∴ Value of the Game 𝑣 = (𝑎11 +𝑎22 )−(𝑎12 +𝑎21 )
= (4+8)−(0+0)
= 3
Optimal strategy for player A is
I II III IV
0 0 2 1
3 3

Optimal strategy for player B is


I II III IV
0 0 2 1
3 3

INVENTORY MODELS

The word inventory means simply a stock of ideal resources of any kind having an economic
value. Inventory means a physical stock of goods which is kept in hand for smooth and efficient
running of an organization it may consists raw materials, spare parts, finished goods, human
resources, financial resources etc.

The problems faced by different organizations have necessitated the use of scientific
techniques in the management of inventories known as inventory control or inventory
management.
The Inventory Cost

Various cost associated with inventory control are classified as follows.

1. Set up cost: - The cost associated with setting of the machinery before starting the production
and is independent of the quantity ordered for production.

2. Ordering Cost: - This cost associated with ordering of raw material for production.
Advertisement, consumption of stationary and postage, telephone charges, rent for space used by
purchase department, travelling expenditure etc. constitute the ordering cost.

3. Purchase or Production cost: - This is cost of purchasing or producing a unit of item. This is
very important when discounts are allowed.

4. Carrying or holding cost: - The cost associated with the storage costs like rent, interest on
the money locked up , insurance of stored equipment ,production, taxes, depreciation of
equipment etc.

5. Shortage or stock out cost: - The penalty cost for running out of stock .This cost includes the
loss of potential profit through sales of items and loss of good will, in terms of permanent loss of
customers and its associated loss profit in future sales.

6. Salvage cost or selling price: - When demand for certain commodity is affected by quality
stocked, decision problem based on profit maximization criterion includes the revenue from
selling.

Economic order quantity: The quantity to be ordered in the most economical way is called
Economic Order Quantity (EOQ).

Characteristics of Inventory system:

(i) Demand: The nature of demand must be known for inventory control. Demand
may be deterministic or Probabilistic.
(ii) Order cycle: The period of time between two consecutive placements of orders is
called order cycle.
(iii) Lead Time: When an order is placed the supplier takes some time to supply the
item. Lead time is the time to between the placement of an order and receipt of
materials.
(iv) Replenishment of Stock: Replenishment of stock may occur instantaneously
without lead time or uniform replenishment may occur when the goods are
manufactured by the factory itself.
(v) Planning horizon(Period): The duration of time over which a particular inventory
level will be maintained is called planning horizon.

Inventory Models:-
Model 1: Demand rate uniform and production rate is infinite(or) Economic Order
quantity model with uniform demand (or) Wilson Harris Simple EOQ:

In this model demand is assumed to be constant. We have to determine the quantity to be


ordered in the most economical manner (Economic Order Quantity). We make the following
assumptions in this model:

i) Demand is known and uniform


ii) Shortages not allowed. As soon as the level of inventory zero, the inventory is
replenished.
iii) Production or supply of materials is instantaneous.
iv) Lead time is zero.
v) Order cost per order is same irrespective of order size.

Assumptions for finding EOQ for constant demand:

C0= Order cost per Order

Cc= Carrying cost (or) Holding cost per unit per year

D= Annual demand on requirement per year.

P= Price of item

Q= Order quantity per year


𝑇𝑜𝑡𝑎𝑙 𝑑𝑒𝑚𝑎𝑛𝑑 𝑖𝑛 𝑡ℎ𝑒 𝑦𝑒𝑎𝑟 𝐷
1)Annual order cost(AOC): Total no. of orders in the year= 𝑂𝑟𝑑𝑒𝑟 𝑞𝑢𝑎𝑛𝑡𝑖𝑡𝑦 𝑝𝑒𝑟 𝑂𝑟𝑑𝑒𝑟 = 𝑄

Total annual order cost=Order cost per order× No. of orders per year
𝐷
i.e AOC= C0 × 𝑄

2) Annual Carrying cost(ACC):


𝑄
Average quantity stocked= 2

ACC=Carrying cost per unit period × Average quantity stocked


𝑸
i.e ACC= Cc× 𝟐

3) Annual purchase cost(APC):

APC= Item price × Demand


= P× D

Annual Total Cost(ATC)= AOC+ ACC+ APC


C0 𝐷 Cc Q
= + + P.D
𝑄 2

Let Q0 (Economic Order Quantity) be optimum Q which is given by

2𝐶0 𝐷
Q0=√ 𝐶𝑐

Optimum annual inventory cost=√2𝐶0 𝐶𝑐 . 𝐷

Note: If carrying cost or Holding cost is expressed in terms of the value of stock held then

Cc=i.p

𝟐𝑪𝟎 𝑫
Q0=√ Here ‘i’ is interest
𝒊.𝒑

Problems:

1. A company uses 10000 units per year of an item. The purchase price is Rs. 25 per
order. The interest on carrying cost per year is 12% of the inventory value and cost of
item is Rs.1. Then find the EOQ, the no. of orders per year, if the lead time is 4 weeks
and assuming 50 working weeks per year find re order point.

Sol. Given

Annual demand(D)= 10000 units

Co (Order cost per order)= Rs. 25

Interest on carrying cost= i = 12% = 0.12

P (Price of item)= Rs. 1

𝟐𝑪𝟎 𝑫
(i) EOQ= √ 𝒊𝑷

𝟐×𝟐𝟓×𝟏𝟎𝟎𝟎𝟎
=√ 𝟎.𝟏𝟐×𝟏

=2041.24≅ 2041
𝐷 10000
(ii) The no. of orders per year = 𝑄= 2041 =4.899 ≅ 5
50
(iii) 𝑤𝑒𝑒𝑘𝑠= 10 weeks
5
10-4 = 6 weeks
𝑁𝑜.𝑜𝑓 𝑤𝑜𝑟𝑘𝑖𝑛𝑔 𝑤𝑒𝑒𝑘𝑠 𝑝𝑒𝑟 𝑦𝑒𝑎𝑟
Re order period= 𝑁𝑜.𝑜𝑓 𝑜𝑟𝑑𝑒𝑟𝑠 𝑝𝑒𝑟 𝑦𝑒𝑎𝑟

=10 weeks

Given lead time=4 weeks

∴Re oeder point is at the end of 6 weeks= 10-4

2. A company uses 24000 units of raw material which costs rupees 12.5 per unit. Placing
each order costs Rs.22.5 and carrying cost is 5.4% per year of the average inventory.
Find the economic order quantity including the cost of material.

Sol. Annual demand (D)= 24000 units.

Order cost= C0=22.5

Interest on carrying cost= i = 5.4% = 0.054

P (Price of item)= Rs. 12.5

𝟐𝑪𝟎 𝑫
(i) EOQ= √ 𝒊𝑷

1264.91≅ 𝟏𝟐𝟔𝟓

Total inventory cost=AOC+ ACC+ APC


𝐶0 𝐷 𝐶𝑐 𝑄
= + + P.D
𝑄 2

Cc =i ×P

= 0.054× 12.5 =0.675


22.5×24000 0.675×1265
Total inventory cost= + +12.5+ 24000
1265 2

300853.15

Model -2: Demand rate not uniform and Production rate infinite:
In this model assumptions are same as model-1with exception that in place of uniform
demand. We were given total demand. And since annual ordering cost equal to annual
carrying cost will have same formulae as in model-1.

Model -3:

Demand rate uniform and production rate finite. (Manufacturing model)

This model is similar to that of model-1 except that of manufacturing rate(replenishment).

This model assumes to replenish continuously till the maximum quantity is reached due to
the fact that in many situations the amount ordered will not be receivable all at once. But will
be available at finite supply rate. The following points are assumed in this model.

- Supply is continuous till Q max is reached


- The production rate say ‘r’) is greater than demand (d)

i.e r >d

-Production begins immediately after production setup

Important formulae of this model:

𝟐𝑪𝟎 𝑫
(ii) 1) EOQ= √ 𝒊𝑷

2𝐶0 𝐷 𝑟
1) Economic batch quantity(EBQ)= 𝑄𝑜𝑝𝑡 = √ (𝑟−𝑑)
𝐶𝑐

𝑟−𝑑
2) Total annual inventory cost= √2𝐶0 𝐷𝐶𝑐 ( )
𝑟

𝑄𝑜𝑝𝑡
3) Optimum length of production run= 𝑑

𝑄𝑜𝑝𝑡
4) Production cycle time= 𝑟

Problem.

5) A contractor has to supply 10,000 bearings per day to an automobile manufacturer. He

finds that when he starts production run he can produce 25000 bearings per day. The cost
of holding a bearing in stock for ayear is Rs. 2 and setup cost production run is Rs. 180.

Find the economic batch quantity and frequency of production run and production cycle

time. (Assume 300 days per year).

Sol. Given D= 10,000 per day

r=25,000 per day

Holding cost=Rs. 2 per year(Cc)

Set up cost (C0)= 180 per year.

2𝐶0 𝐷 𝑟
EBQ= 𝑄𝑜𝑝𝑡 = √ (𝑟−𝑑)
𝐶𝑐

Annual demand (D)=10,000× 300

=30,00,000 per year

r= 75, 00,000 per year

Qopt=30,000

𝑄𝑜𝑝𝑡
Frequency of production run= 𝑑

30000
=10000 = 3 𝑑𝑎𝑦𝑠

𝑄𝑜𝑝𝑡
Production cycle time= 𝑟

=3000000/25000

=1.2 days
UNIT-5
WAITING LINES
INTRODUCTION

A Flow of customers from infinite/finite population towards the service facility form a queue
(waiting line) on account of lack of capability to serve them at a time. The queue may be of
person waiting at a doctor’s clinic or at railway booking office, we wait to eat at restaurant and
this waiting phenomenon is not limited to human beings only. Jobs wait to process on machine,
cars and other vehicles wait at traffic signals. In the absence of a perfect balance between the
service facilities or for the customers, waiting is required either of the services facilities or for
the customer arrival.

By the term customer we mean the arriving unit that requires some service to be performed.
The customers may be persons, machines, vehicles etc. Queue stands for number of customers
waiting to be served. The queue does not include the customer being served. The process or
system that performs the service to the customers is termed by service channel or service facility.

The subject of Queuing is not directly concerned with optimization (max/min), rather if
attempts to explore understand and compare various queuing situations and thus indirectly
achieves optimization approximately.

Elements of Queuing System

The basic components of queuing system as follows.

1. Input Process (Arrival Process)

2. Queue Discipline

3. Service Mechanism

4. Capacity of the system

1. Input process: - This element of Queuing system is concerned with the pattern in which the
customers arrive for service. Input source can be described by following three factors.

a) Size of the Queue :- If the total number of potential customers requiring service are only few,
then size of the input source is said to finite .If the potential customers requiring service area
sufficiently large in number, then the input service is considered to be infinite. When more than
one arrival is allowed to enter the system simultaneously the input is said to be occur in bulk or
batches.
b) Arrival Distribution: - If the pattern at which customers arrive at the service system or time
between successive arrivals ( inter arrival times ) is uncertain , the arrival pattern is measured by
either mean arrival rate or inter arrival time. These characterized by the probability distribution
associated with this random process. The arrival rate follows Poisson distribution and inter
arrival time follow an exponential distribution.

c) Customers behavior: - It is also necessary to known the reaction of customer upon entering
the system. A system be decide to wait no matter how long the queue becomes or if the queue is
too long to suit him may decide not to enter it .If a customer decides not to enter the queue
because of its huge length ,he is said to have balked. A customer may enter the queue but after
some time loses patience and decides to leave. In this case he is said to have reneged. In the case
when there are two or more queues, customer may move one queue to another for his personal
economic gains that is jockey for position.

2. Queue Discipline: - This is referred to the order in which the numbers of the queue are
selected for service. It may be first come first served, last come first served, Service in random
order or priority service.

First come First server (FCFS): - The order of service of customer is in the order of their arrival
or FIFO.

Ex: - Ticket at cinema hall.

Last come first served (LCFS):- The order of service is the reverse of the order in which the
customers arrived (LIFO).

Ex:- Loading and unloading a truck or godown,wearing socks and shoes, dressing a shirt and
coat over it.

Served in random order (SIRO):- The selection of customers for service is done a random way.

Ex: - lottery system from which one is picked up.

Priority service: - The customers are chosen on some priority procedure.

Ex: - Priority given at APSRTC reservation counter who buy a CAT Card.

3. Service mechanism: - The service mechanism is concerned with service time and service
facilities. Service time is the time interval from the commencement of service to the completion
of services. If there is infinite number of servers then all the customers are served
instantaneously on arrival and there will be no queue.

If the number of servers is finite then customers are served according to a specific order.
a)Single queue one serve:- One queue one service channel, where in the customers waits till the
service point is ready to take him in for servicing.

b) Single queue several servers: - The customers wait in a single queue until one of the service
channels is ready to take them in for servicing.

c) Several queues one server: - There are several queues and the customer may join any one of
these but there is only one service channel.

d) Several servers: - When there are several service channels available to provide service much
depends upon their arrangements. They may arrange in parallel or in series or a more complex
combination of both, depending on the design of the systems service mechanism.

By parallel channels, we mean a number of channels providing identical service facilities. It


consists of multiple servers with single or multiple queues.

For series channels, it consists of multiple servers which are connected in series.

4) Capacity of the system: - The source from which customers are generated may be finite or
infinite. A finite source limits the customers arriving service. i.e., there is a finite limit to
maximum queue size. The queue can also be viewed as one with force balking where a customer
is forced to bulk if he arrives at a time when queue size is at it limit. Alternatively an infinite
source is forever abundant as in the case of telephone calls arriving at a telephone exchange.

Characteristics of Queuing system

Some of the operational characteristics of queuing system that are of a general interest for
the evaluation of the performance of an existing queuing system and to design a new system are
as follows.

1. Expected number of customers in the system denoted by E (n) or 𝐿𝑠 is the average number of
customers in the system, both waiting and in service. Here n= number of customers in the
queuing system.

2. Expected number of customers in the queue denoted by E (m) or 𝐿𝑞 is the average number of
customers waiting in the queue. Here m=n-1 i.e., excluding the customer being served.

3. Expected waiting time in the system denoted by E (v) or 𝑊𝑠 is the average number of
customers in the system .It is generally taken to be waiting time plus servicing time.

4. Expected waiting time in the queue denoted by E (w) or 𝑊𝑞 is the average time spent by
customers in the queue before the commencement of his service.
ƛ
5. The server utilization factor (or busy period) denoted by 𝜌(= 𝜇 ) is the proportion of time that
a server actually spends with the customers. Here ƛ stands for the average number of customers
arriving per unit of time and 𝜇 stands for the average number of customers completing service
per unit of time.

The server utilization factor is also known as traffic intensity.

Terminology

1. Customers: - The persons or objects that required certain service are called customers.

2. Server: - The person or an object or a machine that provides certain defined service is known
as server.

3. Service: - The activity between server and customer is called service that consumes some
times.

4. Queue or waiting line: - A systematic or a disciplined arrangement of a group of persons or


objects that wait for a service is called queue or waiting line.

5. Arrival: - The process of customer coming towards service facility or server to receive certain
service is arrival.

6. Mean rate of Arrival: - The average number of customers arriving per unit time is called
mean rate of arrival and it is denoted byƛ.

𝑇𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑎𝑟𝑟𝑖𝑣𝑖𝑛𝑔


ƛ=
𝑇𝑜𝑡𝑎𝑙 𝑡𝑖𝑚𝑒 𝑡𝑎𝑘𝑒𝑛
7. Mean inter arrival time: - It is the average time gap between two consecutive arrivals of
customers. It is the inverse of mean arrival rate.

1 𝑇𝑜𝑡𝑎𝑙 𝑎𝑟𝑟𝑖𝑣𝑎𝑙 𝑡𝑖𝑚𝑒


=
ƛ 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑎𝑟𝑟𝑖𝑣𝑎𝑙𝑠

8. Rate of service: - It is the average number of customers served per unit time and is denoted
by𝜇.

𝑇𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑠𝑒𝑟𝑣𝑒𝑑


𝜇=
𝑇𝑜𝑡𝑎𝑙 𝑠𝑒𝑟𝑣𝑖𝑐𝑒 𝑡𝑖𝑚𝑒
9. Mean service time: - It is the average time taken by the server to serve a customer and equal
to inverse of service rate.

1 𝑇𝑜𝑡𝑎𝑙 𝑠𝑒𝑟𝑣𝑖𝑐𝑒 𝑡𝑖𝑚𝑒


=
𝜇 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑠𝑒𝑟𝑣𝑒𝑑
Classification of Queuing Models (Kendall’s notations used for representing the Queuing
models)

The representation of Queuing model as given below. It is introduced by D.G. Kendall and
then by A. Lee.

(a / b / c): ( d / e / f )

Where a: arrival distribution (inter arrival)

b: departure distribution (inter service time)

c: Number of parallel service channels in the system

d: service discipline

e: Maximum number of customers allowed in the system

f:calling source or population

If we specify the following letters are

Symbols for a and b

M=Mark𝐸𝐾 ovian (Poisson) arrival or departure distribution.

𝐸𝐾 = Erlangen or gammas inter arrival for service time distribution.

𝐺𝐼 =General input distribution

𝐺= General Service time distribution

Symbols for d

FCFS: first come first serve

LCFS: last come first serve

SIRO: service in random order

GD: general service discipline

Symbols for e

N: finite number of customers in the system

∞: Infinite number of customers in the system


Symbols for f

N: finite number of customers in the calling source.

∞: Infinite number of customers in the calling source.

Method1: Single server model with infinite population (M/M/1) :( FCFS/∞/∞)

(Single channel unrestricted queue with infinite population)

Let ƛ = mean or expeted number of arrivals per time period (mean arrrival rate)

μ = Mean or expected number of customers served per time period

(𝑚𝑒𝑎𝑛 𝑠𝑒𝑟𝑣𝑖𝑐𝑒 𝑟𝑎𝑡𝑒)


ƛ
1. Probability that the service facility is busy 𝜌 = 𝜇 , where 𝜌 = 𝑢𝑡𝑖𝑙𝑖𝑧𝑎𝑡𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟 = 𝑃

2. Probability that the service facility is idle.


ƛ
𝑃0 = 1 − 𝜇

3. Probability that there are n units in the system


ƛ
𝑃𝑛 = 1 − (µ)𝑛

4. Probability that there are more than or equal to k units in the system i.e., queue ≥ k
ƛ ƛ
𝑃(𝑛 ≥ 𝑘) = (𝜇)𝑘 Also 𝑃(𝑛 ≥ 𝑘) = (𝜇)𝑘+1

ƛ 𝑘 ƛ
𝑃(𝑛 = 𝑘) = (𝜇) (1 − µ)

ƛ
5. Expected or average number of units in the system or system length 𝐿𝑠 = 𝜇−ƛ

6. Expected or average number of units in the Queue waiting for service or length of queue
ƛ2
𝐿𝑞 = µ(µ−ƛ)

µ
7. Expected or average length of non-empty queue 𝐿𝑞 1 = µ−ƛ

1
8. Expected or average waiting time in the system 𝑊𝑠 = 𝜇−ƛ
ƛ 1
9. Expected or average waiting time in the queue 𝑊𝑞 = µ (𝜇−ƛ)

10. Probability that the waiting time in the queue is ≥ 𝑡.


ƛ
𝑃(𝑤𝑎𝑖𝑡𝑖𝑛𝑔 𝑡𝑖𝑚𝑒 ≥ 𝑡) = µ 𝑒 −(µ−ƛ)𝑡

Problems

1. Arrival of cars at a filling station is considered to be a Poisson variant with an average time of
4 minutes between one arrival and the next. The length of filling petrol is assumed to be
distributed exponentially with mean of 0.05 hours. Find

i) Average time an arrival spends in the system.

ii) Average time an arrival spends waiting in the queue.

iii) Average length of the system.

iv) Average length of the queue.

V) The new arrival rate if the management of petrol pump is determined to install second pump
is determined to install second pump under the condition that the average waiting time with the
present exceeds 12 minutes.
60
Sol: Arrival Rate ƛ = = 15 𝑐𝑎𝑟𝑠/ℎ𝑜𝑢𝑟
4

60
Service Rate μ = = 20 𝑐𝑎𝑟𝑠/ℎ𝑜𝑢𝑟
3

i) Average time an arrival spends in the system.


1 1
𝑊𝑠 = 𝜇−ƛ = 5 = 0.2ℎ𝑜𝑢𝑟𝑠 = 12𝑚𝑖𝑛

ii) Average time an arrival spends waiting in the queue.


ƛ 1 15
𝑊𝑞 = ( )= = 0.15ℎ𝑜𝑢𝑟𝑠 = 9𝑚𝑖𝑛
µ 𝜇−ƛ 20(20−15)

iii) Average length of the system.


ƛ 15
𝐿𝑠 = 𝜇−ƛ = 20−15 = 3 𝑐𝑎𝑟𝑠

iv) Average length of the queue.


ƛ2 15.15
𝐿𝑞 = µ(µ−ƛ) = 20(20−15) = 2.25𝑐𝑎𝑟𝑠

v) Let new arrival rate be ƛ

Given arrival waiting time in the system = 12 min

Hence the new arrival rate is same as the given arrival rate.

Method-II:

Single Server model with finite population .i.e., (M/M/1) :( FCFS/∞/𝑴)

(Single Channel unrestricted queue with finite population)

Let there be M units in the source population.


1
∶ average inter arrival time between successive arrivals.
ƛ

µ: service rate

𝑀! ƛ 𝑖
1. Probability that the system shall be idle 𝑃0 = [(𝑀−𝑖)! (𝜇) ]−1
ƛ 𝑛 𝑀!
2. Probability that there n customers in the system 𝑃𝑛 = 𝑃0 (𝜇) (𝑀−𝑛)!
if 0<𝑛<𝑚
𝑃𝑛 = 0 If 𝑛>𝑚
ƛ+𝜇
3. Expected length of the queue 𝐿𝑞 = 𝑀 − (1 − 𝑃0 )
ƛ

𝜇
4. Expected number of customers in the system 𝐿𝑠 = 𝑀 − ƛ (1 − 𝑃0 )

1 𝑀 ƛ+𝜇
5. Expected waiting time of a customer in the queue 𝑊𝑞 = 𝜇 [1−𝑃 − ]
0 ƛ

1 𝑀 ƛ+𝜇 1
6. Expected time a customer spends in the system 𝑊𝑠 = 𝜇 [1−𝑃 − + 1] = 𝑊𝑞 + 𝜇
0 ƛ

Problem

1. In a factory having 5 machines, breakdown occurs randomly and the average time
between the breakdowns is 2 days. Assuming that the repairing capacity of the workman
is one machine a day and the repairing is distributed exponentially. Find
i) Probability that the service facility is idle.
ii) Probability of various numbers of machines to be and being repaired.
iii) Expected length of the queue.
iv) Expected number of machines waiting to be and being repaired.
v) Expected time a machine shall wait in the queue to be repaired.
vi) Expected time a machine shall be idle for reason, for waiting to be and being repaired.
Sol: M = 5 machines

Service rate 𝜇 = 1 machine/day


1
ƛ = 2 Machine/day

ƛ
∴ 𝜌 = 𝜇 = 0.5

Since 𝑃0 is basic to all formulae for this model, we shall first determine its value.

i 𝑀! ƛ 𝑖 𝑀! ƛ 𝑖
(𝑀 − 𝑖)! ( ) ( )
𝜇 (𝑀 − 𝑖)! 𝜇
0 1 1.0000 1
1 5 0.5000 2.5
2 20 0.25 5.00
3 60 0.125 7.5
4 120 0.0625 7.5
5 120 0.03125 3.75
27.25

𝑀! ƛ 𝑖
𝑃0 = [ ( ) ]−1 = 0.0367
(𝑀 − 𝑖)! 𝜇

i) Probability that the service facility is idle.


𝑀! ƛ 𝑖 −1
𝑃0 = [ ( ) ] = 0.0367
(𝑀 − 𝑖)! 𝜇
ii) Probability of various numbers of machines to be and being repaired.
ƛ 𝑛 𝑀!
𝑃𝑛 = 𝑃0 (𝜇) (𝑀−𝑛)!
𝑃1 = 0.09175 , 𝑃2 = 0.18350, 𝑃3 = 0.27525, 𝑃4 = 0.27525, 𝑃5 = 0.13762
iii) Expected length of the queue.
ƛ+𝜇
𝐿𝑞 = 𝑀 − ƛ (1 − 𝑃0 ) = 2.11 𝑚𝑎𝑐ℎ𝑖𝑛𝑒𝑠
iv) Expected number of machines waiting to be and being repaired.
𝜇
𝐿𝑠 = 𝑀 − ƛ (1 − 𝑃0 ) = 3.07 𝑚𝑎𝑐ℎ𝑖𝑛𝑒𝑠
v) Expected time a machine shall wait in the queue to be repaired.
1 𝑀 ƛ+𝜇
𝑊𝑞 = 𝜇 [1−𝑃 − ] = 2.19𝑑𝑎𝑦
0 ƛ
vi) Expected time a machine shall be idle for reason, for waiting to be and being
repaired.
1
𝑊𝑠 = 𝑊𝑞 + = 3.19𝑑𝑎𝑦𝑠
𝜇
Method-III:
Single channel, restricted queue with infinite population .i.e., (M/M/1) :( FCFS/𝑵/∞)

In this system the maximum size of the queue permissible is N. This model is also known
as finite queue model.

Ex: - A Doctor who gives appointment of a maximum number N per day.

Assumptions of this model

1. The arrivals follow Poisson distribution with a mean arrival rate.

2. The service time has exponential distribution with an average service rate.

3. Arrivals are from infinite population.

4. The customers are served on FCFS basis.

5. There is only single service station.

6. The mean arrival rate is less than mean service rate.

7. The waiting space available for the customer in the rate queue is finite given by N.

Let the maximum permissible number of units in the system be N.


1−𝜌 ƛ
1. Probability that the service facility is idle 𝑃0 = 1−𝜌𝑁+1 , here 𝜌 = µ

(1−𝜌)𝜌𝑛
2. Probability that there are n units in the system 𝑃𝑁 = = 𝑃0 𝜌𝑛 , 𝜌 ≠ 1, 𝑛 = 0,1,2, − −
1−𝜌𝑁+1
−−−−𝑁

𝜌 (𝑁+1)𝜌𝑁+1
3. Expected or average number of customers in the system 𝐿𝑠 = 1−𝜌 − 1−𝜌𝑁+1

𝐿
4. Expected or average waiting time in the system 𝑊𝑠 = ƛ1𝑠 where ƛ1 = ƛ(1 − 𝑃𝑁 ) effective
arrival rate
1
5. Expected or average waiting time in the queue 𝑊𝑞 = 𝑊𝑠 − 𝜇

1
Where 𝜌 = 1 → 𝑃𝑛 = 𝑁+1

𝑁−1 𝑁(𝑁−1)
𝐿𝑠 = , 𝐿𝑞 =
2𝜇 2(𝑁+1)

Problem
1.In a bank with single server , there are 3 chairs for waiting customers. on a average one
customer arrive every 7.5min and each customer takes 6 min for getting served. Making suitable
assumptions. Find

a) The probability that arrival will get a chair to sit down.

b) The probability that an arrival will have to stand.

c) Expected waiting time of a customer.

Sol: The assumptions are

a) Arrivals follows Poisson Distribution.

b) Service time follows Exponential Distribution.

c) Customers are served on FCFS basis.


60
G.T 𝐴𝑟𝑟𝑖𝑣𝑎𝑙 𝑅𝑎𝑡𝑒 ƛ = 7.5 = 8 𝑝𝑒𝑟 ℎ𝑜𝑢𝑟

60
𝑆𝑒𝑟𝑣𝑖𝑐𝑒 𝑅𝑎𝑡𝑒 𝜇 = = 10 𝑝𝑒𝑟 ℎ𝑜𝑢𝑟
6

Max permissible capacity N=3

a) The probability that arrival will get a chair.

i.e. The probability that the service facility is idle.


8
1−𝜌 1−
10
𝑃0 = 1−𝜌𝑁+1 = 8 = 0.339
1−( )3+1
10

b) The probability that an arrival will have to stand.

i.e., Probability that there are N units in the system

𝑃𝑁 = 𝑃0 𝜌3 = 0.174

c) Expected waiting time of a customer.


1 𝑠𝐿
𝑊𝑞 = 𝑊𝑠 − 𝜇 where 𝑊𝑠 = ƛ(1−𝑃
𝑁)

𝜌 (𝑁+1)𝜌𝑁+1
𝐿𝑠 = 1−𝜌 − = 1.225
1−𝜌𝑁+1

𝑠 𝐿
𝑊𝑠 = ƛ(1−𝑃 = 0.185
𝑁)
1
𝑊𝑞 = 𝑊𝑠 − 𝜇 = 0.085ℎ𝑟𝑠 = 5.1𝑚𝑖𝑛

Method-IV:

Multi server, unrestricted queue with infinite population

Multi channel queuing theory treats the conditions in which there are several service
stations in parallel and each element in the waiting line can be served by more than one station.
Each service facility is prepared to deliver the same type of service. The new arrival selects one
station without any external pressure.

When a waiting line is formed, a single line usually breaks down into shorter line in front of
each service station. The arrival rate ƛ and service rate 𝜇 are mean values from Poisson and
exponential distribution respectively.

Service discipline is first come first served and customers are taken from a single queue. If
there is any empty channel it is filled by the next customers in line.

Let n = number of customers in the system

𝑃𝑛 = 𝑃𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑜𝑓 𝑛 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠 𝑖𝑛 𝑡ℎ𝑒 𝑠𝑦𝑠𝑡𝑒𝑚

k = number of parallel service channel (𝑘 > 1)

ƛ=arrival rate of customers

µ = service rate of individual channel

When 𝑛 < 𝑘 there is no queue and all arrivals are being served and rate of service will be n.µ
as only n channels are busy. When n=k all channels will be working.

When 𝑛 > 𝑘 there will be (n-k) persons in the queue and rate of service will be kµ as all the k
channels are busy.

1. Probability that there are no customers or units in the system


1 ƛ 𝑛 1 ƛ 𝑘 1
𝑃0 = [∑𝑘−1
𝑛=0 𝑛! (µ) + 𝑘! (µ) ( ƛ )]−1
1−
𝑘µ
2. The probability that there are n units in the system
ƛ 𝑛
( )
µ
𝑃𝑛 = 𝑃0 𝑖𝑓 𝑛 < 𝑘
𝑛!
ƛ 𝑛
( )
µ
𝑃𝑛 = 𝑘!𝑘 𝑛−1
𝑃0 𝑖𝑓 𝑛 > 𝑘
1 ƛ 1
3. Probability that customers have to wait 𝑃(𝑛 ≥ 𝑘) = 𝑘! (µ)𝑘 ƛ 𝑃0
[1− ]
𝑘µ

4. Average queue length (Expected number of customers waiting in the queue)

ƛ 𝑘
ƛµ( )
µ
𝐿𝑞 = (𝑘−1)!(𝑘𝜇−ƛ)2 𝑃0

5. Average (expected) number of customers in the system

ƛ 𝑘
ƛµ( ) ƛ ƛ
µ
𝐿𝑠 = (𝑘−1)!(𝑘𝜇−ƛ)2 𝑃0 + µ = 𝐿𝑞 + µ

6. Average waiting time of an arrival (expected time a customer spends in the queue waiting for
service)

ƛ 𝑘
µ( ) 𝐿𝑞
µ
𝑊𝑞 = (𝑘−1)!(𝑘𝜇−ƛ)2 𝑃0 = ƛ

7. Average time on arrival spends in the system

ƛ 𝑘
ƛµ( ) 1 𝐿𝑠
µ
𝑊𝑠 = (𝑘−1)!(𝑘𝜇−ƛ)2 𝑃0 + 𝜇 = ƛ

ƛ
8. Utilization factor 𝜌 = 𝑘𝜇

Problem

1. A tax consultant firm has 4 service stations (counters) in its office to receive people who have
problems and complaints about their income, wealth and sales taxes. Arrivals average
100persons in 10 hours service day. Each tax adviser spends an irregular amount of time
servicing the arrivals which has been found to have an exponential distribution. The average
time is 20min. calculate 𝑃0 , 𝐿𝑞 , 𝐿𝑠 , 𝑊𝑞 , 𝑊𝑠 , 𝑃(𝑛 > 𝑘).
100
Sol: G.T ƛ = = 10𝑝𝑒𝑟 ℎ𝑜𝑢𝑟
10

60
𝜇 = 20 = 3𝑝𝑒𝑟 ℎ𝑜𝑢𝑟

k=4
ƛ 10
𝜌 = 𝑘𝜇 = 12

i) Probability of no customers in the system

𝑘−1
1 ƛ 𝑛 1 ƛ 𝑘 1
𝑃0 = [∑ ( ) + ( ) ( )]−1 = 0.0208
𝑛! μ 𝑘! μ ƛ
𝑛=0 1−
𝑘μ

ii) Average queue length

ƛ 𝑘
ƛµ( )
µ
𝐿𝑞 = (𝑘−1)!(𝑘𝜇−ƛ)2 𝑃0 = 3.24 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠

iii) Average (expected) number of customers in the system


ƛ
𝐿𝑠 = 𝐿𝑞 + = 6.567 𝑐𝑢𝑠𝑡𝑜𝑚𝑒𝑟𝑠
µ
iv) Average waiting time of an arrival (expected time a customer spends in the queue
waiting for service)
𝐿𝑞
𝑊𝑞 = = 0.3324ℎ𝑟𝑠
ƛ

v) Average time on arrival spends in the system


𝐿𝑠
𝑊𝑠 = = 0.6567ℎ𝑟𝑠
ƛ

1 ƛ 1
vi) Probability that customers have to wait 𝑃(𝑛 ≥ 𝑘) = 𝑘! (µ)𝑘 ƛ 𝑃0 = 0.618
[1− ]
𝑘µ

Dynamic Programming
Identify the decision variables


Specify the objective function


Decompose in to smaller problems (Stages)


Identify the static variables


State the general recursive relations. Also decide forward /forward


Calculate required values of return function by appropriate formulae at each
stage


Repeat and determine the overall optimal policy

Introduction

Dynamic Programming is an optimization technique of multistage decision process. The word


Dynamic means to the situations of change in several stages, such as every week, every month
etc and programming is the term used in the mathematical sense of selecting an optimum
allocation of resources.

In Dynamic programming a large problem is split into smaller sub problems each of which
involve in a few variables. These sub problems are sequentially optimized so as to make the total
problem optimal.

The solution of a Dynamic Programming Problem is based on Bellman’s principle of


optimality (Recursive optimization technique) which states “An optimal policy has the property
that, whatever the initial state and initial decision, the remaining decisions must constitute an
optimal policy with regard to the state resulting from the first decision”.

Discrete and continuous, deterministic as well as probabilistic models can be solved by this
method. Thus Dynamic Programming method is very useful in obtaining optimal solution of
various problems such as inventory, replacement, allocation, linear programming, Reliability
calculations etc. A Single constraint problem is relatively simple, but in the problem of more
than two constraints more complexities may appear.

Features and their definitions of Dynamic Programming Problem

1. Recursive Function: - A Dynamic Programming Problem is solved by an objective


function which is recurring in nature.
a) Forward recursive approach: - In this approach the main problem is attacked by
dividing it from initial stage to final stage.
b) Backward recursive approach:- In this approach the problem is dealt by sub dividing
the problem from last to first and then solving from final to initial stages.

2. Stage; - A large problem is decomposed in to smaller and sequential sub problems. Each
sub problem is referred to as a stage where a decision is called for. Thus each stage can be
considered to have a starting and ending. The ending of one stage will be the beginning of its
next stage.

3. State:-Any stage just before starting or just after completion will have certain status called
the state and the variable that links up to stages is called state variable.

4. Return function: - The decision of a stage described in the form of an algebraic equation
with state variables to explain the worth or benefit or cost is known as Return function or
transformation equation.

Characteristics of Dynamic Programming Problem

The basic characteristics of the Dynamic Programming Problem are given below.

1. The problem can be sub divided into sub problems referred to as stages with policy
decision at each stage.
2. A stage is a device to sequence the decisions. This means that the sequential summation
of these decisions obtained from optimal solution of sub problems will be the part of the
optimal solution of the total problem.
3. Every stage consists a number of states associated with it. The states are different
possible conditions in which system may find itself at that stage of the problem.
4. Decision at each stage convents the current stage into state associated with the next stage.
5. The state of the system at stage is described by a set of variables called state variables.
6. When the current state is known, an optimal policy for the remaining stages is
independent of the policy of the previous stages.
7. To identify the optimum policy for each state of the system. A recursive equation is
formulated with n stages remaining given the optimal policy for each state with (n-1).
8. Using recursive equation approach each time the solution moves either backward or
forward by stage obtaining the optimum policy of each state for that particular stage, till
it attains the optimum policy beginning at final or initial stage respectively.

Problems:

1. A Medical representative located at city 1 has to travel to city 10. He knows the
distance of alternative routes from city 1 to city 10 and has drawn the network
map based on the distance between the cities as in the following table. Draw the
network and find the shortest possible route. Also, find the shortest routes from
any city to city 10. (March-2017 R-13)

From To Corresponding
city city distance in km
1 2,3,4 4,6,3
2 5,6,7 7,10,5
3 5,6,7 3,8,4
4 5,6,7 6,10,5
5 8,9 4,8
6 8,9 3,8
7 8,9 8,4
8 10 7
9 10 9

The given problem can be divided into 5 stages as shown in the figure. It is sived in the
tabular form method as shown below.

At stage 1 if the medical representative is at city 10, it is the destination itself, he need
not travel any more.

Stage 2: At stage 2 i.e if he is at city 8 or city 9 the shortest path calculation is shown
below.
Decisions Destinations Minimum Optimal
State Stage distance Cumulative distance decision
variable(S1) distance
Entering 8 7 7 7 8-10
state 9 9 9 9 8-10

Stage-3:

Now using these optimal results of stage 2 we proceed stage 3

Decisions Destinations Minimum Optimal


(Optimal ) decision Cumulative
State City8 (7) City 9(9) distance optimum
variable(S2) decision
Stage Cumulative Stage Cumulative
distance optimum distance Distance
decision
Entering 5 4 11 8 17 11 5-8 5-8-10
state
6 3 10 7 16 10 6-8 6-8-10
7 8 15 4 13 13 7-9 7-9-10

Stage-4: With the above optimal decision we complete stage-4 decisions


Decisio Destinations Minim Optimal
ns um decision Cumulativ
State City5 (11) City 6(10) City 7(13) (Optim e
variable Cumula Stage Cumulati al ) optimum
Stage distanc decision
(S2) Stage Cumulative tive distanc ve
distan e
distance optimum Distanc e Distance
ce
decision e
Ente 7 18 10 20 5 18 18 2-5 2-5-8-10
ring 2 (or)2-7 (or)2-3-9-
state 10
3 3 14 8 18 4 17 14 3-5 3-5-8-10
4 6 17 10 20 5 18 17 4-5 4-5-8-10

Stage-5: With the optimal distance decisions we now proceed to calculate decision table of
stage-5

These two optimal solutions are obtained as follows.

Solution set-1: Sequence (City)1-3-5-8-10

Distance(km): 6+3+4+7=20 km
Solution set-2: Sequence(City) 1-4-5-8-10

Distance: 3+6+4+7=20 km

We can thus state that medical representative may have to travel minimum number of 20 km
from city 1 to city 10.

2) Consider the problem of allocating salesmen to three zones in such a way that the total
profits of the company are maximized. The following table gives the details of the zones,
number of salesmen and the corresponding profits.

No. of sales men Profits (In thousands)


Zone1 Zone 2 Zone 3
0 30 35 42
1 45 45 54
2 60 52 60
3 70 64 70
4 79 72 82
5 90 82 95
6 98 93 102
7 105 98 110
8 100 109 110
9 90 100 110

Sol. Stage-1:

x 0 1 2 3 4 5 6 7 8 9
F1(x) 30 45 60 70 79 90 98 105 100 90

This table gives the profit corresponding to different salesmen allocated to zone 1

F1(x)= f1(x)

Stage-2(Zones1 &2):

If S denotes the total profits from each combination

We have the general form

S= f2 (x)+ F1(A-x)

Where A is the total no. of salesmen.


We have to maximize

F2 (A)=Max(f2 (x)+ F1(A-x))

This equation can be applied to determine the optimal distribution of any number of salesmen
(A=9,8,7,6,5) as shown in table

Zone 1 0 1 2 3 4 5 6 7 8 9 No. Of
sales men
Zone 2 30 45 60 70 79 90 98 105 100 90 Profits
0 35 65 80 95 105 114 125 133 140 135 125

1 45 75 90 105 115 124 135 143 150 145

2 52 82 97 112 122 131 142 150 157

3 64 94 109 124 134 143 154 162

4 72 102 117 132 142 151 162

5 82 112 127 142 152 161

6 93 123 138 153 163

7 98 129 143 158

8 109 130 145

9 100 130

The maximum profit for each value of A are represented. The above table shows the profits for
all combinations.

A 0 1 2 3 4 5 6 7 8 9
S 65 80 95 105 115 125 135 143 154 163
0+0 0+1 0+2 1+2 1+3 0+5 1+5 3+4 3+5 6+3
(or) (or)
0+3 1+6
We have

F2 (A)=Max(f2 (x)+ F1(A-x))

Therefore S= f3 (x)+ F2(A-x)

F3 (A)=Max(f3 (x)+ F2(A-x))

Zone 1+zone 2 0 1 2 3 4 5 6 7 8 9 No. Of


sales men
Zone 3 65 80 95 105 115 125 135 143 154 163 Profits
0 42 107 122 137 147 157 167 177 185 196 205

1 54 119 134 149 159 169 179 189 197 208

2 60 125 140 155 165 175 185 195 203

3 70 135 150 165 175 185 195 205

4 82 147 162 177 187 197 207

5 95 160 175 190 200 210

6 102 167 182 197 207

7 110 175 190 205

8 110 175 190

9 110 175

A= (x1+x2)+x3

A 0 1 2 3 4 5 6 7 8 9
S 107 122 137 149 159 169 179 190 200 210
0+0 0+1 0+2 1+2 1+3 1+4 1+5 5+2 5+3 5+4

Thus for A=9 we find that the profit is maximum for x1+x2=4 and x3=5. From stage II we find
that for A=4, the profit has maximum value 115 with x1=3, x2=1.
Therefore the optimal distribution of salesmen is

Zone-1→3.

Zone-2→ 1

Zone-3→ 5

Max profit =Rs. 2,10, 000

3.A truck can carry a total of 10 tons of a product. Three types of product are available for
transport. Their weight and values are given below. Determine the loading such that the total
value of the cargo is maximum (At least one unit of each type must be loaded)

Type Value(Rs.) Weight(Tons)


A 20 1
B 50 2
C 60 2

Ans. We have to determine how many units of A,B,C is to be loaded. We take the loading of
type C as stage-1.

Stage-I:

State x1
No. of units Optimal value
(Weight)
1 2 3 Value Decision
C11=60 C12=120 C13=180 F1A)
W=2 W=4 W=6
2 60 - - 60 1
3 60 - - 60 1
4 60 120 - 120 2
5 60 120 - 120 2
6 60 120 180 180 3
7 60 120 180 180 3

One unit in each of A and B must be taken.

Hence x1 cannot exceed 7 tons.

We can take C and B for stage-II


F2(A)=Max{f2(x)+F(A-x)}

Stage-II:

State x2
No. of units Optimal value
(Weight)
1 2 3 Value Decision
C21=50 C22=100 C33=150 F2(A)
W=2 W=4 W=6
4 50+6 100 - 110 1
5 50+60 100 - 110 1
6 50+120 100+60 150 170 1
7 50+120 100+60 150 170 1
8 50+180 100+120 150+60 230 1
9 50+180 100+120 150+60 230 1

Stage-III: C,B and A are taken.

F3(A)=Max{f3(x)+F2(A-x)}

State x3 No. of units


Optimal value
(Weight)
1 2 3 4 5 6 Value Decision
C31=20 C32=40 C33=60 C34=80 C35=100 C36=120 F3(A)
W=1 W=2 W=3 W=4 W=5 W=6
5 20+110 - - - - - 130 1
6 20+110 40+110 - - - - 150 2
7 20+170 40+110 60+110 - - - 190 1
8 20+230 40+170 60+110 80+110 - - 210 2
9 20+230 40+170 60+170 80+110 100+110 - 250 1
10 20+230 40+230 60+170 80+170 100+110 120+110 270 2

We find that the max. value is Rs. 270

The loading is A-2 units, B-1 unit, C- 3 units.

Weight= 2+2+6=10 tons.

You might also like