KENYA METHODIST UNIVERSITY

Department of Economics and Applied Statistics

BUSS 320

QUANTITATIVE METHODS IN

BUSINESS
NJOROGE CHRISTOPHER

Open and Distance Learning Instructional

Material
 Published by Kenya Methodist University
P.O. BOX 267 – 60200, MERU
Email: info@kemu.ac.ke
TEL: 254 – 064 – 30301, 31146/0736752262

Quantitative Methods in Business

BUSS 320
© KeMU, 2014
All rights reserved.
No part of this module may be reproduced, stored in any retrieval
system or transmitted in any form or by any means,
electronically, mechanically, by photocopying or otherwise;
without the prior written permission of Kenya Methodist
University.

3
COURSE OVERVIEW
I welcome you to the study of Quantitative techniques. In this course we shall study the application of
mathematical models in solving business related problems. The course borrows from economics and
statistics. An understanding of basic economic and statistical concepts is essential. The purpose of this
course is to equip students with quantitative techniques that will enable them to make sound and
competent management decisions.
The topics are arranged in 12 weeks and each week has one or more lectures depending on the topic’s
depth. There are 15 lectures and each lecture has its own objectives that you should achieve. At the end of
every lecture, you will find a series of SAQs that are meant to help you to evaluate your understanding
of the concepts presented. You should attempt all the questions and activities once you have finished
studying the relevant work. A summary of each lecture is also provided at the end with a list of further
resources that you are expected to read and make notes from.
By the end of this course the student should:
1. Understand basic mathematic concepts and appreciate statistical calculations
2. Be in a position to identify specific quantitative techniques that are applicable to particular
organizational problems
3. Apply various quantitative models in solving business problems. Students are encouraged to apply
concepts learnt in class so as to master their use. They are expected to have an analytical mind.
4. Interpretation of results derived from computer solutions arrived at through application of specific
mathematical models

Kindly, make sure that:

 You complete each lecture at a time before proceeding to the next one.
 Refer to the suggested additional resources to get further information on each topic
 Make notes so as to simplify your study
 Complete all activities and questions as you progress
 Spend at least 3(three) hours to complete each topic for you to understand and apply the knowledge
and skills acquired
Once again welcome and let us begin. Good luck!!!!

4
COURSE OBJECTIVES

By the end of this course the student should:

1. Understand basic mathematic concepts and appreciate statistical calculations
2. Be in a position to identify specific quantitative techniques that are applicable to particular
organizational problems.
3. Apply various quantitative models in solving business problems. Students are encouraged
to apply concepts learnt in class so as to master their use. They are expected to have an
analytical mind.
4. Interpretation of results derived from computer solutions arrived at through application of
specific mathematical models

5
REFERENCES & SUPPLEMENTARY MATERIAL
Barry Render, Ralph Stair, Michael Hanna, (2006) Quantitative Analysis for Management,
Prentice Hall Kothari, C. R. (2000), Quantitative Techniques, Vikas Publishing House, New Delhi
Lucey T , Quantitative Techniques (2002), Thomson Learning, London, UK
Saleemi N A (1997), Quantitative Techniques Simplified, N A Saleemi publishers, Nairobi, Kenya
Srivastra U, Shenoy G and Sharma S (1999) Quantitative Techniques for Managerial Decisions,
New Age International

Rendelinghuis A, Julyan F, Steyn B and Benade F (1992), Quantitative Methods for Managerial
Decision making, Butterwrths, Durban

6
SYMBOLS

Objectives

Activity

! Key note

Summary

Self-Assessment Questions

7
 Suggested further reading

EVALUATION & GRADING

This course is evaluated in line with other courses offered in the Bachelor of Business
Administration degree at Kenya Methodist University. The total marks are categorized into 2
parts – the coursework & the final examination
The coursework is made up of a sit-in CAT and an individual assignment. This caters for 30% of
the final mark. The final exam is made up of 5 questions. Only three questions are to be attempted
in total. Question one carries 30 marks and is compulsory. It is structured to have a broad
coverage of the course. Only two questions are to be attempted out of the other four. Each of
these questions carries 20 marks. The Course Assessment may be summarized as below:

Coursework:
CATs & Assignments 30%
End of trimester exam 70%
Total 100%

The final score, made up of the coursework and the exam is then graded using a GPA system as
below:

GRADE MARKS GPA POINTS

A 75 and above 4.0
B+ 68-74 3.5

8
 B 60-67 3.0
C+ 55-59 2.5
C 50-54 2.0
D+ 45-49 1.5
D 40-44 1.0
E Under 40 0.0

NOTE!
For Distant Learning Students, the coursework is made up of a comprehensive assignment. This
should be submitted in time through the ODLM office so that grading is done in time. There is
revision of the assignment during the tutorial at the end of the trimester, just before the end of
trimester exams.

COURSE OUTLINE
The purpose of this course is to equip students with quantitative techniques that will enable them to
make sound and competent management decisions. The course is covered in 11 lectures over the 12
weeks in the trimester as outlined below:
Week Topics Content
1 Lecture 1  Definition of Quantitative Methods
Introduction to  Components of a mathematical model
Quantitative techniques  Important models in quantitative analysis

Lecture 2  Definition and types of matrices

Types of Matrices:  Order of a matrix
Addition, Subtraction and  Addition and subtraction of matrices
multiplication of matrices  Multiplication of matrices
2 Lecture 3  Determinant of a matrix
Matrix inversion and  Inverse of a matrix including order 3x3 matrix

9
Week Topics Content
application of matrix  Application of matrix algebra:
algebra o Solving Linear / simultaneous equations:
Inverse method and crammers rule
o Input-output analysis
3 Lecture 4  Terminology in network analysis
Network Analysis  Rules for drawing network diagrams
 Critical path analysis [CPA]
 Advantages of network analysis
 Distinction between PERT and CPA
 Limitations of PERT and CPA
4 Lecture 5  Introduction
Linear Programming  Requirements of linear programming
 Assumptions of linear programming
 Formulation of linear programming problems
 Graphical solution to LP problems
5 Lecture 6  General structure of the problem
Transportation Model  Requirements of a transportation problem
 Deriving the initial solution using various methods
6 CAT [THERE WILL BE NO MAKE-UP CATS]
7 Lecture 7  Evaluating whether initial transportation solution is
Transportation Model… optimal: Computation of shadow price.
 Deriving the optimal solution.
Assignment Model  The assignment problem including both maximization
and minimization problem
8 Lecture 8.  Differential calculus
Calculus and its  Integral calculus
application in business  Economic applications of differentiation
decisions  Economic application of integration
9 Lecture 9  Assumptions in game theory
Game theory  Basic concepts in game theory
 Use of pure strategies to solve problems
 Dominance
 Limitations of games in competition
10 CAT 2 CAT [THERE WILL BE NO MAKE-UP CATS]
11 Lecture 10  Components of a queuing system
Queuing theory  Queuing situations
 Queue elements

10
Week Topics Content
 Classes of queue systems
 Simple queue characteristics
 Terminologies used in queuing theory
 Assumptions of the model
 Solutions of problems involving single channel
Poisson arrivals with exponential service rate
12 Lecture 11  The process of simulation
Simulation  Importance of simulation to managers
 Areas of application of simulation in an organization
 Limitations of simulation
13-14 End of trimester exams

Teaching Methodologies
1 Lectures/instructional Notes
2 Assignments / Questions
3 Discussions/Case Studies Analysis

LECTURE 1
INTRODUCTION TO QUANTITATIVE METHODS
Chapter Objectives:
By the end of this topic, students should be able to:
o Understand what quantitative techniques is all about
o Describe the phases of managerial decision process
o Explain the components of a mathematical model for decision making

11
 o Relate different quantitative models to problem solving scenario

Definitions of Quantitative analysis

There are different definitions of quantitative methods. A few have been sampled below:
 “Process of analyzing information and making decisions using numerical data”
 ‘Methods incorporating multiple conflicting criteria into the management planning process’
 ‘A decision-support tool allowing decision makers to solve their problem by evaluating, rating,
and comparing different alternatives’
 ‘Methods that convert data to numerical indicators, and then analyze these numbers using
statistics to establish relationships among the concepts’
 ‘Methods that seek to obtain information about the social world that is already in or can be
converted to numeric form’
 ‘Numerical means to drawing sociological conclusions using powerful statistical tools to help
understand patterns in which the behaviors, attitudes, or traits under study can be translated
into numerical values’

It is also referred to in different terms:’

o Quantitative techniques
o Management mathematics
o Operations research
o Business mathematics
o Management science

 Quantitative analysis assists the management of an organization in:

o Planning the allocation of resources at their disposal
o Control the use of resources in an efficient / effective way in order to achieve
predetermined objective
 Resources at the disposal of management may include:
o Personnel
o Materials

12
 o Machinery
o Money
o Time
In short QM is the range of mathematical and statistical techniques used to analyze data.
Quantitative methods have been widely used by political scientists in a range of contexts,
including, for example, the study of arms races, of political stability, of political violence, and of
the behaviour of legislators, but by far their most prominent application has been in the area of
electoral attitudes and behaviour, where data are easily quantified.
While such methods have enhanced the study of politics, there have been criticisms of quantifying
for the sake of it, of equating results obtained with the results of scientific experiments
(misapplying the methods of the natural sciences to social data), and overemphasizing numbers at
the expense of explanation (the establishment of the existence of a statistically significant
correlation or regression coefficient may say little about its meaning). Such criticisms have led
some to a more restrained and cautious use of quantitative methods.
Levels of managerial decisions
The levels at which management decisions are made are generally categorized into three:

 Strategic decisions (Top Management)

 Managerial decisions (Mid-level Management)

 Operating decisions (Low-level Management)

Top Management's Critical Decisions

According to the Free Management Library, managers should approach decision making in an
organized manner. Top managers in a company include the board of directors or a chief executive
officer; these executives make critical company decisions related to the corporate strategic planning
and organizational development of the company. These top managers may decide how to handle a
major crisis and what product to launch or manufacture. They also identify competitors, create a

13
corporate vision for the company, decide on mergers and acquisitions, develop budgets and set
long-term goals.

Middle-Level Management Decisions

Most non-critical decisions are delegated to middle level management. Top management relies
heavily on middle management to make the right decisions. An effective leader allows his
management team to make decisions without micromanaging them and fully supports their
decisions. Middle management may handle tactical decisions, oversee the regional market and
decide how to meet the company's short-term goals. Middle management decisions might include
marketing a new product, communicating with and managing lower management and determining
what issues need to be addressed with top-level managers. Each individual middle management
department develops a strategy to meet its inner-departmental goals.

Lower-Level Management Decisions

According to the U.S. Small Business Administration, common decision making mistakes occur
when a manager only hears or sees what she wants to. Operational decisions affect daily tasks and
are generally handled by lower-level management. Lower-level managers should identify what
impact their decisions will have on themselves and others. Supervisors or team leaders may decide
employee-related issues, such as pay rates, training, evaluations, raises, overtime, promotions,
hiring and disciplining or terminating employees. A supervisor at this level may decide to reward
the most productive employee with an employee of the month award, or offer incentives such as
movie tickets or gift certificates.

Identify the three levels at which managerial decisions are made. Outline three types of
decisions are made at each level

Phases of Managerial Decision Problem

14
Excellent managerial decision-making accounts for the difference between businesses that grow fast
and businesses that don't. Managers and entrepreneurs who rise to the top and take their
companies with them have usually developed habits and systems for making difficult decisions.
The system of managerial decision-making can be narrowed down to 7 core steps:

1. Identify the core problem/core opportunity

The apparent complexity of business problems usually obscures the fact that most visible
"problems" are merely symptoms or effects cascading from one root problem (the cause). This is a
general phenomenon of systems. Complex chains of cause and effect in a system mean that the
greatest changes can be brought about when adjustments are made at the root.
On the other hand, identifying the core opportunity in a marketplace can be just as complex as
identifying a core problem. Often, it requires the ability to mine the customer landscape to
identify their "core problem", before you reverse-engineer a product or solution for it and offer
that solution to the customer.

2. Brainstorm your options

In this phase, your job becomes either to know or to discover the full breadth of alternatives
available to you. 2 factors that differentiate successful business leaders start to show up in this
phase.
A manager with a great deal of experience is able to draw up a wider set of potential directions to
choose from. Separately, a manager who is disciplined at execution is able to research unknown or
new options at a much faster pace, and replicate the advantages of much deeper experience.

3. Analyze Options
Management decision-making in today's world is an established science in its own right. There are
literally hundreds of strategic thinking tools applied by managers to analyze options and make
decisions. However, many of these tools are useless if the core assumptions under-girding the
analysis are wrong or misguided. You must make sure that your managerial decision-making
process includes ways to test assumptions as well as a system for cycling back if assumptions turn
out to be wrong.

15
4. Make a Decision
The point of analysis is for managerial decision-making to be a faster and better process that leads
to the attainment of business goals. However, even with excellent analytical tools, a leader's
personal decision-making style can affect the speed at which decisions are made, and even the
quality of the final decision. Decisive managers accept that bad outcomes can result from good
decision-making procedures and take solace in the quality of their preparation and their process.

5. Take Action
Depending on the size of the organization you lead, the gap between resolution and taking action
can be quite wide and time consuming. A bias for action is the single biggest determinant to
making better and better decisions over the long run. There are just too many details and too
much complexity in the business environment for a manager to attempt to be completely
prescient.
To paraphrase business adviser Dan Kennedy, the best chance you have is to figure out the things
that don't work (and won't work) as quickly and as cheaply as possible. That requires a culture of
rapid decision and action.

6. Review Results
As a consultant, one of the most pervasive phenomena I have encountered in business is that of the
business leader who hates to go back; who hates to review past work. Whether due to personality
or habit, many entrepreneurs find the monitoring and testing of mundane business details to be
distasteful at best.
Constant monitoring is one of the commonalities found in fast growing businesses. Recently, I
spoke with a software manufacturer who found increased sales by monitoring buyer behavior. He
and his team observed that a few web visitors who were abandoning the web site during the
shopping cart process were coming back 2 or 3 days later to make a purchase. When they provided
web shoppers with a "Save shopping cart" feature, purchases increased significantly.

16
7. Implement changes
The point of monitoring and tracking decisions and their results is to make improvements. Not
just improvements in products and processes, but improvements in overall decision making.
Making a habit of systematic change, will help you build a more nimble and profitable
organization.

What do you consider to be the key stages in the decision making process?
Mathematical Models
A mathematical model is a description of a system using mathematical concepts and language. The
process of developing a mathematical model is termed mathematical modeling. Mathematical
models are used not only in the natural but also in the social sciences such as economics,
psychology, sociology, and political science. Models may also help to explain a system and to
study the effects of different components, and to make predictions about behaviour.
Mathematical models can take many forms, including statistical, differential equations, game
theory e.t.c. In many cases, the quality of a scientific field depends on how well the mathematical
models developed on the theoretical side agree with results of repeatable experiments. Such
mathematical models need to be compared with experimental measurements in order to advance
the theories and test their validity.
Components of a Mathematical Model in Solving Business Problems
 An objective function – Mathematical statement of what the management wishes to achieve
once the problem is solved optimally
 Decision variables - the unknowns to be determined by the solution to the model.
 Constraints - the physical limitations of the system, especially applying to resources
The objective of the model is to generate a solution (or optimal solution) to the problem. It involves
identification of a set of the decision variable values which are feasible (i.e. satisfy all the
constraints) and selecting the best which lead to the optimal value of the objective function.
Mathematical models covered
Matrix algebra – This model will be applied in solving linear simultaneous equations and the
input output analysis

17
Network Analysis – This model deals with planning for long-term projects by b
Linear Programming – This model deals with allocation of scarce resources so as to either
maximize / minimize
Transportation and Assignment Models – The transportation model deals with transporting
goods from several sources to given destinations. In the assignment model, personnel are to be
assigned tasks in order to optimize results.
Calculus – This model deals with application of microeconomics in making business decisions. It
applies principles of differential and integral calculus
Game theory – This model deals with strategies to be used by players in a competitive
environment
Queuing theory – This model deals with service delivery so as to balance between service costs
and waiting costs where there are queues
Simulation – This is the representation of the behavior or characteristics of a real system through
the use of a hypothetical system which replicates the real system.

Identify any four mathematical models and explain how each can be applied in
business
LECTURE 2
TYPES OF MATRICES AND COMPUTATION OF THE DETERMINANT
Definition
A matrix is a rectangular array of numbers of the order (size) m x n, where m is the number of
rows and n is the no of columns. E.g.
2 4 3 4 1 12 5 8 6
A= 1 6 B= 17 2 4 -6 C = 3 -4 1
3 10 2 6 3

Order 3x2 Order 2x4 Order 3x3

The numbers in the matrix can be positive or negative, whole numbers or decimals and are
referred to as elements. A matrix is denoted by a capital letter.

18
Types of Matrices
Column matrix – A matrix with a single column. The order of the matrix is mx1

8 7
e.g. C = 12 of order 3x1 or D = 4 of order 2x1
9
Row matrix – A matrix with a single row. The matrix is of order 1xn

E = 7 9 16 order 1x3 OR F = 4 71 3 5 of order 1x5

Square matrix - A matrix where the number of rows is equal to the number of columns i.e. m =
n 5 8 6
e.g. G= 3 4 1 of order 3x3
2 6 3
Null or zero matrix – A matrix in which all the elements are zero
0 0 0 0
H= 0 0 0 0 order 2x4

Diagonal matrix – A square matrix in which all the elements except at least one in the leading
(principal) diagonal are zeros

0 0 0
J = 0 2 0
0 0 6

Transpose of a matrix – Matrix obtained by interchanging rows and columns. E.g., for the square
matrix G, the transpose GT is given as
5 3 2
GT = 8 4 6

19
 6 1 3

Identity matrix [I] – A diagonal matrix in which all the elements in the leading diagonal are equal
to 1 while all the other elements are zeros

1 0 0
I= 0 1 0
0 0 1

Symmetric matrix - A matrix in which the elements are ‘reflected’ along the leading diagonal. The
transpose of the matrix is the same as the original matrix. Example:
5 3 9
S= 3 8 4
9 4 7

Sub matrix - A matrix obtained by deleting any number of rows and columns from a given matrix
for example, with respect to matrix S above, sub matrices may be:

5 9
S= 3 4 or 3 8 or 5 9
9 7 9 4 9 7
Eliminating column 2; eliminating R1 & K3; eliminating R2 and K2
Operating With Matrices
Matrices A and B are said to be equal if and only if:
i) they are of the same order
ii) the corresponding elements (ai,j)are equal to (bi,j) for all values of i and j

Are the set of matrices, A & B below equal?

20
i) A = 1 0 and B = 1 0 0
0 1 0 1 0
0 0 1

ii) A = 6 -9 and B = 6 -9
15 4 15 4

Addition and subtraction rules

Matrices can be added or subtracted only if they are of the same order. Corresponding elements
are added / subtracted
Example:
10 6 3 -7 6 6
A = -7 5 B = 9 10 C = 4 -5
12 15 6 8 8 4

19 5 4 0
Then A + B + C = 6 10 and A – C = -11 10
26 27 4 11

Multiplication of matrices
A matrix of any dimension can be pre-multiplied by a scalar. In this case, each element in the
matrix is multiplied by the constant (scaler). If matric B above is pre-multiplied by a scaler of 4,
then 4B will be given by:
3 -7 12 -28 A scaler can be a fraction e.g. 1/5 or
4 9 10 = 36 40 3/8
6 8 24 32
Matrix multiplication

21
2 matrices are said to be compatible or conformable if and only if the number of columns in the
pre-multiplying matrix are equal to the number of columns in the other matrix, If for example
Matrix A is of the order 3x2 and matrix B is of the order 2x4, then:

A B In multiplying matrices, the row of

3 x 2 2 x 4 the pre-multiplying matrix is
multiplied by each of columns of the
Conformable other matrix and the sum obtained
Order of product matrix AB
Note
When dealing with matrices, AB ≠ BA, even when the matrices are square matrices and hence
conformable
Example

A= -7 5 B= 9 10 5
6 8 6 AB = -33-30 -5
Order 1 x 2 2 x 3 1 x 3

Note BA is not possible!

22
 LECTURE 3
MATRIX INVERSION AND APPLICATION OF MATRIX ALGEBRA
The inverse of a matrix A is denoted as A-1. It is a matrix such that A. A-1 = A-1.A = I (Identity
matrix). An inverse is only defined for a square matrix. The inverse of a matrix is required to
perform ‘division’, which is not possible conventionally
The procedure in getting the inverse of a matrix A is:
i) Determine the determinant of the matrix denoted as A . If the determinant is equal to
zero, then there is no defined inverse, and such a matrix is referred to as a singular
matrix
ii) Get the matrix of minors of matrix A
iii) Get the matrix of cofactors, where the cofactors for each element (ai,j) is given as is Cij
= (-1)i+j(ai,j)
iv) Transpose the matrix of cofactors to get the Adjoint matrix of matrix A, AT
v) A-1 = AT
A

Note:
For a 2x2 matrix A = a b
c d , the determinant is computed as a.d – b.c and the inverse is
computed as:
d -b
A-1 -c a

Example
Matrix B = 4 6

23
 2 8 , B = 32 – 12 = 20

Then B-1 = 1 8 -6 0.4 -0.3

20 -2 4 . = -0.1 0.2 Test that B B-1 = B-1B = I

Inverse of a 3x3 matrix

4 2 6
Assume we have a matrix, E = 2 5 0
0 3 9 , the stages in getting the inverse are as follows:
E

 Compute the determinant of the matrix: There are alternative methods to this below:
(0 + 0 + 36) = 36 (Sum of the product of ‘other’ diagonal )
4 2 6 4 2
= 2 5 0 2 5
0 3 9 0 3 Sum of the product of 3 leading diagonals
(180 + 0 + 36) = 216
The determinant is the difference between the sums of the ‘leading’ diagonal and the
Sum of the product of the ‘other diagonal: i.e. the determinant = 216 – 30 = 180
 Matrix of minors for each element Eij is given as the determinant of the 2x2 matrix resulting
after crossing out the ith row and the jth column. For element E1,1 [4], this is given as 9*5 – 0*3
= 45. The entire matrix of minors is given as:

45 18 6
0 36 12
-30 -12 16
 Matrix of cofactors – Each element in the matrix of minors is multiplied by the factor (-
1)i+j

24
 45 -18 6
0 36 -12
-30 12 16

 Transpose the matrix of cofactors:

45 0 -30
. -18 36 12
6 -12 16


45 0 -30 0.25 0 0.167
. E-1 = 1 -18 36 12 = -0.1 0.2 0.067
180 6 -12 16 0.033 -0.067 0.089

Test that EE-1 = E-1 E = I.

Note:
The determinant can be computed by summing up the product of the elements in a given row (or
column) in the original matrix and the corresponding values in the similar row (or column) in the
matrix of cofactors – illustrated below using only rows 1 and 2 elements:
Row 1: 4 *45 + 2*-18 + 6*6 = 180
Row 2: 2*0 + 5* 36 + 0*-12 = 180

Application of matrix Algebra –

a) Solution to linear simultaneous equations
Two alternative matrix approaches may be used in solving a set of simultaneous equations: the
inverse method and crammer’s rule
Irrespective of whether the inverse method or the crammer’s rule is used, we need to restate the
equations to be solved in matrix format

25
The inverse method
When this method is used, we simply apply the relationship: AA-1 = I in solving for the unknown
variables. The procedure is:
 Put the original equation in matrix form:
A[x] = B, where A is the matrix and [x] represents the unknown variables in form of a
column matrix
 Pre-multiply both sides by the inverse A-1. This translates the equation into:
I[x] = A-1B, which defines the values of each of the variables when solved
Example
Solve the equations below using the inverse method:
5 + 8 = 139
3 + 66 = 193
Solution
In matrix format
5 8 x 139
3 6 y = 93

The inverse of the matrix is given by:

1 6 -8 and therefore 1 6 -8 5 8 x 1 6 -8 139
6 -3 5 6 -3 5 3 6 y = 6 -3 5 93

Therefore 1 0 x 15 OR x 15
0 1 y = 8 y = 8

Therefore x=15 and y =8

Crammer’s rule
This involves working only with determinants. The steps are:
(i) After putting the equations to be solved in matrix format, determine the determinant of
the matrix A, pre-multiplying the column matrix of the variables.

26
 (ii) Make matrices, in each case replacing the column of each unknown variable with the
RHS, and compute the respective determinants
(iii) The solution to each variable is the ratio between the determinant computed when each
variable is replaced and the original determinant

Applying this to the example above:

5 8 139 8 5 139
3 6 = 30 – 24 = 6: 93 6 = 139*6 -93*8 = 90; 3 93 = 465 - 417 = 48

x = 90 /6 = 15: y = 48 /6 = 8

b) Input-Output Analysis
This model was developed by Prof Leontief in 1951.It analyses production within an economy and
the interdependencies between the sectors in an economy. The model recognizes that the output
of one sector may be used as input in another sector as well as input in the sector itself. There are
two models in input-output analysis
 A closed model in which the entire production is used entirely within the production
process by the industry itself together with other productive sectors within the economy
 The open model, which also recognizes that part of the output, will be used for
consumption, thus representing the final demand.

Terminologies
Intermediate demand – Part of the total output that is used within the productive sectors of the
economy
Final demand – Part of the production [total output] that is consumed or used outside the
productive sectors of the economy
Input-output matrix – This is also referred to as the matrix of technical coefficients. It shows
the interdependencies among the productive sectors. This matrix gives the proportion of output
from a given sector required as input in the production process in a given sector.

27
Input-output analysis may therefore be used to determine the amount of production required in
each industry in order to meet the intermediate demand and the final demand.

Application of the model

Given some level of final demand, input-output analysis can be used to determine the total output
that will be required in each sector to meet both intermediate and final demand. We illustrate
using a two-sector economy with Agriculture and Manufacturing as the sectors. The open model
may be illustrated as below:
From To
A M Demand Total
A 800 750 450 2 000
M 900 600 1 500 3 000

Interpreting row for Agricultural sector:

Total output is 2 000, out of which agriculture itself uses 800, manufacturing uses 750 and the final
consumers take the balance of 450. Note that the total output [X] equals the sum of intermediate
demand, [Id] of 1 550 and the final demand [D} of 450 – i.e., X = Id + D. A similar analysis may
be done for the manufacturing sector. In our illustration above, the input-output matrix, A, may
be shown as below
X Y X Y
A = X 800/2 000 750/3 000 A 0.4 0.25
Y 900/2 000 600/3 000 = M 0.45 0.2 ,

Notice that intermediate demand, ID = AX - i.e., 0.4 0.25 2 000 1 550

0.45 0.2 3 000 = 1 500
The input-output matrix may be interpreted as follows:
Agriculture requires an input 0.4 of its output and 0.25 of the output from manufacturing to
produce 1 unit of output. At the same time, manufacturing requires an input of 0.45 from
agriculture and 0.2 from itself to produce one unit of output.
The matrix of technical coeffients is assumed to be stable

28
Leontief matrix
This is the matrix [I – A], where I is the identity matrix. It comes as a result of the following
relationship: X = AX + D, i.e., Total output [X] = intermediate demand [AX] + final demand
[D]. This implies that:
X – AX = D or [I – A] X = D.
Therefore, pre-multiplying both sides by the leontief inverse: [I] X = [I – A] -1 D

In our example, if final demand changes to 800 for Agriculture and 1 800 for manufacturing, then
we can determine the total output and the associated intermediate demand in each sector:
Total output will be:

-1
A 0.6 -0.25 800 1 0.8 0.25 800 2 965.99
M = -0.45 0.8 1 800 = 0.3675 0.45 0.6 1 800 = 3 918.37

Output required from agriculture is 2 965.99 and from manufacturing is 3 918.37.

Intermediate demand will be:

2 965.99 800 2 165.99 0.4 0.25 2 965.99 2 165.99

3 918.37 - 1 800 = 2 118.37 OR 0.45 0.2 3 918.37 = 2 118.37
Three sector economy – illustration
An economy consists of 3 sectors P, Q and R. Given below are the are the inputs and outputs
from the 3 sectors. The amounts are in Billions of shillings

Sector PQ R External Demand Total Output

P 18 30 45 15 108
Q 27 30 60 3 120
R 54 40 60 26 180
Required:
a) Form a technical coefficients matrix, A

29
b) If the external demand was to change to Sh10 billion, Sh 2 billion and Sh. 20 billion from the
current levels respectively for A, B and C, calculate the total output of
sectors P, Q and R. You are given that:

1 336 216 234

[I – A] -1 = 109 288 372 294
396 348 486 , where I is the identity matrix

c) What will the intermediate demand for the sectors P, Q and R be in this case

Solution
a) The matrix of technical coefficients is given as:
18/108 30/120 45/180
A = 27/108 30/120 60/180
54/108 40/120 60/180

b) The total output matrix is given by: X = [I – A] -1D

P 1 336 216 234 10 77.72

Q = 109 288 372 294 2 = 87.19
R 396 348 486 20 131.89

c) The intermediate demand = X – D or AX

77.72 10 67.72 .167 .25 .25 77.72 67.75
87.19 - 2 = 85.19 OR .25 .25 .333 87.19 = 85.14
131.89 20 111.89 .5 .333 .333 131.89 111.81

30
 Self-Assessment Questions

Question1
In an economy there are 2 industries, A & B in an economy. Given below are the supply &
demand in millions of shillings.
User
Producer A B Final demand Output
A 15 10 10 35
B 20 30 15 65

Required
a) the matrix of technical coefficients
b) Determine the following:
i)the total output of each industry if the final demand changes to Sh 12 million for A and Sh 18
million for B
ii) the intermediate demand associated with the output in (i) above

Question 2
You are provided with the following input-output matrix in a 3-sector economy represented as X,
Y & Z respectively

0.2 0.3 0.2

A= 0.4 0.1 0.2
0.1 0.3 0.2

31
Required
a) Interpret each of the elements in the first column of the matrix
b) Given that the final demand vector for the 3 respective sectors in millions of shillings is
100; 50 & 60 for sectors X, Y and Z respectively
i) Determine the total output of each sector
ii) The corresponding intermediate demand

c) In a 2 sector economy made up of A and B, you are provided with the following input-output
matrix for the respective sectors

0.25 0.3
0.5 0.4
The final demand for the sectors amount to Sh 150 million and Sh 90 million respectively for A
and B
Required
(i) Interpret the ‘0.25’ and the ‘0.4’ in the input-output matrix
(ii) Compute the total output for each of the sectors [6 marks]

Question 3
An economy consists of 3 sectors P, Q and R. Given below are the are the inputs and outputs
from the 3 sectors. The amounts are in Billions of shillings

Sector PQ R External Demand Total Output

P 18 30 45 15 108
Q 27 30 60 3 120
R 54 40 60 26 180
Required:
a) Form a technical coefficients matrix, A

32
b) If the external demand was to change to Sh10 billion, Sh 2 billion and Sh. 20 billion from the
current levels respectively for A, B and C, calculate the total output of
sectors P, Q and R. You are given that:

1 336 216 234

[I – A] -1 = 109 288 372 294
396 348 486 , where I is the identity matrix

c) What will the intermediate demand for the sectors P, Q and R be in this case

Question 4
An economy consists of three industries A, B and C with each industry producing a
single product. The table below describes the interaction of the use of A, B and C
production over some fixed period of time.
TO A B C consumer demand Total output

FROMA 90 80 35 95 300
B 30 160 70 140 400
C 60 120 140 30 350
Suppose final demand for A, B and C changes to 100, 150 and 25 respectively. Determine the new
total output by various industries so as to satisfy the new final demand.(14 marks)

Question 5
a) A restaurant sells beverages as follows;
2 cups of tea, 3 glasses of red wine and a bottle of soda for shs. 490.
3 cups of tea, 4 glasses of red wine and two bottles of soda for shs. 700
A cup of tea, 2 glasses of red wine and a bottle of soda for shs. 330
Required;
i)Formulate the above information into three simultaneous equations
ii) Determine the price of a cup of tea, a glass of wine and a bottle of soda.

33
iii)Determine the price of 4 cups of tea and two glasses of wine from the restaurant. ( 10 marks)

b) Solve the following system of simultaneous equation using a matrix-based method of your
choice:
6X + 4Y + Z = 162
2X + 6Y + 5Z = 194
X + 3Y + 4Z = 112 [8 marks]

QUESTION 6
a) Briefly explain what you understand by each of the following terms:
i)Identity matrix
ii) Singular matrix
iii)Inverse matrix
iv) Redundant constraint
v) Shadow price
vi) Dummy activity [12 marks]

b) An economy is made up of 2 sectors, P and Q. The total output, intermediate as well as final
demand for each of the sectors is given below. All the values are in millions of shillings
Producing Total Intermediate Demand Final
Sector Output P Q Demand
P 2 500 750 800 950
Q 4 000 1 250 1 600 1 150

Required:
i)The matrix of technical coefficients, A [4 marks]
ii) Determine the total output for each of the sectors and the related intermediate demand if the
final demand for sectors P and Q changes to Shs 800 million and Shs 900 million respectively
[6 marks]

34
Question 7
a) Outline any three models that are frequently used by managers and explain how each of the
models may be of use in solving business problems [6 marks]

b) Solve the following set of equations using a matrix method of your choice:

8X – 3Y = 245
6X + 4Y = 90 [6 marks]
Question 8
a) Explain each of the following terms:
i)Intermediate demand
ii) Infeasibility
iii)Primary inputs
iv) Singular matrix
v) Dummy activity [10 marks]

b) Solve the following equations using crammers rule:

4X + 6Y + 2Z = 184
2X + 5Y + 2Z = 132
5X + 2Y - Z = 108 [6 marks]
c) An economy consists of 3 interdependent sectors: Agriculture, Mining and Manufacturing. The
flow of inputs between the industries is represented in the table below:
INPUTS (SHS 000) FINAL
DEMAND
OUTPUT (SHS 000] AgricultureMining Manufacturing
Agriculture20 000 32 500 37 500 10 000
Mining 30 000 65 000 37 500 30 000
Manufacturing 40 000 32 500 12 500 40 000

35
Required:
i)The technical coefficients matrix, a [3 marks]
ii) The related Leontief matrix [3 marks]
iii)The change in output for each of the three sectors if the final demand for agriculture increases
by Shs 1 million (i.e., to Shs 11 million), that of mining remains unchanged and the final demand
for manufacturing falls by Shs 500 000
You are given that Leontief inverse is as below:
0.480.240.24
I–A -1 = 1/216 0.390.6 0.33
0.3 0.240.42 [8 marks]

b) An industrial system has the industries with the following input for industry A to produce one
pounds worth output it requires £0.32 of its own products and 0.40 of industry B’s product.
Similarly, industry B requires £0.21 of its own product and £0.40 of industry A’s product to
produce one pound worth product.

If the final demand is given by

E = 50,000
30,000

Determine
i)The technical coefficient matrix for this system.
ii) Using the matrix in (i) above determine the total output to be produced so as to satisfy the
final demand. (10 marks).

36
 LECTURE 4
NETWORK ANALYSIS
Introduction
 Network analysis is applied in planning and controlling large projects which are made up of
interrelated jobs.
 The technique is useful both in ensuring that the ‘jobs’ are done in time and in helping
management to allocate scarce resources.

Concepts / terminologies in network analysis

Project – combination of all jobs to be performed for the entire task to be completed. It is
represented in form of a network diagram. Activities should be sequentially / logically arranged in
a network diagram, taking into account their precedence.
Network diagram – a graphical presentation of a project showing the various activities and events
Activity – an individual ‘job’ within the project which takes time, uses money and other
resources. Each activity is represented by an arrow in the network diagram

37
Event – represents a reference point in time where an activity starts / ends. Each event is
represented by a node (circle) in the network diagram. Events in the diagram are interlinked by
activities. Normally, events are numbered sequentially
Dummy activity – an activity appearing in the network diagram so as to maintain logic in the
sequencing of activities and in order not to violate rules of drawing network diagrams. Dummy
activities use no time and no resources. Dummy activities are represented by a dotted arrow. They
become necessary as the network diagram is drawn
Critical path – the set or combination of activities on the network diagram which has the longest
duration. This path gives the shortest completion period. All activities on the critical path are
referred to as critical activities
Earliest start time [EST] – the earliest point in time at which an activity can commence, after
taking into account the preceding activities
Earliest completion time [ECT] – the earliest point in time that an activity can be completed, after
taking into consideration the duration of al the preceding activities
Latest starting time [LST[ - latest point in time at which an activity can be started without
increasing the overall project duration
Latest completion time [LST] – latest point in time that an activity can be completed without
increasing the overall project duration
Float or slack – the time (duration) with which an activity can be delayed or its duration increased
without delaying the completion of the project. Critical activities do not have any float
Crashing / accelerating a project – reducing the duration of a project by decreasing the time it
takes to complete individual activities. This requires that more resources be allocated to the project
Critical path analysis / method [CPA / CPM] - a network analysis in which the duration of each
activity can be established with reasonable certainty
Programme evaluation and review technique [PERT] – a network technique in which the
duration of activities therein cannot be determined with certainty.
Beta distribution – a probability distribution used in PERT to determine the expected duration of
each activity as well as the standard deviation

Rules in drawing network diagrams

38
 The project has one start point [start event] and one end point [finish event] – with various
activities / events sequentially interlinked in between - no dangling activities / events
 The network diagram is drawn from left to right
 Each activity has a preceding [tail] event and a succeeding [head] event
 Looping activities are not allowed
 Two or more activities cannot start in a given tail event and have the same head event
 The length of an arrow does not in any way indicate the duration of the activity or the
resources used by the activity
NOTE
1. Activities in a network may be ‘described’, named alphabetically, numerically or with
reference to the tail-head events
2. Dummy become necessary to maintain the ‘flow’ and logic of the network
Critical path analysis [CPA] method
 After the network diagram has been constructed, it becomes necessary to determine how long
a project will take to be completed
 An analysis of the time required to complete each activity is undertaken so as to determine the
time at which each succeeding activity can commence. The duration required for each activity
is presumed to be known with certainty [deterministic]
 The critical path is the longest path in the network and is made up of activities that can not be
delayed.
 To determine the critical path, the following steps are necessary:
o Draw the network diagram
o For each event, determine the earliest start time [EST] – earliest time that the next activity
can start after taking into consideration all the preceding activities – done through the
‘forward pass’. The EST of the finish event represents the duration of the project
o Compute the latest start time [LST] for the succeeding activity – Latest time that an activity
should start so as not to delay the project beyond the critical time. Computations are done
using the ‘backward pass’
o The critical path is made up of those activities for which EST=LST. All the activities on
this path should be completed in time if the project is not to be delayed

39
 Note:
If management wants to reduce the duration of the project, then ways have to be looked into
of reducing the duration of at least one of the critical activities

Example1
Activity Preceding Duration [Weeks]
A - 3
B A 5
C A 4
D A 4
E B 2
F B 4
G E 6
H F 3
J C,F 8
K D 7
L G,H 4
M J,K,L 5

E[2]
G[6]

B[5] F[4] H[3]

A[3]
L[4]
C[4] M[5]
J[8]
D[4]

40
 K[7]

Example 2
Activity Predecessor Duration [weeks]
A - 5
B - 4
C A 1
D B 1
E B 5
F B 5
G C,D 4
H F 3

Required
a) draw a network diagram for the project
b) determine the critical activities and the project duration

Example 3
Activity Predecessor Duration [Months]
A - 3
B - 4
C A 3
D A,B 6
E A 5
F A,B 3
G C,E 3
H D,E 2
J F 4
K G,H,J 3

41
Required
a) Draw a network diagram for the project
b) Determine the project duration and the critical activities

Concept of float / slack

 Refers to slack or free time by which a non-critical activity can be delayed without affecting
the project duration – i.e., without the activity itself becoming critical
 All activities on the critical path have zero or no float
 There are 3 categories [types]of float:
o Total float – duration by which an activity could be delayed without affecting total
project time [TFi = LSTj – ESTi - Di
o Free float - duration by which an activity could be delayed without affecting the start
time of the subsequent activities and total project time. Preceding activities are ignored
in determining free float [FF = ESTj - LSTi - Di
o Independent float - duration by which an activity could be delayed without affecting
the total project time, time for preceding and that of subsequent activities [IFi = ESTj –
ESTi - Di
Note:
In the above computations, the notation is as follows:
LSTi = latest start time of activity i for which we are determining the float
ESTi = earliest start time of activity i
LSTj and ESTj are similar but relating to the subsequent activity j.

Project Evaluation and Review Technique [PERT]

 Unlike CPA which presupposes that the expected duration of each activity can be predicted in
advance, PERT applies 3 time estimates for each activity, based on which the expected
duration of each activity is computed using the beta probability distribution

42
 These durations are:
o The optimistic [shortest] duration [represented as a]
o The most likely time [represented as m]
o The pessimistic [longest] duration [represented as b]
 The expected duration an activity i, is given as:
Ei = [a + 4m +b] /6
 Once the expected duration for each activity has been determined, the analysis continues just
as in CPA. An advantage with PERT is that since thee are multiple time estimates, further
analysis can be performed to determine the probability of completing the project within a
given duration.
 This is done as follows:
o For each of the critical activities, calculate the standard deviation, using the beta
probability distribution. For activity i, his is given as:
Σi = [b – a] / 6
o Calculate the standard deviation of the critical path. This is given as:
Σe = [Σ(b – a)2/6]1/2, using the normal distribution, calculate the Z-score value for the
scheduled completion date [S} from the expected duration per the critical path [Pe].
i.e., Z = [S – Pe] / Σe
o The normal distribution can be used to determine the confidence interval for the
completion time of the project or determine the probability of completing the
project in a given time

Example – illustration to be done in class

Activity Predecessor Expected duration
am b
A - 579
B - 1 1.5 5
C - 2 34
D A 2.5 3.5 7.5
E A 1 23

43
F B 0.5 4.5 5.5
G B,C 8 17 20
H E,F 7 14 15
J D 5 8 11

Required
c) Draw a network for the project
d) Compute the expected duration for each activity, determine the critical path and the expected
duration of the project
e) Compute the probability of completing the project within
i. 18 weeks
ii. 25 weeks

Project crashing / acceleration

 Once the critical path has been identified, it may be found out that the project duration is
more than required
 The overall project duration may be reduced by reducing or ‘crashing’ the time required on
one or more of the critical activities
 To crash any activity that is not on the critical path would be a waste of resources since it will
not result in reduction of overall project duration
 To reduce the duration of an activity, it requires that extra resources e.g., labour, material,
equipment e.t.c., be allocated and hence increases the costs of completing the project in less
time
Note:
 Initially, all the non-critical activities are irrelevant and should not be considered for crashing.
However, after crashing commences, some activities which were initially not critical may
become critical
 When evaluating the critical activities that are crashable for reducing their duration, we take
into account the ‘slope’ – i.e., it is the activity with the least crash cost per unit of time that
should be crashed first

44
 Sometimes, it may be necessary to crash some activities on different critical paths
simultaneously

Example
The management accountant of Mapeni Ltd. is planning the annual closedown of one of the
departments for maintenance purposes. The shutdown process has 6 phases. You are provided
with the following information about the phases.
Activity / phase Preceding Duration
Normal Crash Normal Crash
A - 8 4 8 000 20 000
B - 6 5 6 000 7 000
C B 2 1 10 000 14 000
D - 4 2 4 000 7 000
E A 3 2 1 000 2 600
F C,E 2 2 16 000 16 000
G B,D 5 3 12 000 16 800

Required
a) A network for the project, the critical activities and the normal cost of the project
b) Determine the duration beyond which the project duration should not be crashed and the total
costs for that duration
c) The management accountant contends that although each week saved in the project would
increase costs, there would be an additional contribution of Sh. 4 800 per week if the
department is not closed. Determine the optimal duration of the closure for the department

Self-Assessment Questions

Question 1

45
a) Kilifi Ltd has won the tender on a major job. Nine activities have been identified by the
engineers as illustrated below. The activity duration is in weeks
Activity Predecessor Duration [weeks]Cost (Shs 000)
A - 5 450
B A 2 270
C A 4 750
D A 7 240
E B 6 150
F D 8 320
G C,E 1 690
H F,G 3 475
Required
i)Draw a network diagram for the project [5 marks]
ii) Determine the expected project duration, the critical activities and project cost[5 marks]
iii)Compute the total float for each of the non-critical activities [4 makes]

b) Explain each of the following terms

i)Dummy activity
ii) Earliest start time
iii)Beta probability distribution [6 marks]

Question 2
Mr. Kalidi is a manager of Construction Company. The company is bidding on contract to repair
the road network in a small town. It has identified the following activities.
(Time weeks)
Activity Optimistic Most likely pessimistic

1-2 4 6 10
1-3 3 5 9
2-4 7 12 20

46
 2-5 3 5 8
3-4 6 11 15
4-5 4 6 11
4-6 3 9 14
5-6 2 4 8
6-7 3 5 9
Mr. Kalidi has agreed with the client that the project should be completed in the shortest
time possible.
Required:
(i) Draw a network for the project (4 marks)
(ii)Determine the critical path and the expected project completion time. (4 marks)
(iii) Determine the standard deviation of the project completion time. (3 marks)
(iv) Explain any five Limitations of games in competition (10 marks)

Question 3
The following activities relate to a project to be undertaken by a certain organization.
Activity Preceeding activity Duration in weeks
A - 6
B A 9
C A 15
D A 24
E B 18
F C 3
G C 6
H C,D 9
I E,F 21
J G,H 12
K I,J 15
Required:
i) Draw a network for the project. (5marks)

47
ii) Determine the critical path using the diagram (5marks)
iii) Calculate any floats for the activities. (5marks)

Question 4
A project consists of nine jobs (A, B, C, D, E, F, G, H and I with the following preceding
activities and time estimates in days.
Job Preceding Optimistic most likely pessimistic
activity time time time
A _ 4 7 16
B _ 3 6 15
C A, B 7 13 30
D A,B 3 6 9
E B 6 12 18
FD,E 4 7 16
G C,F 4 10 28
H D, E 2 5 8
I G,H 3 6 9
Required:
(i) Determine the expected duration and variance for each job
(ii)Draw a network diagram for the project.
(iii) Identify the critical path
(iv) Calculate the variance and standard deviation of project length
(v) Determine the effect on both project time and the critical path if the expected duration of job D
were to be 10 days

Question 5
Given below are the activities required in a project, the related duration and costs.
Activity Duration (Weeks) Optimistic Most likely Pessimistic
1-2 3 5 7
1-3 2 5 14

48
1-4 5 7 21
2-6 5 8 11
3-4 0 0 0
3-6 2 3 4
4-5 4 6 14
5-6 0 0 0
5-7 3 4 5
6-7 1 3 5

Required:
a) Determine the expected duration in weeks and the variance for each activity [8 marks]
b) Draw a network diagram for the project and use it to determine the critical activities and the
expected duration of the project [10 marks]
c) Explain the nature of activities ‘3-4’ and ‘5-6’ [2 marks]
d) Compute the standard deviation of the expected project duration [6 marks]
e) Outline any four rules used in drawing a network project [4 marks]

NOTE: Di = [a+4m+b]/6; Σ2 = [(b-a)/6]2 and σc = [∑(b-a)/6]2]0.5 – using only critical activities

QUESTION 6
a) Explain the following terms as used in network analysis:
i.Activity and event
ii. Looping and dangling activities [6 marks]

b) A project is made up of 11 activities, A – K, with the sequencing and time estimates of the
activities as shown below:
Activity Estimated Duration (weeks)
Optimistic Most likelyPessimistic
1-2 4 5 6
2-3 3 5 19

49
2-4 1 3 5
3-4 0 0 0
3-5 1 3 11
5-6 2 3 4
4-5 6 9 18
4-6 2 5 8
Required:
i)Determine the expected duration and variance for each of the activities[7 marks]
ii) Draw a network diagram for the problem [5 marks]
iii)Determine the expected project duration and the critical activities [5 marks]
iv) As the project manager, determine the project duration you would give your
client to be 95% confident of delivering on time [7 marks]

Question 7
a) Briefly explain the following terms:
i)Decision variables
ii) Critical activity
iii)Redundant constraint
iv) Queuing model [8 marks]
b) A project is made up of eight activities as detailed below:
Activity predecessor time estimates in weeks
A - 5
B - 12
C A 14
D B 5
E B 9
F C, D 17
G E, F 5
H E 13
Required:

50
a) Draw a network diagram for the project [6 marks]
b) Determine the critical activities and the expected project duration [7 marks]
c) Determine the float for activities ‘C’ and ‘H’ [4 marks]

LECTURE 5
LINEAR PROGRAMMIMG

Chapter Objectives:

At the end of this topic, students should be able to:

o Explain the components of a linear programming model

o Formulate linear programming problems

o Determine the solution to linear programming problems using the graphical method

51
 o Do sensitivity analysis on linear programming optimal solution

INTRODUCTION
 In short-run management decisions, products can be ranked according to contribution per unit
of the scarce resource [if there is a limiting factor].
 When there is more than one scarce resource, this becomes ineffective as rankings will differ.
 As a result, we resort to the use of linear programming. It’s a mathematical technique used by
management in order to assist in the allocation of scarce resources when there are more than
one limiting factor
 Linear programming can be applied by management in evaluating:
 Blending
 Production planning
 Oil refinery management
 Distribution
 Financial and economic planning
 Manpower planning
 Blast furnace burdening
 Farm planning
 Capital rationing (financial management)
Requirements for linear Programming
 Variables should be capable of being expressed mathematically.
 There should be a linear relationship between the variables i.e. if it requires 3 hours to produce
1 unit of x then 300 units require 900 hours
 Different alternative courses of action should be feasible
 Negative solutions are unacceptable.

Formulation of Linear Programming Problems

1. Identification / specification of decision variables
As said earlier, these are the variables in respect of which a management decision is required;
e.g., the products in a product mix decision

52
2. Setting up an objective function
This is expressed mathematically using decision variables with a basic objective function
3. Constraints
Each resource usage is expressed mathematically and related to two decision variables.

Solution to liner Programming Problems

Two kinds of approaches can be used:
 Graphical method
 Evaluation of corner points
 Use of iso-contribution / cost line
 Simplex method – this is beyond the scope of this unit

Graphical method
 This method can be used if there are two decision variables. The decision variables are plotted
on each axis and the constraints on resource usage are drawn as straight lines.
 The feasible region is identified as that which does not contravene any of the constraints (ABC
in diagram)
 The optimum solution is determined by either;
 Evaluation, the value of the objective fn for the decision variable values at each of the
corners of the feasible region and identifying the highest value as the optimum solution or,
 Determine the corner within the feasible region which is the furthest from the origin on
the basis of the ISO-contribution (the objective fn line) as the optimal solution.

Illustration
XYZ Ltd produces two products Alpha and Beta using the same raw material, labour and
machinery. Details of the products are given below:
Alpha Beta
Price per unit 24 000 16 000
Variable costs per unit 21 000 12 000
Contribution per unit 3,000 4,000

53
 Resources/Unit Maximum/week
Machine hours 4 2 1000 hours
Labour hours 2 6 1500 hours

Material input [Q] 1 1 300 kg

The maximum weekly demand of product Beta is 200 units and weekly fixed costs $480 000.

Required
(a) Formulate the above as a LP problem
(b) Determine the optimal level of output of each product per week.
(c) Evaluate the resource constraints in order to establish,
(i) The limiting factors
(ii) The non-scarce resources and the unused amounts of such resources.

Solution
Let a equal the # of units of product alpha produced and sold
Let b equal the # of units of product beta produced and sold

a) Objective fn
Maximise contribution (z) = 3000a + 4000b
A B A B
Machine hour = 4a + 2b < 1000 [2500] [0 500]
Labour hour = 2a + 6b < 1500 [7500] [0 250]
Material Q = a+b< 300 [3000] [0 300]
DD for b = b < 200 [2000] [0 200]
A, b Σ 0

b) Corner point method

 Evaluate the objective function value for each of the corner points in the feasible region

54
 For maximisation problems, corner point giving the highest value is the optimal solution. For
minimisation problems, the corner point with the least value gives the optimal solution

b) Alternatively – use iso-contribution / cost line

 This method super-imposes the objective function [iso-cost or iso-contribution] into the graph
 For maximisation LP problems, if you try to push it inwards, the point furthest from the
origin within the feasible region will give the optimal – for minimisation the point within the
feasible region nearest to the origin using the super-imposed iso-cost line gives the optimal
solution

c) (i) Limiting factor

This is determined by the constraints that intersect at the optimal point - DD for Beta and material
B input
(ii)Others will be not fully used
Machine hours available 1 000
Used at optimal level
Alpha (100 x 4) 400
Beta (200 x 2) 400 800
Unused 200

Labour hours available 1 500

Used
Alpha (100 x 2) 200
Beta (200 x 6) 1 2001 400
Unused 100

Exercise

55
A company produces two products x and y with respective contribution/unit of $40 and $30.
There are 3 resource constraints i.e.
 Processing hours
 Skilled labour hours
 Material a, which is used in the production of x and y

The table below identifies the resource requirements in the production of x and y.

Resources x y Total available

Processing hour/unit 4 4 12 000
Skilled labour/unit 4 2 10 000
Material a 1 2 5 000

Total FC are 70 000

Illustration
XYZ Ltd produces two products Alpha and Beta using the same raw material, labour and
machinery. Details of the products are given below:
Alpha Beta
Price per unit [Sh] 24 000 16 000
Variable costs per unit [Sh] 21 000 12 000

Resources/Unit Maximum/week
Machine hours 4 2 1000 hours
Labour hours 2 6 1500 hours
Material input [Q] 1 1 300 kg
The maximum weekly demand of product B is 200 units and weekly fixed costs Sh480 000.

Required
(a) Formulate the above as a LP problem

56
(b)Determine the optimal level of output of each product per week.
(c) Evaluate the resource constraints in order to establish,
(i) The limiting factors and the respective shadow prices
(ii)The non-scarce resources and the unused amounts of such resources

Question 1
(a) A linear programming problem has three characteristics (parts) in its formulation.
State them (3 marks)

(b)A firm engaged in producing two models X1 and X2 performs only three operations – painting,
assembly and testing. The relevant data are as follows:

Unit sale price Hours required for each unit

Sh. Assembly Painting Testing
Model X1 50 1.0 0.2 0.0
Model X2 801.5 0.2 0.1

Total number of hours available each week are as follows:

• Assembly - 600
•Painting - 100
•Testing - 30
The firm wishes to determine its weekly product-mix so as to maximize revenue.
Required:
(i) Formulate the above Linear Programming (5 marks)
(ii) Solve using graphical method (7 marks)

Question 2

57
A company manufactures two different kinds of chemicals / products, A and B. There are three
departments / processes through which the chemicals have to pass. You are provided with the
following information.
Product Process 1 (kg) Process 2 (kg) Process 3 (Hrs) Profit/ unit
A [per unit] 0.4 0 0.6 Sh 400/unit
B [per unit] 0.5 0.2 0.3 Sh 300/unit
Max avail / week200kg 50kg 210kg

Required
a) Formulate as a linear programming problem
b) Determine the optimal weekly production mix
c) Which process(es) are a real constraint in the production process? What are the respective
shadow prices
d) The non-binding process constraint(s) and the excess of recourse

Question 3
Maxwell Ltd. produces two products, X and Y. The standard contribution per unit of each
product is computed as below:
Product x Product y
Selling price 110 88
Standard costs
Material (8kg @ $4) 32 (4kg @ 4) 16
Labour (6 hours @ $10) 60 (5 hrs @ 10) 50
v. overheads (4hours @ $1) 4 96 (6 hrs @ 1) 6 102
Contribution/unit 14 16

The maximum availability of scarce resources per month is given below:

Maximum available
Materials (kilograms) 3 440
Labour (hours) 2700

58
Machine (hours) 2 760

The maximum demand for product Y is 400

Required:
(a) Formulate the problem as an LP and determine the optimal solution using the graphical
method.
(b)The company is contemplating on producing a new product, Z. The selling price will be
$84/unit and the production of each unit would use resources as indicated below:
Material [kilograms] 6
Labour hours 5
Machine hours 3

Advise the management on whether to introduce product Z

Question 4
Maji Fabricators makes two types of water pumps, the manual model and the automatic model.
The company makes a profit of Shs 750 for each manual model pump and Shs 600 for each
automatic pump. A unit of each pump uses labour and machine hours as given below:
Type of pump Time required per unit (minutes)
Labour Machine
Manual 30 20
Automatic 15 30

You are also provided with the following information:

i)Total labour hours available in a week are 4,800. The total machine hours are also 4,800 in a
week
ii) The maximum demand for each in a week is 140 units
Required
a) Formulate this as a linear programming problem [7 marks]
b) Using graph paper, plot the constraints and shade off the unwanted region for each
[5 marks]

59
c) Determine the optimal weekly output of each product and the maximum profit for the period.
[5 marks]
d) Determine the binding constraints and compute the shadow price of each [6 marks]
e) Outline two requirements for a linear programming model to be applied [2 marks]

Question 5
a) The formulation of a linear programming problem has three distinct parts. Identify each of the
three parts clearly explaining the importance of each [6 marks]
b) Mtwapa Ltd makes 2 products Delta and Gamma. Given below are the details of each of the
products and the constraining resources per week
PRODUCT
Delta Gamma
Profit per unit (Shs) 750 500
Resource requirements per unit Maximum Available
Skilled Labour Hours 2 1.5 1,200
Material (Kilograms) 2 4 2,000

In addition you are provided with the following information:

i)The maximum demand of product Delta per week is 450 units
ii) Mtwapa Ltd has to produce at least 100 units of product Gamma per week to satisfy a
contractual agreement with the Department of Defense
Required:
i)Formulate this as a linear programming problem [7 marks]
ii) Using graph paper, plot the constraints and shade off the unwanted region for each [5 marks]
iii)Determine the optimal weekly output of each product and the maximum profit for the period
[5 marks]
iv) Determine the non-binding constraints and the unused amount of each such constraint
[2 marks]

60
 LECTURE 6
THE TRANSPORTATION MODEL
This model has been constructed to assist management in allocating resources, just as in linear
programming. The transportation problem deals with distributing a product from a number of
sources to a number of destinations. Typically, each source is supply limited, each destination has
a known demand and the shipping costs between sources and destinations are given. The objective
is to find the cheapest shipping schedule that satisfies demand without violating supply constraints.

Characteristics and assumptions of the transportation problem

The following are the characteristics and assumptions of the transportation problem: -
1. The supply. A limited quantity of one commodity such as cement, oil or oranges is available
at certain sources or origins such as factories, refineries or groves.
2. The demand. There is a demand for the commodity at several destinations such as
warehouses, distribution centers or stores.

61
 3. The quantities. The quantities of supply at each source and the demand or requirements at
each destinations are constants.
4. The shipping cost. The per unit costs of transporting the commodity from each source to each
destination are assumed to be constant
5. It is assumed that no shipments are allowed between sources or between destinations.
Allowing such transshipments would require special adjustments in the model.
6. All supply and demand quantities are given in whole numbers.
7. The problem is to determine how many units should be shipped from each source to
each destination so that all demands are satisfied at the minimum total shipping cost.

This is special case of linear programming which deals with distribution of goods from various
sources (production points) to various destinations (consumption or deman points). Each source
has a capacity which it can produce whereas a destination’s demand is known in advance. The
objective is to ship the goods from the various sources to the various destinations so as to
minimize the transportation costs. It can be used to help in the determination of a suitable site for
locating production. It is presumed that the transport cost/unit is known in advance and is
constant from one point to another. If it costs kshs x to transport a unit of goods from A to B,
then it would cost kshs 20x to transport 20 units from A to B.
The iterative process in solving transportation problem may be illustrated as below:

62
 TABULATE & ENSURE SUPPY =
BALANCE DEMAND

USE ANY OF 3 METHODS:

GET INITIAL SOLUTION  NWC RULE
TEST FOR OPTIMALITY  LEAST COST METHOD
USING EITHER:  VOGEL’S APPROX.
 STEPPING STONE OR
 MODIDIED
DISTRIBUTION IS THE DETERMINE OPTIMAL
SOLUTIO YES
TRANSPRTATION
N
SCHEDULE & COSTS
OPTIMAL

NO

GENERATE
IMPROVED
SOLUTION

Developing an initial solution

This may be developed using these different methods.
(i) Northwest corner rule.
(ii) The Vogels approximation technique.
(iii) The least cost method.
Consider the following example:
Alpha Ltd has 3 production depots, A with a capacity of 500 units; B with a capacity of 1 200 units
and C whose capacity is 2 300 units. The units are marketed through 3 destinations X Y and Z
with demand of 800, 1 000 and 3 100 units respectively. The production costs in the factory are the
same so that the transportation cost between the factories will determine which factory supplies to
which market area. These costs are given in the table below
Destinations

X Y Z Capacity

A [7] [6] [2] 500

sources B [3] [5] [9] 1200
C [1] [8] [4] 2300

63
 Requires 800 100 3100 4000

Methods of getting an initial allocation

 The Northwest corner rule method
The N.W. corner rule allocates production from N.W. corner of the table and exhausts each
source/destination before proceeding along the unexhausted row/column.
Destinations

X Y Z Capacity

A 500 500
sources B 300 100 800 1200
C 2300 2300
Requirements 800 100 3100 4000

The transportation costs then are,

Cell Cost
AX 3 500
BX 900
BY 500
BZ 7 200
CZ 9 200
Total cost 21 300

Note: To proceed to the next step one would have to take into account two issues:
o Consider the number of filled cells in the table. This should be equal to m+n-1, where m=
total number of sources and n= total no of destinations. This is referred to as a test for
degeneracy. When the number of filled cells is less than m+n-1, the problem is said to be
degenerate. We tackle how to deal with degeneracy later
o Compute the shadow prices for the unfilled cells. This helps in evaluating whether an
optimal solution has been attained, i.e., whether costs can be reduced further if some units
were allocated to the unfilled cells. Since the procedure for computing shadow prices is the

64
 same irrespective of the method used, we first consider other methods of giving an initial
solution and have a look at computation of shadow prices later
 Vogel’s Approximation Technique (VAT or VAM)
This technique gives a better initial solution than the northwest corner rule method. The
procedure is as follows:
(i) For each source/destination (read row/ column respectively), determine the 2 least
costs.
(ii) Complete the difference and place it at the end of the centre: row [source] / column
[destination]. A centre refers to a source/destination. This difference represents an
opportunity cost
(iii) Determine the largest opportunity cost. The first initial allocation is made to the
cell which has the least cost on that row/column.
(iv) After this allocation, close out the other cells of the centre [source or destination]
whose allocation has been exhausted.
(v) Recalculate the differences for centers [sources / destinations] not yet exhausted,
using only costs related to unfilled cells and continue as in (iii) above until all
allocations are made. We can then compute shadow prices to determine if an
optimal solution has been obtained.

Just like in the northwest corner rule, tests for degeneracy / optimality of the solution would have
to be made
With VAM our initial allocation would have been:
X Y Z Capacity
A X X 500 500
4
B 800 100 300 1200 2
C X X 2300 2300 3
Requirements 800 100 3100 4000
Opportunity /cost2 1 2
2 3 5

65
 Cell cost
AZ 1 000
BX 2 400
BY 500
BZ 2 700
CZ 9 200
15 800
As can be seen, the solution obtained gives less initial cost than the northwest corner rule.

 Least cost method.

Under this method the initial allocation is made by allocating the largest number of units possible
to the source/destination cell with the least cost. That order is maintained in allocating units to all
other cells.
In case 2 cells have the same cost/unit and an allocation is to be made in either, then select the cell
that allows the larger amount of units to be transported. Similarly, shadow prices can be compared
to determine if the solution is optional.
In our case, the allocation would have been.

X Y Z Capacity
A X X 500 500
B X 100 1100 1200
C 800 X 1500 2300
Requirements 800 100 3100

i.e. allocate to CX; then AZ; next CZ, BY and last BZ.
The cost would be:
Cell Cost
AZ 1 000
BY 500
BZ 9 900

66
 CX 800
CZ 6 000
18 200
It can be seen that this initial solution is somewhere between Vogel’s approximation method and
the northwest corner rule.

Procedure in solving:
The following procedure is then used to solve the problem.

Test for / coping with degeneracy

Degeneracy occurs if the filled cells are less than m+n-1. Testing for degeneracy is done after every
re-allocation. If we have a situation whereby the filled cells are less than m + n – I, then we must
locate an epsilon [quantity of zero (0)], to a cell located in such a way as to enable us to compute
the shadow prices. Once this has been achieved, we proceed as usual.

Evaluation of whether solution is optimal

This is done by computing shadow prices for each unfilled cell [source/destination combination. It
represents the reduction in costs [negative shadow price] or increase in costs if a single unit is
allocated to the unfilled cell. Degeneracy needs to be resolved before we compute shadow prices.
(i) These may be computed using the stepping stone method or the modified distribution
(MODI) method. (All shadow prices should be zero or greater than zero when the problem
is optimal.
(ii) If the problem is not optimal (existence of negative shadow prices), select the cell with the
most negative shadow price [representing the maximum cost reduction if units are allocated
to the cell] and allocate the largest number of units possible. This will usually be the
number of units in the cell with the least no of units (and a negative) in the computation of
shadow prices.
(iii) Continue with re-allocations until all the shadow prices are equal to greater than zero.
NB: A shadow price of zero indicates there is another solution to the problem. i.e., if units are
moved to this cell, the cost won’t change.

67
Methods of computation of shadow prices
Stepping stone method of computing shadow prices
When this method is used, shadow prices for each unfilled cell are computed by “stepping” on or
using only the filled cells. Reallocations will then have to be made by shifting units from one of
the filled cells to the unfilled cell as in (ii) above.
Let us improve on the solution obtained using the least cost method [which is nearer to the
optimal solution than the northwest corner rule initial allocation] in order to avoid many
iterations. The solution had a cost of Sh. 21 300 and was as in the table below:

X Y Z Capacity
A 500 500
B 100 1100 1200
C 800 1500 2300
Requirements 800 100 3100 4000

Computation of shadow prices for un-occupied cells.

Cell Shadow Price
AX 7 – 1 + 4 – 2 = 8 [We step on AZ; CX and CZ]
AY 6–5+9–2=8
BX 3 – 1 + 4 – 9 = -3 [We step on CX;CZ and BZ]
CY 8–4+9–5=8

Since BX has a shadow price of -3, costs can be reduced [by Sh. 3 per unit] if some units were
allocated to that cell. The maximum no of units that can be allocated to that cell is 800 [the
number of units in the cell with the least no of units (and a negative) in the computation of
shadow prices]. We shall end up with an allocation like below.
X Y Z Capacity
A 500 500
B 800 100 300 1200

68
 C 2300 2300
Requirements 800 100 3100 4000

Note:
This is the same as the first or initial solution using Vogel’s approximation technique. We need to
recheck for optimality by computing the shadow prices of unfilled cells.
Cell Shadow Price
AX 16 – 5 = 11
AY 15 – 7 = 8
CX 10 – 7 = 3
CY 17 – 9 = 8
All the shadow prices are positive and therefore the solution is optimal

MODI (Modified Distinction) method of computing shadow prices

The procedure is as follows:
o We set Ri + Ki = Cij for the cells which are filled, where i represents the row number and
j represents the column number, and Cij the cost of transporting a unit from source i to
destination j
o After writing all equations, we shall have more unknowns than the number of equations
implying that we can’t solve for any Ri or kj. We set, Ri = 0
o Solve the equation for Ri and Ki for all the other values of Ri and all values of Kj
o Compute shadow price for each unfilled cell as:
Shadow price = Cij – Ri – Kj

Let us now use the initial solution obtained using the northwest corner rule, and then set R1 = 0.
The values will be:
R1 + K1 = 7 →K1 = 7
R2 + K1 = 3 →R2 = -4
R2 + K2 = 5 K2 = 9
R2 + K3 = 9 K3 = 13

69
 R3 + K3 = 4 R3 = -9

This shadow prices for un-occupied cells are:

AY : 6 – R1 – K2 = 6 – 0 – 9 = -3
AZ : 2 – R1 – K3 = 2 – 0 – 13 = -11
CX : 1 – R3 – KI = 1 – (-9) – 7 = 3
CY : 8 – R3 – K2 = 8 – (-9) -5 =12

Shadow prices of -3 on cell AY and -11 on cell AZ implies that the solution is not optimal. We
allocate some units to AZ [with least opportunity cost. A maximum of 500 units can be
transferred there. This will actually give us the optimal solution.

Unbalanced problems.
This occurs when source totals [capacity] is not equal to the destination [requirement] totals.
Under such circumstances, we add a source/ destination, referred to as a dummy in order to
equate total capacity to total requirements and proceed as usual. The transportation costs in all the
cells along the dummy source / destination are zero since these represent slack variables as we used
in linear programming

THE ASSIGNMENT MODEL

The assignment problem deals with finding the best one-to-one match for each of a given number
of ‘candidates’ to a number of positions e.g. assigning teachers to classes or workers to machines.
Different benefits or costs are involved in each match and the goal is to maximize the total reward
or minimize the total expense.

This occurs when we need to allocate personnel to jobs in order to minimize costs, e.g., repairmen
Adam, Brown and Cooper would be allocated jobs of repairing a radio, oven, & coffee table with
the following costs.

RadioOven Table

70
 Adams 11 14 6
Brown 8 10 11
Cooper 9 12 7

Procedure in solving for the optimal allocation

o Make an opportunity cost table. This is done by:
i) Subtracting the smallest number on each row [personnel] from every number in the
row and then
ii) Subtracting the smallest number in each column [equipment to be allocated] of the table
obtained after (i) above from every number in that column
o Test for optimality by drawing lines (as few as possible along the row/ column to cross out all
the zeros. If the number of lines is equal to the number of rows/ columns, then an optimal
assignment can be made. If the number of lines is less than number of rows/ columns, then
revise the o/cost table as below.
o Subtracting the smallest number (opportunity cost) not covered from every
other uncovered number and
o Add this number to any number(s) lying at the intersection of horizontal &
vertical lines obtained in (ii above).
o Test for optimality again by drawing the fewest number of lines to cross out the
zeroes.
o If no optimal allocation can be made still, proceed as above until the number of lines is equal to
the number of assignments required, which implies an optimal solution
o In order to allocate the personnel, check whether we have a only one zero along a given row
[personnel] and or a single zero along a column [equipment] and allocate accordingly
Solving our assignment problem above
Step 1
Row opportunity cost table:
R O T
A 5 8 0
B 0 2 3

71
 C 2 5 0
Step 2
Total opportunity cost table:
R O T
A 5 6 0
B 0 0 3 not optimal, a minimum of 2 lines to cross C
2 3 0 out the zeroes

Step 3
R O T
A 3 4 0
B 0 0 5 3 lines →
optimal solution
C 0 1 0
To allocate:
Check any row /column with only one zero.
→Row for Adam has only one zero on tables: allocate the tables to Adams
→Column for oven has zero only with Brown allocated: Allocate accordingly
→Thus, allocate the radio to Cooper
The cost will be: Personnel Equipment Cost
Adam Table 6
Brown Oven 10
Cooper Radio 9
25
NB: If the no. of people to be assigned or the item is not enough, we use a dummy row
[personnel] /column [equipment] will be created. The problem is then solved as before. The
personnel allocated a dummy eventually will be redundant or the equipment allocated a dummy
personnel is not to be allocated to any of the existing ‘real’ personnel

Maximisation assignment Problem

72
Sometimes we may want to maximize the pay-off. Under such circumstances we need to convert
the problem into a minimization problem before solving. The following initial step converts the
maximization problem into a minimization problem:
 Identify the highest value in the payoff table.
 Subtract each pay-off in the table from this value.
 Then proceed as before, since the problem will have been converted into a minimization
assignment problem

Example:
The US army at the gulf has 4 types of planes to patrol the Iraq airspace. These planes have
different ratings in covering different sectors classified as I to IV according to locality, height of
flying etc. The ratings are given below,

Planes I II III IV
A 20 60 50 55
B 60 30 80 75
C 80 100 90 80
D 65 80 75 70

Required:
Determine what type of plane should be allocated what sector and the total points scored over the
surveillance.
Solution
Conversion of problem from maximization to minimization to problem
I IIIII IV
A 80 40 50 45
B 40 70 20 25
C 20 0 10 20
D 35 20 25 30

73
Row opportunity costs
I II III IV
A 40 0 10 5
B 20 50 0 5
C 20 0 10 20
D 15 0 5 10

Total opportunity cost

I II III IV

A 25 0 10 0
B 5 50 0 0
C 5 0 10 15
D 0 0 5 5

Thus the optimal allocation is:

Type of fighter Sector Score
A IV 55
B III 50
C II 100
D I 65
270
Question
The chairman of the business dept in a college wanted to assign four proffers courses to teach
during the forthcoming semester. The chairman evaluates the teaching evaluations (done by
students) over the last 2 years since each professor can take any of the courses.
Statistics Mngt Finance Economics
Professor
Anderson 90 65 95 40
Sweeney 70 60 80 75
Williams 85 40 80 60

74
Kennedy 55 80 65 55
Required:
Determine the course allocations to the professors in order to maximize overall teaching ratings,
and what will be the maximum rating.

QUESTION 1
a) Given below is the transportation schedule used by Pagomo Ltd in distributing product jito
from its 3 plants A, B and C to the markets coded as D, E, F & G

D E FG PRODUCTION
A 200 400 200 200 1000
B 400 100 - 100 600
C 100 - 900 200 1200
DEMAND700 500 1100 500 2800

The transshipment costs from each of the plants to each of the market warehouses are summarized
as below:

D E FG
A 2796
B 5438
C 3124

Required:
i)Determine the transportation costs currently incurred by Pagomo [4 marks]
ii) Using the least cost method, generate an initial transportation schedule and determine the costs
associated with your schedule [6 marks]

75
iii)Evaluate whether the table generated in (ii) above is optimal and comment on your evaluation
(you may use any method of your choice in computing shadow costs)
[6 marks]

b) The management of a ZESA, a utility company wants to allocate / assign three service (S1, S2
& S3) teams to three geographical areas (Z1, Z2 and Z3). Due to each team’s familiarity with each
zone, their efficiency varies, a fact that is reflected I the service costs in thousands incurred by each
team per week of service in a given zone as shown below:
Sercice Tean Geographical zone
Z1Z2Z3
S1 20 15 31
S2 17 16 33
S3 18 19 27
Required:
Advise the management of ZESA on which service team to be allocated each of the three regions I
order to minimize weekly costs and the associated weekly cost[9 marks]

QUESTION 2
a) A factory has 3 plants P1, P2 and P3 producing goods for export. These goods are stored in 3
alternative warehouses W1, W2 and W3. The production capacities in each of the plants and the
storage capacity in the warehouses over a one month period are summarized as below:

Plant Capacity Warehouse Capacity

P1 250 W1 260
P2 330 W2 300
P3 220 W3 240
The table below summarizes the transportation cost per unit from each plant to each warehouse
WAREHOUSE
W1 W2 W3
P1 80 70 100

76
 PLANT P2 60 80 40
P3 50 60 90
Required:
Using a method of your choice, determine the cost minimizing transportation schedule
[12 marks]
b) A firm selling computers has four technicians contracted to undertake after sales service to
clients. Three clients have requested for service.
The costs of sending the technician to each of the clients is given as below,

Technician Client
Gotech Ltd Mon Bureau Hotmal Ltd
Sh. Sh. Sh.
Otieno 7,500 5,400 4,200
Beth 11,400 5,100 6,900
Cate 4,500 4,800 9,000
Damaris 7,800 8,400 8,700
The management assigns technicians to clients at the at the minimum cost possible.
Determine;
(i) the technician to be assigned each of the clients
(ii)the minimum cost to be incurred (10 marks)

QUESTION 3
a) The TamuTamu products company ltd is considering an expansion into five new sales districts.
The company has been able to hire four new experienced salespersons. Upon analysing the new
salesperson’s past experience in combination with a personality test which was given to them, the
company assigned a rating to each of the salespersons for each of the districts .These ratings are as
follows:

77
 Districts
Salespersons 1 2 3 4 5
A 92 90 94 91 83
B 84 88 96 82 81
C 90 90 93 86 93
D 78 94 89 84 88
The company knows that with four salespersons, only four of the five potential districts can be
covered.
Required:
i)The four districts that the salespersons should be assigned to in order to maximize the total of the
ratings (8marks)
ii) Maximum total rating. (2marks)

b) Three ware houses P, Q and R can supply 900, 400, and 800 items respectively. Three stores at
A, B and C require 300; 1200 and 600 items respectively. What is the minimum cost of allocating
the items from the warehouses to the stores if the unit transportation costs are as shown in the
table below:
Transport costs “Sh /item to strore
A B C Total Available
From Warehouse: P Q
R
10
2
1
20
10
20
5
8
7

78
9
4
8
Total Required 300 1 200 600 2 100

Required:
i.How will the items be transported so as to incur the least cost. (14marks)

ii. Define the following terms as used in linear programming:

i)Feasible solution
ii) Transportation problem
iii)Assignment problem (6marks

QUESTION 4
The Steadman research group has been notified that it has received government research grants to
undertake four major projects. The managing director has to assign resrach officer to each of these
projects. Currently there are five research officers-Adams, Ben, Charlie, Dan and Eliud who are
available to carry out these duties. The amount of time required to complete each of the research
projects is dependent on the experience and ability of the research officer who assigned to the
project. The managing director has been provided with an estimate of the project completion time
(in days) for each officer and each project.
Research Officer Project
1 2 3 4
Adams
Ben
Charlie
Dan
Eliud 80
72
96

79
60
64 120
144
148
108
140 60
48
72
52
60 104
110
120
120
92
96
As the four projects have equal priority, the managing director would like to assign research
officers in a way that would minimize the total time (in days) necessary to comlete all four projects
Required:
a) Determine an optimal assignment of research officers to projects and hence determine the total
number of days allocated to these four projects (12marks)
b) State any further allocations that would result in the same total number of days. If research
officers Brown, Charlie and Dan express a preference for projects 2 or 3, while officers Adams and
Eliud express their preference for projects 1 o 4, which of the optimal allocation seems to be the
most sensible for the managing director to make. (8marks)

Question 5
Mtwapa Ltd produces product sigma at 4 plants- A, B, C and D. These are sold at four sales offices
– P, Q, R and S. the production cost differs from one plant to another so that it is Shs 10, Shs 12,
Shs 8 and Shs 9 at plants A-D respectively. Similarly, the selling prices per unit are Shs 30, Shs 32,

80
Shs 28 and Shs 25 respectively for sales point P-S. Given below are the unit transportation costs
from each of the plants to the respective sales offices.

PQ R S CAPACITY
A 5 8 2 2 400
B 7 6 3 4 750
C 10 50 7 6 520
D 4 8 5 4 330
DEMAND800 250 650 300 2 000

Required
a) Determine the profit matrix for Mtwapa Ltd
b) Determine the optimal allocation so as to maximize profits and the maximum PROFIT

6: CALCULUS
DIFFERENTIAL CALCULUS
The relationship between the independent variable and the dependent variable may be presented
in form of a graph. In some business applications, this relationship may be linear, in which case it
is represented by:
Y=a+bX
Where Y is the dependent variable (plotted on the vertical (Y) axis
X is the independent variable (plotted on the horizontal (X) axis,
‘a’ is a constant, the Y-axis intercept or value of dependent variable when
independent variable is equal to zero (0)
‘b’ is a constant, the slope of the function, representing change in the
dependent variable (Y) as a result of a change in the dependent variable
Sometimes however, the relationship between the independent and dependent variables may not
be linear. For such curvi-linear functions, the slope changes with the level of independent variable.
Differential calculus deals with the determination of the slope of such non-linear functions. The

81
derivative (differential) of a function is given as dy/dx, representing change in Y (dependent
variable) as a result of changes in X, (independent variable).

General rules of differentiation

1. The derivative of a function Y =axn+bxm+c is given by:
dy / dx =anxn-1+bmxm-n
e.g.; If Y = 5X3 +10X2 – 16X + 50
Then, dy/dx = 15X2 +20X - 16

2. Product rule:
For a function a product function such that Y = U.V, where both U and V are expressed
in form of the dependent variable, then;
Dy/dx = Udv/dx + Vdu/dx
e.g., If Y = (3X2 + 5X) (9X3 - 10X2)
Then U = 3X2 + 5X and du/dx = 6X + 5
V = 9X3 – 10X2 and dv/dx = 27X2 – 20X

dy/dx = Udv/dx + Vdu/dx = (3X2 + 5X) (27X2 – 20X) + (9X3 – 10X2) (6X + 5)
When expanded, this comes into:
dy/dx = 81X4 – 60X2 + 135X3 – 100X2 + 54X4+ 45X3 – 60X3 – 50X2
And when the like terms are brought together, it comes into;
dy/dx =135X4 + 120 X3 - 210 X2
NB:
If the original function was put in its expanded form by opening the brackets and then
differentiating it, we would end up with the same result. The expanded form of the
function will be given by:
Y = 27X5 – 30X4 + 45X4 - 50X3
Y = 27X5 + 15X4 - 50X3

dy/dx = 135 + 60 – 150

82
 3. The chain rule
When Y = (U)n, where U is a function of X, the independent variable, then:
dy/dx = dy/du.du/dx
e.g., If Y = (3X2 + 8X)4, then:
Y = U4 where U = 3X2 + 8X
dy/du = 4U3 and du/dx = 6X + 8
dy/dx = 4U3(6X + 8) = 4(3X2 + 8X)3. 6X + 8

4. Quotient rule
When Y is a quotient function such that Y = U/V, where both U and V are functions of X
the independent variable, then;
dy/dx = (Vdu/dx – Udv/dx) / V2
e.g., If Y = (5X2 + 15X) / (2X – 5), then:
U = 5X2 + 15X and du/dx = 10X + 15
V = 2x – 5 and dv/dx = 2
Therefore dy/dx = [(2X – 5) (10X + 15) – 2 (5X2 + 15X)] / [(2X – 5)]2

Applications of differential calculus

Differential calculus has applications in business in the following areas.
a) Determining marginal costs and the cost minimizing quantity of output as well as the
related costs
b) Determining marginal revenue and the revenue maximizing quantity of output as well as
the related revenue
c) Determining the marginal profit and the profit maximizing quantity of output as well as
the related profits
d) Analyzing the effects of price changes on revenue i.e., analysis of price elasticity of demand.

Cost applications

83
For a total cost function, the first derivate (dTc/dQ) gives the marginal cost. Costs are minimized
when the marginal cost is equal to zero. The second order condition required for minimization is
that the second derivative d2TC/dQ2 > 0
Revenue applications.
The derivative (differential) of the total revenue, i.e., dTR/dQ is the marginal revenue. Revenue is
maximized when the marginal revenue is equal to zero. The second order condition for
maximization is that the second derivative should be less than zero (0) or negative; i.e.;
d2TR / dQ2 < 0

Profit (Σ) application

Profit (Σ) = Total revenue – total costs. This implies that we get the profit function by deducting
the cost function from the revenue function. The 1st derivative of the profit function (dΣ/dQ) gives
the marginal profit. Profits will be maximized when the marginal profit is equal to zero. The 2nd
order condition for maximization should be met i.e., the second derivative d2Σ / dQ2 < 0
Alternatively profit is maximized when marginal costs = marginal revenue i.e., dTC/dQ
(Marginal cost)=dTR/dQ (Marginal revenue)

Analysis the effects of price changes on revenue

The price elasticity of demand is defined as the proportionate change in quantity demanded as a
result of a proportionate change in the price of the commodity. This is given by:
|ep| = proportionate (%) μ in Qty dd / proportionate (%) μ in price.
=dQ/Q/dP/P
= dQ/Q *P/dQ dQ/dP * P/Q

|ep| = Σ

|ep| > 1

|ep| = 1

|ep| < 1

84
 Q* |ep| = 0 Quantity
Note:
Total revenue (TR) is maximized when |ep| = 1, at quantity Q*
If |ep| > 1 prices should be reduced in order to maximize total revenue
If |ep| < 1 prices should be increased in order to maximize total revenue
ILLUSTRATION
Kaloleni Ltd manufactures a single product. The demand curve for the product is given by the
equation:
P = 5 600 – 4Q, where P is the selling price per unit in shillings and Q is the number of units sold
The average cost per month is given by the equation:
AC = 164 000/Q – 700 + 2Q
Required:
a) Determine the marginal revenue function and the marginal cost functions expressed in terms of
the output, Q [5 marks]
b) Determine the quantity that Kaloleni should produce in order to:
i) Maximize revenue and the maximum revenue per month [4 marks]
ii) Minimize the total costs and the minimum cost per month [4 marks]
iii) Maximize the profit and the maximum profit in a month [4 marks]
iv) The selling price per unit, the average cost per unit and the total fixed costs at the quantity
determined in (iii) above [3 marks]
c) Kaloleni is operating at the revenue maximizing level of output determined in (i) above. If the
objective of the company is to maximize profits, explain whether the company should consider
a price increase or decrease, and the change in price per unit you would recommend and its
effect on the profit [5 marks]

Solution:
a) i) Total revenue (TR) = Price x Qty. (P*Q)
TR = (5 600 – 4Q).Q

85
 TR = 5 600Q - 4Q2
Total cost is given as: TC = AC*Q
= (164 000/Q – 700 + 2Q)*Q
= 164 000 – 700Q + 2Q2
Marginal cost (MC) is given by = dTC/dQ
= -700 + 4Q
i) To max Total revenue (TR), Marginal revenue (MR) = 0
MR = dTR/dQ = 5600 – 8Q = 0
8Q = 5 600
Q = 700
For second order condition: d2TR/dQ2 = -4; (-ve (less than zero) implies it is a maximum)
TR = 5 600 *700 – 4*(700)2
= 1 960 000
ii) To minimize total costs (TC), marginal cost (MC) = 0
Therefore -700 + 4Q = 0
4Q = 700
Q = 175 units
For second order condition: d2TC/dQ2 = 4; (+ve (more than zero) implies it is a minimum)
Minimum costs by substituting are given by:
TC = 164 000 – 700*175 + 2*(175)2
= 102 750
iii) In order to maximize profits (Σ);
Σ = TR –TC
= 5 600Q – 5Q2 – (164 000 – 700Q + 2Q2)
= 5 600Q – 5Q2 – 164 000 +700Q - 2Q2
= -7Q2 + 6 300Q - 164 000
Profit (Σ) is maximized when dΣ/dQ = 0
= -14Q + 6 300 = 0
14Q = 6 300
Q=450 units

86
 D2/dQ2 = -14 (-ve, implies that it is a maximum)

Maximum profits (Σ) by substitution:

Σ = -7Q2 + 6 300Q – 164 000
At Q = 450
Σ = -7(450)2 + 6 300 (450) -164 000
Max Σ = Shs 1 253 500

OR ALTERNATIVELY
For profit (Σ) maximization, Marginal cost (MC) = Marginal revenue (MR)
This then implies that: -700 + 4Q = 5600 – 8Q
14Q = 6 300
Q = 450 units
Profit (Σ) = TR -TC
TR = 5 600*450 – 4* (450)2 = 1 710 000
TC = 164 000 – 700*450 + 2* (450)2 = 254 000
Profit (Σ) 1 456 000

iv) At an output level of 450 units:

Selling price per unit = 5 600 – 4*450 = Shs 3 800
Average cost per unit = 164 000/450 – 700 + 2*450 = 564.44
Total fixed cost don’t change = Shs 164 000

b) Currently, the company is operating at Q = 700 (Revenue maximizing) at a price of:

P = 5600 – 4*700 or Shs 2 800 per unit.
In order to maximize profit, Q = 450 and price is Shs 3 800 as per (iv) above
The company should therefore increase price by Shs 1000 which will reduce the quantity
sold to 450 units

87
 GAMES THEORY & PROBLEMS
In a competitive environment, the competing parties devise different strategies for success.
Amongst the various possible alternatives of strategies, the selected for the purpose of
making an effective decision. In a business situation where two competitors are involved,
the decision maker has to study the move of the competitor. Therefore, a ‘game’ between
the concerned competitors arises as a result of actions and interactions that are involved.
Business firms competing with other another are of various categories. There are those
firms that try to find the best amongst the various media of advertising such as radio,
newspapers, TV, etc. also, business firms dealing with more or less similar products try to
explore various strategies for capturing each other’s markets and thereby attract customers.
In matters of defence involving Army, Navy and Air force, there can arise various
‘competing strategies’ or games evolving attacks between the forces at war.

88
Preliminary Details:
1) The participants in a game who take decisions are called ‘Players’.
2) A finite number of all the possible outcomes of actions that are open to
a player are called ‘Strategies’.
3) When a player selects one of the strategies from among those available to
him, then a ‘Pay off’.
4) An outcome which is a result of every combination of strategies adopted
by players is called a ‘Pay Off’.
5) When the different games that result from a game are presented in a
tabular form, it is called a ‘Play off Matrix’.
6) The maximum element from amongst the minimum elements pertaining
to each respective row is called ‘Maxmin’.
7) The minimum element from amongst the maximum elements pertaining
to each respective row is called ‘Minimax’.
8) When the maximum value of a player and the minimax value of his
opponent happen to result in an identical element then that element is called ‘Saddle Point’.
9) If two persons play a game and when the algebraic sum of the games of
both the players at the end of the play is zero, then such a game is called ‘Two person-zero-
sum game’. It should be noted that the gain of a player is loss of his opponent. The
following example relates to a payoff matrix of a zero sum two persons’ game.
10) Pure strategy: A player ‘A’ adopts minimax strategy and B adopts
maxmin strategy. Here saddle point is said to be a pure strategy game when there is a saddle
point. Game without a saddle point necessitate the use of mixed strategies. A player’s best
strategy would be to maximize his gain or minimize his loss.

Problem – 1:
Player B

I II
I 28

89
II 89

Find the best strategy for each player and the value of the game.

Solution:
If Player A adopts strategy I and B also adopts strategy I, then the gain to A would be 2 (B
pays A 2 units).
Here, by adopting strategies A finds the minimum gains and tries to obtain the maximum
amongst these minimum gains.
By adopting strategies 1 and II B finds his maximum losses and tries to obtain the minimum
amongst those.

Player B Row minimum

28 2
Player A II 8 9 8 Maximum

Column Maximum 8 9

Minimax

Minimax = Maxmin = 8
The 8 units represents the element which is called the ‘Saddle Point’ because the Minimax
= Maxmin.

The value of the game is 8 as it represents the game of the player A. The optimal strategies
of players are written as

Player A : (0, 1)
Player B : (0, 1)

90
The optimal value of the game is 8.

Solve the following game.

Player B

I II
Player A I 68
II 8 9

Solution:
Player B Row minimum
I II
Player A I 6 8 8 Maxmin
II 8 9

Column Maximum 8 9

Minimax

Saddle Point = 8
Value of the game= 8

Problem 3:
Solve the following game.

Player B
I II II IV V
I 12345
II 3 7 5 4 2

91
 III 1 5 4 3 7
IV 6 7 5 8 5
V 14352

Solution
Player B Row Minimum
123451
Player A 375422
675851
143251
Column Maximum 6 7 5 8 7

Minimax

Saddle Point is 5.
Optimal strategies for A : (0,0,0,1,0)
Optimal Strategies for B : (0,0,1,0,0)
Value of the game = V = 5

Mixed Strategy Games

If in a game, there does not exist a saddle point, then such a game is called a Mixed Strategy
Game’. The following theoretical procedure explains the method for solving a Mixed
Strategy Game.

Suppose the game is:

Player B
g11 g12
Player A g21 g22

92
We assume the probability that player B commences with the 1st row to be ‘b’ and that for
commencing with the 2ndrow to be 1-b
If B plays the 1st column then A gains.
g11 g21 (1-a) ………………………… (1)
If B plays the 2nd column, then A gains

g12 a + g22 (1-a) ………………………… (2)

Taking results (1) & (2) as identical, we get

g11 a + g21 (1-a) = g12 a + g22 (1-a)

Solving and simplifying, we get

g22 - g21
a=
(g11 - g21) + (g22 - g12)

G11 - g12
And 1 - a =
(g11 - g12) + (g22 - g12)

Proceeding in a similar manner, we can find the optimal strategies for B. If A plays the 1st
row then the gain to B would be:

g11 b + g12 (1-b) ………………………… (3)

If A play the 2nd row, then gain to B would be:

93
 g21 b + g22 (1-b) ………………………… (4)

From results (3) & (4) if follows

g11 b + g12 (1-b) = g21 b + g22 (1 – b)

Solving and simplifying, we get

g22 - g12
b=
(g11 - g21) + (g22 - g12)

g11 - g12
And 1 - a =
(g11 - g21) + (g22 - g12)

Now, if A plays(a, 1-a)

Then value of game is
g11a + g21 (1-a)
Substituting the value of ‘a’ & ‘1-a’ in the above result, we get the value as:
g11 g22 - g11 g21 + g21 g11 –g21 g12
=
(g11 - g21) + (g22 - g12)

Problem -4:
Find the value of the following game:
Player Y Row Minimum
94 4
Player X 5 6 5 Maxmin

Column Maximum 9 6

94
 Minimax
Here Maxmin Minimax
So there is no saddle point.
Therefore the game is a mixed strategy game.

If 9 4 gg
5 6 gg

Then g11 = 9, g12 = 4, g21 = 5, g22 = 6

Putting these values in the formulae stated earlier, we get
g22 - g21 = . (6 - 5)
a= (9 – 4) + (6 -5)
(g11 - g21) + (g22 - g12)
1 =1
=
5+1 6

1 – a = 1- 1 = 5
6 6
Further,

g22 - g21 = . (6 - 4)
b= (9 – 5) + (6 -4)
(g11 - g21) + (g22 - g12)

2 =2
=
6 3

95
 1 =2
1-b = 1-
3 3

Value of the game

g11 g22 - g21 + g21 g12
=
(g11 – g12) + (g22 – g22)

(9 x 6)-(5 x 4)
=
(9 - 4) + (6 -5)
54 - 20 = 2 = 5 2 Units
=
5+1 6 3

Interpretation
From the above working it follows that the player X can play the first row for (1/6)Th of
the time and the second row for (1/6)TH of the time and the second row for (5/6)TH of
the time. By doing so he can gain at least 5 2/3 units. At the same time Y can play the first
column for (1/3)rds of the time and the second row for (2/3)rds of the time. By doing so,
he can restrict his loss to 5 2/3 units.

Simplified method
The method of solving a mixed strategy problem can be simplified by using an alternative
algebraic as explained below:

Player B Row Diff. Ratios Interchanging Ratios

Player A 9 4 9 – 4 = 5 5/6 1/6
5 6 6 – 5 = 1 1/6 5/6

96
Column Difference 9 – 5 6-4
Sum = 4 + 2= 6
Ratios = 4/6 2/6
Interchanging
the ratio 2/6 4/6
i.e. 1/3 2/3
Strategy for A is (1/6, 5/6)
Strategy for B is (1/3, 2/3)
Value of the game is
= g11a + g21 (1 – a)
= 9(1/6) + 5(5/6) = 9/6 + 25/6 + 34/6 + 5 ½ Units.

Mixed strategy games of sizes 2Xq and pX2

If a game, the strategies available to a player X are two and the strategies of his opponent Y
are more than two (say q) then the game is called a game of size 2Xq. Similarly, we can
have a game of size pX2.

Procedures for solving a 2Xq or pX2 game

(I) If a saddle point exists, then the strategies that pertains to the saddle
point became the optimal strategies of the players. The saddle point value is the value of the
game.
(II) If a game has no saddle point, then depending upon the nature of the
game, one of the following methods can be used.
1. Dominance method
2. sub-game method
3. Graphic method

Dominance Method

97
According to this method, if every element of the row is greater than or equal to the
corresponding element of every other row, then such a row is called the dominating row
and the row over which it dominates is called the dominated column.
By elimination of such dominated rows and columns (strategies) in a game, the pay off
matrix in a game gets simplified for arriving at a solution.
This procedure of solving by dominance is explained in the following problems.
Problem -5:
Player Y
I II III
I 9 7 4
Player X II 3 4 1
III 6 8 7
Here every element of the 1st row is greater than the corresponding element of the second
row. Therefore the dominated row is 2nd row. By eliminating this row, we get
Player Y
I II III
I 9 7 4
Player X III 6 8 7
Here every element of the 3rd column is less than the corresponding element of the 2nd
column, we get
Player Y
I II
I 9 7
Player X III 6 8

As there is no saddle point, this game can be solved as a mixed strategy game

Player Y Row Diff. Ratios Interchanging Ratios

I II
I 9 7 9–7=2 2/4 1/2

98
Player X
III 6 8 8–6=2 2/4 1/2
Sum 4

Difference 9 – 6 = 38 – 7 = 1
Sum of diff. = 3 + 1= 4

Ratios =¾¼
Interchanging
The ratio ¼¾

Value of the game is

= 9 (1/2) + 6 (1/2)
= 9/2 + 3 = 7 ½ Units.

Sub games Method

If the dominance method becomes inapplicable in certain types of games, the Sub-games
method is suggested. The following are the various steps for using this method.

(i) it should be assumed that the saddle point does not exist and that the
dominance method is inapplicable.
(ii) The given matrix should be broken into games of 2X2 matrices and each
sub-games should be solved separately.
(iii) We select the sub-game which provides the optimal gain to a player.

The procedure of sub-games method has been explained with the help of problem No. 6v
below.
Problem - 6.
Player B

99
 138
Player A
642
This matrix can be arranged to form the following sub-games:-
1 3 38 18
6 4 42 62
(i) (ii) (iii)

Solving each sub-game, as under:

(i)
Player B Row minimum
13 1
Player A 6 4 4 Maxmin

Column Maximum 6 4
Minimax
Saddle point exists
The value of the game is 4.
(ii)
3 8
4 2
This sub-game has no saddle point. Therefore we solve it by the mixed strategy method.
Player B Row Diff. Ratios Interchanging Ratios
3 8 -5 5/7 2/7
Player A
4 2 -2 2/7 5/7
Sum -7

Difference 1 6
Sum of diff. = 7 6/7

100
Ratios = 1/7 6/7
Interchanging
The ratio 6/7 1/7

Value of the game is

= 3 (2/7) + 4 (5/7)
= 6/7 + 20/7 = 26/7 = 3 5/7 Units
(iii)
1 8
5 2

As there is no saddle point, we solve it by the mixed strategy method.

Player B Row Diff. Ratios Interchanging Ratios
1 8 7 7/11 4/11
Player A
6 2 4 4/11 7/11
Sum 11
Difference 56
Sum of diff. = 11
Ratios = 5/11 6/11
Interchanging
The ratio 6/11 5/11

Value of the game is

= 1 (4/11) + 6 (7/11)
= 4/11 + 42/11
= 46/11 = 4 2/11 units.
Among all the 3 values i.e. 4, 3 5/7 and 4 2/11, we find that the optimal solution with
respect to B is the value which gives minimum loss. This value of the game is decided in
terms of the minimum loss to B.

101
Problem – 7:
Player B
32
Player B 1 4
57
This matrix can be arranged to form the following sub games:-
3 2 14 32
1 4 57 57
(i) (ii) (iii)

Solving each sub game, as under:

(i) Player B Row Minimum
32 2 Maxmin
Player A 14 1

Column Maximum 3 4
Minimax
Here there is no saddle point.

Therefore using mixed strategy procedure, we get

Player B Row Diff. Ratios Interchanging Ratios
3 2 3-2 = 1 ¼ ¾
Player A
1 4 4-1 = 3 ¾ ¼
Sum 4

Difference 2 2
Sum of diff. = 4
Ratios = 2/4 2/4
Interchanging

102
The ratios ½ ½

Value of the game is

= 3 (3/4 + 1 (14)
= 10/4 = 2 ½ Units.

(ii) Player B Row Minimum

14 1
Player B 57 5 Maxmin

Column Maximum 5 7

Minimax

Here a saddle point exists.

Therefore the value of the game is 5.

(iii) Player B Row Minimum

32 2
Player A 57 5 Maxmin

Column Maximum 5 7

Minimax

Here the saddle point is 5.

Therefore the value of the value of the game is 5.

103
From the three strategies that are the available to A, we find that the maximum value of the
game for A is 5. Hence the strategy of the game is decided in terms of the maximum gains
to A, which is 5 units.

QUESTION 1
You are provided with the following pay-off table
PLAYER Y
Y1Y2Y3
X1 3 5 -3
Player X X24 8 6
X3 -1 -4 3

Required
a) Interpret the values ‘5’ and ‘-1’ in the table
b) Using the payoff table, explain the concept of dominance as used in
game theory, clearly indicating whether there is any strategy that is dominated for each of
the players
c) Using the minimax and the maximin rules, establish whether there is a
saddle point for this game

QUESTION 2
Solve the following pure strategy game for players X and Y and determine the game value
PLAYER Y
Y1Y2
X1 6 10
Player X X28 12
X3 5 7

104
 QUESTION 3
You are provided with the following pay-off matrix for players X and Y
PLAYER Y
Y1Y2
X1 5 8
Player X
X2 11 6
Required
Solve the game to determine the proportion of the tie each player uses each of his strategies
and determine the game value

QUESTION 4
You are provided with the pay-off table. Use the table to determine the game value and the
optimal strategy for each of the competitors
PLAYER Y Player B
Y1Y2Y3 B1 B2
X1 9 7 4 A1 8 9
Player X X23 4 1 Player A A2 12 8
X3 6 8 7 A3 7 10
QUEUING THEORY
This theory deals with problems that involve queuing or waiting. There are many areas where
queues are experienced in the service industry eg supermarkets banks airports etc. queues form /
arise because resources are limited. The essence of queuing theory is to minimize the queues (the
no of units in the system and the costs associated with the queue.) ie determination of the
optimum service rate so as to minimize the costs associated with the queue. This may be illustrated
graphically as follows.

105
 Total costs
Service costs

Costs

Waiting costs

Slow (longer queues) Service rate Fast (Shorter queues)

Components of a queuing system

The queuing system can be broken down into various sub-components or sub-systems. These
include:
Arrival
This describes the pattern of arrival of units or persons in the queue. Arrivals are assumed to be
random and to follow a Poisson probability distribution.
Service mechanism
The time taken by a server to service a customer is referred to as service time. The time spent in
the system is made up of two components i.e., the time of waiting in the queue (waiting time) and
service time. Queuing theory presumes that the time taken to service each customer is
independent in each case and follows an exponential distribution. It is not discrete and therefore it
can take some fraction time.
Queuing discipline
This deals with how the queue is formed and the determination of the customer to be served next.
Terminologies of importance with respect to queuing discipline include:
Balking – a customer decides not to join the queue since it is too long.
Reneging - this occurs when a customer/person /unit joins the queue but leaves it if they
have waited for too long for the service.
Jockeying – this occurs when customers switch between queues if they think they will get
served faster.

106
NB:
In queuing theory we assume that there is no balking, reneging or jockeying
Servers
Queues can be classified according to the number of service points and how service is organized.
Important queuing systems include
Single server – single queue (single channel facility)
Multiple servers – Multi server (multiple channel facility)
Sequential queues
In this course, we are interested in single channel queuing model
Single channel Queuing model
The conditions required for a single channel facility include:
1. The number of arrivals per unit of time follows a Poisson probability distribution with a mean
arrival rate μ (lambda)
2. The service time has an exponential distribution. The average service rate is given by Σ (mu)
3. Arrivals are from an infinite population
4. Queuing discipline is first in first out.
5. There is only one single server or service station.
6. The mean arrival rate (μ) is less than the means service rate (μ) i.e., μ < μ
7. The waiting space available for customers in the queue is infinite and can therefore take any
number of customers

In a single channel queuing model, the following formulae are important:

Question 1
Customers arrive at a sales counter at an average rate of 20 customers per hour. The arrival rate is
random and may be assumed to follow a Poisson distribution. The time required to serve each
customer has an exponential distribution with a mean of 2 minutes.

107
Required:
a) The average waiting time in the queue
b) The average waiting time in the system
c) The average number of customers in the queue
d) The average number of customers in the system
e) Probability that there are no customers in the system

Solution
a) Average waiting time in the queue = μ/ μ (μ -μ)
μ = 20 customers/hr
1 hr = 60 minutes: Σ = 60/2 = 30 customers per hour =
Average waiting time in the queue = 20/30(30-20) = 20/300 hours =
0.067 hrs or 4 minutes = 1/15
The average waiting time in the system (Ws) = 1/ (Σ -μ)
= 1/ (Σ -μ) = 1/(30 -20) = 1/10 = 0.1hours = 6 minutes
b) The average no of customers in the queue (Lq) = μ2/ Σ (Σ-μ)
= (20)2/20(30 – 20)
= 400 / 200
= 2 customers
c) The average no of customers in the queue (Ls) = μ/( μ - μ) = 20/10 = 2 customers
d) Probability that the system is idle;
P(idle) = 1 – (μ/ μ)
= 1 – (20/30) = 1 – 0.6667 = 0.333

Exercises
Question 1
The arrival rate of customers in a barber shop follows a Poisson probability distribution with a
mean of 5 customers per hour. It takes the barber an average of 8 minutes to do a haircut, with the
average time following an exponential distribution.
Required:

108
a) The average number of customers in the system
b) The average waiting time in the system
c) Compute the probability of having 3 customers in the queue

Question 2
Aero planes arrive at Jomo Kenyatta International Airport at an average rate of 20 planes per
hour. The arrival is random and may be assumed to follow a Poisson distribution. The rate at
which the planes are given permission to land depends on the weather, with a mean landing rate of
40 planes per hour in good weather and 25 planes per hour in bad weather.
Required:
Under each of the weather conditions, determine
a) The average number of planes circling the airport
b) The average number of planes in the system: i.e., in the process of landing or circling the
airport
c) The average waiting time in the system
d) The average waiting time in the queue
e) The time a plane spends in the ‘landing’ process itself

9: SIMULATION
 The process of simulation
 Importance of simulation to managers
 Areas of application of simulation in an organization
 Limitations of simulation

109
Tusian & Pandey (2008) define simulation as a numerical technique for conducting experiments ….
which involve certain types of mathematical and logical relationships necessary to describe the
behavior and structure of a complex real world system over extended period of time. They see it as
a quantitative model that uses computerized mathematical model to represent actual decision
making under conditions of uncertainty. Alternative courses of action are evaluated on the basis of
facts and assumptions. Repeated experiments are performed to re-create a real world system
representing an observed phenomenon. The model uses probabilistic input variables (at times
together with controllable inputs) to generate the output. This may be illustrated as below.

CONTROLLABLE INPUTS

PROBABLISTIC SIMULATION MODEL OUTPUT

INPUTS

The values for the probabilistic input variables are generated randomly. The model for generating
these inputs uses random numbers as well as probabilities and is referred to as the Monte Carlo
System. After a series of experiments using the simulation model, the analysis can show the
relationship between the controllable inputs that will provide desired output of the real system.

PROCEDURE FOR THE MONTE CARLO SIMULATION

The procedure may be summarized as follows:
1. Define the controllable variables involved in the model.
2. Generate a probability table for each of the probabilistic inputs, using data collected relating to
the problem being investigated,
3. Generate a series of random numbers to be applied in the model relating to each of the random
or probabilistic inputs. This can be generated from a calculator, computer or from random
number tables.

110
4. Determine the output associated with each combination of probabilistic inputs and
controllable inputs and record the simulated result.

This procedure is iterative in nature and many repeated trials have to be performed before a real
world situation representing the phenomena under analysis is presented. It may be illustrated as
follows:

Define the
controllable variables

Generate probability
table for each
probabilistic inputs

Use random numbers to allocate

values the probabilistic inputs

generate/compute value of the

objective(apply the model)

Record Simulated result

111
Illustration
A small garage was opened along the Mombasa Malindi road. An observation was made of the
number of vehicles that stopped by for refueling. You are provided with the following
information on the statistics for the first 50 days of operation
No of vehicles refueling 5 10 15 20 25 30
No of days: 2 6 8 16 12 6

The number of litres of petrol filled into the cars was observed over the same period and is
summarized below:
Qty of fuel (litres/car) 10 15 20 25 30 35 40
No of days 4 6 16 10 6 6 2

Petrol is sold at a price of Shs 75 per litre, with a variable cost of Shs 48 per litre. The company
incurs variable costs of Shs 50 per vehicle refueling at the station, and the fixed costs per week
amount to Shs 15 000.
Required
Simulate the process over the next one week and determine the profit / loss based on the
following random sequence of numbers for the number of vehicles and then the quantity of fuel
respectively: 1749820645786362965338980432

Solution:
No. of vehicles: 5 10 15 20 25 30
No. of days 2 6 8 16 12 6
Probability: 2/50 = 0.04 012 0.160.320.240.12
Cumulative probability 0.04 0.160.320.640.88 1.00
Random numbers allocated 00 -04 05-15 16-31 32-63 64-87 88-99

Quantity of fuel: 1015 20 25 30 35 40

No of days 4 6 16 10 6 6 2
Probability 0.08 0.120.320.20 0.12 0.120.04

112
Cumulative probability 0.08 0.200.520.72 0.84 0.961.00
Random numbers: 00 -09 10-19 20-51 52-71 73-83 84-95 96-99

Model for determining profits:

π = No. of litres fueled * profit or contribution per litre – total variable costs – total fixed
costs

The problem is simulated as below

DA RAND 1.NO. RAND 2.FUE TOTAL FUEL SERVIC NET

Y OM OF OM L FUEL CONTR E COST CONTR
NUMBE CARS NUMBE LTRS LTRS IB CARS*5 IB
R R LRS*27 0
1 17 15 49 20 300 8100 750 7350
2 82 25 06 10 250 6750 1250 5500
3 45 20 78 30 600 16200 1000 15200
4 63 20 62 25 500 13500 1000 12500
5 96 30 53 25 750 20250 1500 18750
6 38 20 98 40 800 21600 1000 20600
7 04 10 32 20 200 5400 500 4900
TOTAL 84 800
FIXED 15000
COSTS/WK
PROFIT 69800

Exercise
In a Quantitative methods class, lateness has been observed for the last few days. You are provided
with the following data:
No. of late students:0 1 2 3 4 5 6
Probability: 0.1 0.130.18 0.20 0.250.070.04

113
Required
Simulate the lateness of the students for the next 12 days and compute the average number of late
students based on the simulated data

Simulation : Its Application

(i) The process of simulation can be applied to a variety of problem that are subject to
quantification and estimation.
(ii)It is a very useful tool that is popularly and reliably used for solving business problems.
(iii) Its use extends to various aspects of military operations, strategic and tactical moves. It can be
used in problems relating to police personnel used for patrolling and in traffic operations. Also,
problems involving the dispatch of hospital-vans in inventory control policies.
(iv) The technique of simulation can be advantageously used in inventory control policies.
(v) In matters of financial planning, pertaining to capital budgeting and selection of portfolios,
these procedures of simulation can be applied.
(vi) Computer simulation processes are for studying the behaviour of nuclear reactors during
operating and accident situations. By the use of this method, the design can be made ‘fail safe’.
(vii) Its applications ion decision making problems is of much significance especially in avoiding
unexpected risks. Computer software packages can also be evolved suitably.

Simulation : its Merits and Demerits.

A) Merits
(i) Through the process of simulation it is possible to overcome complications and complexities
that may arise in certain problems, which require use of rigorous and excessive mathematical
calculations.
(ii)In real problems, if the procedure of experimentation is adopted then it involves costly risks
which can be avoided by the use of simulation techniques.
(iii) The long term implications of managerial decisions can be easily foreseen by a research analyst
through the use of simulation.

114
(iv) It is always preferable to test in an appropriate manner, a simulation solution prior to its
acceptable as the best solution.

B) Demerits
(i) It should be noted that the process of simulation may not guarantee a perfect optimal solution
to a problem. Obviously its methodology does not ensure optimality.
(ii)It is applicable only if some data is provided. In other words, its basis rests on the assumptions
of a certain system that can be fitted to a given situation.
(iii) The procedure of simulation does not provide mathematically accurate results but only yields
estimates based on certain suppositions.
(iv) In certain problems the process of simulation involves sampling error. This can be avoided
only if the sample size is modified.
(v) Its use in certain problems may not prove to be economical.
(vi) Unless the concerned problems are amenable to random phenomena, it may not be possible to
use to Monte Carlo simulation technique.
(vii) The procedure of simulation aims at finding a solution to a problem. However such a solution
may not be precise so as to be considered as the correct and final solution to a problem

Question 1

The demand curve for production Alpha is given by the function

P = 12 600 – 5Q Where P is price /unit and Q the quantity of Alpha

On the other hand, the total cost equation is given as

TC = 150 000 - 900Q + 4Q2

Required

(a) Determine the quantity of Alpha to produce in order to

115
i)maximize total revenue

ii) to maximize total profits

iii) to break-even

b) Currently the production of Alpha is 800. The sales director is of the opinion that a reduction in selling
price per unit will lead to an increase the market share. The Managing director is of the opinion that the
price should be increased in order to increase the total revenue. Advice the management on whether to
increase or reduce the price per unit if their objective is to maximize the total revenue

c) The production director recommends that production should be increased to 1 000 units so as to take
advantage of current low material price. Determine the effect on profits if this recommendation is adopted

Question 2

A Public transportation company has been experimenting on a possibility of a developing a system of

charging fares. The demand functions, which expresses the ridership as functions of fair charged is given
below:

Q=10,000-125P

Where Q equals the average number of riders per hour and p equals the fare in the shillings.

Required:

[a] Determine the fare , which should be charged in order to maximize hourly bus fare revenue.

[b] What is the expected maximum revenue?

[c] How many riders per hour are expected under this figure?

Question 3

A major cosmetic and beauty supply firm which specializes in door-to door sales approach, has found that
the response of sales to the allocation of additional sales representation behaves according to the law of
diminishing returns. For one region sales district, the company estimates that annual profit P, stated in
hundred of shillings, is a function of the number of sales representations x assigned to the district. The
function these two variables is

116
P = f (x) = - 12.5x2 + 1,375x – 1,500

a) What number of representations will result in maximum profit for the district?

b) What is the expected maximum profit?

Question 4

A Company estimates that the demand for its products fluctuates with the price it charges. The demand
function is q = 280,000 – 400p

Where q equals the number of units demanded and p equals the price in dollars. The total cost of producing
q units of the product is estimated by the function.

C = 350,000 + 300q + 0.0015q2

Required:

a) Determine how many units q should be produced in order to maximize annual profit.

b) What price should be charged?

Question 5

The cost accountant of a firm producing colour television has worked out the total cost function for the
firm as TC= 120Q – Q2 + 0.02Q3. A sales manager has provided the sales forecasting function as P= 114 –
0.25Q where P is price and Q the quantity sold Where P = Price in £

Required:

i)Find the level of production that will yield minimum average cost per unit and determine whether this
level of output maximizes profit for the firm.

ii) Determine the price that will maximize profit for the firm.

iii) Determine the maximum revenue for this firm.

Question 6

117
A firm analyzed its operating conditions, prices and costs and has developed the following functions:

Revenue = 400Q – 4Q2 and cost = Q2 + 10Q + 30 where Q is the number of units sold. The firm wishes
to maximize profit and wishes to know:

a) What quantity should be sold?

b) At what price?

c) What will be the amount of profit?

d) Find the point of maximum value of the revenue function?

Question 7

The marginal revenue of a firm is given by:

MR = 1 800 – 4Q, where Q is the quantity sold in a week.

The marginal cost is given by:

MC = 6Q – 900. Total fixed costs per week amount to Sh. 84,000

Required:

a) Determine the additional revenue associated with increase in sales from 300 to 400 units

b) Determine the additional costs associated with increase in sales from 300 to 400 units

c) What will be the change in profits when sales are increased from 200 to 300 units

d) Determine the quantity to produce in order to:

i)Minimize costs

ii) Maximize revenue

iii) Maximize profits

Queuing theory

118
Question 1

Customers arrive at a sales counter at an average rate of 20 customers per hour. The arrival rate is random
and may be assumed to follow a Poisson distribution. The time required to serve each customer has an
exponential distribution with a mean of 2 minutes.

Required:

a) The average waiting time in the queue

b) The average waiting time in the system

c) The average number of customers in the queue

d) The average number of customers in the system

e) Probability that there are no customers in the system

Exercises

Question 1

The arrival rate of customers in a barber shop follows a Poisson probability distribution with a mean of 5
customers per hour. It takes the barber an average of 8 minutes to do a haircut, with the average time
following an exponential distribution.

Required:

a) The average number of customers in the system

b) The average waiting time in the system

c) Compute the probability of having 3 customers in the queue

Question 2

Aero planes arrive at Jomo Kenyatta International Airport at an average rate of 20 planes per hour. The
arrival is random and may be assumed to follow a Poisson distribution. The rate at which the planes are

119
given permission to land depends on the weather, with a mean landing rate of 40 planes per hour in good
weather and 25 planes per hour in bad weather.

Required:

Under each of the weather conditions, determine

a) The average number of planes circling the airport

b) The average number of planes in the system: i.e., in the process of landing or circling the airport

c) The average waiting time in the system

d) The average waiting time in the queue

e) The time a plane spends in the ‘landing’ process itself

Simulation

Illustration

A small garage was opened along the Mombasa Malindi road. An observation was made of the number of
vehicles that stopped by for refueling. You are provided with the following information on the statistics for
the first 50 days of operation

No of vehicles refueling 5 10 15 20 25 30

No of days: 2 6 8 16 12 6

The number of litres of petrol filled into the cars was observed over the same period and is summarized
below:

Qty of fuel (litres/car) 10 15 20 25 30 35 40

No of days 4 6 16 10 6 6 2

Petrol is sold at a price of Shs 75 per litre, with a variable cost of Shs 48 per litre. The company incurs
variable costs of Shs 50 per vehicle refueling at the station, and the fixed costs per week amount to Shs 15
000.

120
Required

Simulate the process over the next one week and determine the profit / loss based on the following random
sequence of numbers for the number of vehicles and then the quantity of fuel respectively:
1749820645786362965338980432

Exercise

In a Quantitative methods class, lateness has been observed for the last few days. You are provided with the
following data:

No. of late students: 0 1 2 3 4 5 6

Probability: 0.10.13 0.18 0.20 0.25 0.07 0.04

Required

Simulate the lateness of the students for the next 12 days and compute the average number of late students
based on the simulated data. Use the following set of random numbers [278145632786940539513771]

121