DEPARTMENT OF ACCOUNTING

UNIVERSITY OF JAFFNA- SRI LANKA

Program Title : Bachelor of Business Administration Honours in Accounting -2021/2022

Course Unit : BBAA 31013- Quantitative Methods for Business
Handout : 02 – Linear Programming I
Prepared by : Dr. V.Sritharan
Issued on : 9th November, 2023
Intended Learning Outcomes:
After completing this lesson, you should be able to:
 define objective function
 formulate the Linear programming (LP) problem
 construct a graph and plot the constraint lines
 identify feasible solution region
 use iso-profit/iso-cost line for solution
 find the optimum point.
 Evaluate the objective function at optimum points to get the required maximum/
minimum value of the objective function
 define shadow price
Introduction
All organizations have to make decisions on how to allocate their scarce resources. Managers
in these organizations are faced with many problems relating to the use of limited resource.
Linear Programming is a mathematical technique for finding the best uses of an organization’s
resource. That is linear programming typically deals with the problem of allocating limited
resources in a way that maximize return or minimizes cost.

Requirements for Linear programming

Linear programming is a technique to optimize the value of some objective when the factors
involved are subject to some constraint or limitation.

Thus linear programming can be used to solve problems which have the following
characteristics.
i) Can be stated in numeric terms.
ii) All factors have linear relationships.
Eg: - If one unit requires 5 man hours, 2 units require 10 man hours and so on. It
must be stressed that every relationship in the problem must be linear or able to
be linearly approximated.
iii) The problem must permit a choice or choices between alternative courses of action.
iv) There must be one or more restrictions on the availability of resources.
Eg: - Only 5000 machine hours are available per week.
v) All decision variables in a linear programming problem must assume non-negative
values.
Eg:- If the decision variables are x and y then we have x  0 , y  0 and these are
referred to as non negativity constraints.

Formulation of a linear programming model

Steps for the formulation of the model can be stated as follows.
i) Identify the decision variables.
ii) Identify the coefficients related to objective function.
iii) Construct the objective function.
iv) Identify the resource limitations.
v) Relate resource limitations to resource utilization.
vi) Find the constraints for the model
BBAA 31013- Quantitative Methods for Business page 1
The Assumptions of Linear programming
1) Linearity:
A primary mathematical assumption of the linear programming model is that of linearity. That
is we assume that all functional relationships between decision variables both in the objective
function and in the constraints to be linear in form. A linear relationship involving n decision
variables x1,x2…………xn is an expression of the form

c1x1 + c2x2 + ………………… + cnxn

where c1,c2,………., cn are real numbers

2) Additively:
The value of the objective function for the given values of decision variable must be equal to
the sum of the contributions (profit or cost) from each decision variable. For example, total
profit earned by the sale of two products A and B must be equal to sum of the profits earned
separately from A & B. Similarly, the total of resources used must be equal to the sum of
resources used by each decision variable. For example, if one unit of product A requires 4
labour hours and 16kg of material and one unit of product B requires 2 labour hours and 9kg
of material then one unit of A and one unit of B will require 6 labour hours and 25kg of
material.

3) Divisibility (or continuity):

The linear programming model assumes that each decision variable is a continuous variable
which can take any value within a certain range. This implies that all activity levels and all
resource usages are completely divisible. Therefore, decision variables can take fractional
values such as 87.5 units of a product, 2/3 units of an item etc.

Solution of linear programming problem

There are two main methods of solving a linear programming problem namely.
i) Graphical method
ii) Simplex method

Graphical method
Graphical analysis in linear programming is restricted to problems with two variables. When
there are only two variables, the graphical method involves in graphing the system of
equations and inequalities.

Graphical solution to maximization problems

A small company produces two types of toy cars, model A and model B. Each car requires in
the manufacture the use of three machines M1, M2 and M3. To produce a model A car it
requires the use of M1 for 2 hours, M2 for 1 hour and M3 for 1 hour and to produce a model
B car it requires 1 hour on M1, 2 hours on M2 and 1 hour on M3. The maximum number of
machine hours available per month for M1, M2, and M3 are respectively 180,160 and 100.
The company can make a profit of Rs. 80 on a model A car and a profit of Rs. 120 on a model
B car. Further the company can sell all the cars it can produce.

Formulate a linear programming model, graph the constraints of the model and show the
feasible region. Determine the maximum possible monthly profit that can be expected.

We can summarize the above data in to a table as follows

BBAA 31013- Quantitative Methods for Business page 2

Machine Model A Model B Available hours
hours/ unit hours/ unit

M1 2 1 180
M2 1 2 160
M3 1 1 100
Profit/unit Rs. 80 Rs.120

Formulation of the linear programming model.

x- number of model A cars produced per month
y- number of model B cars produced per month

The objective function Z = 80x + 120y

Which is the total profit per month.
Where
80x = Profit from x cars of model A
120y = Profit from y cars of model B
The constraint set
for resource M1
2x + y  180
for resource M2
x + 2y  160
for resource M3
x + y  100
for non negativity conditions.
x0
y0
The complete linear programming model for this problem can now be expressed as.

Maximize Z = 80x + 120y

Subject to : 2x + y  180
x + 2y  160
x + y  100
and x,y  0

Graphing the constrains

1. 2x + y  180
Consider 2x + y = 180 and graph this line
x 0 90
y 180 0
2. x + 2y  160
Consider x + 2y = 160 and draw its graph
x 0 160
y 80 0
3. x + y  100
Consider x + y = 100 and draw the line
x 0 100
y 100 0

4. x  0, y  0

BBAA 31013- Quantitative Methods for Business page 3

Iso-profit line
An Iso-profit line is a line each of whose points designates a solution with the same objective
function value. Iso-profit lines can be used to determine the maximum profit strategy in a
linear programming maximization model.
Now, using the objective function we can write
120y = - 80x + Z
y = - 2/3 x + 1/ 120 Z
The slope of this line is - 2/3 and the intercept on the y axis is 1/120 Z . Since the slope is
independent of Z, for different profit values of Z we get parallel profit line.
When we consider Z = 4800 and Z = 7200
We have the profit lines as follows.
for Z = 4800, y = - 2/3 x + 40 …………………(1)
for Z = 7200, y = - 2/3 x + 60 ………………….(2)

These profit lines which are parallel to each other are called Iso- profit lines.
We observe that as Z increases the Iso-profit lines are moving further away from the origin
the object of the problem is to be on the highest attainable Iso profit line. Clearly the optimal
solution can not be in the interior of the feasible region. For any Iso-profit line passing
through interior points of the feasible region, there is another Iso-profit line close by that has
higher Z value. Therefore, the optimal solution must be at a corner point.

This property holds true for any linear programming problem. Therefore, a simple method is
to consider evaluating all corner point solution and comparing them to find the best.

Corner point method for finding the optimal solution

In the feasible region show in this example there are five points O, A, B, C, D on the
boundary of the feasible region. These points are called vertices of corner points of the
feasible region.
In our problem, there are five corner points which are O = ( 0,0 ), A = ( 0,80 ), B = ( 40,60 ),
C = ( 80,20 ), D = ( 90,0)
We can now evaluate the objective function value at these points.
Objective function Z = 80x + 120y
at O = ( 0,0 ) Z0 = 80(0) + 120(0) =0
at A = ( 0,80 ) ZA = 80(0) + 120(80) = 9,600
at B = ( 40,60 ) ZB = 80(40) + 120(60) = 10,400
at C = ( 80,20 ) ZC = 80(80) + 120(20) = 8,800
at D = ( 90,0 ) ZD = 80(90) + 120(0) = 7,200
Thus Z has a maximum value of 10,400 at B where x = 40 and y = 60.
BBAA 31013- Quantitative Methods for Business page 4
Graphical solution for minimization problem
Provided that they only have two unknowns minimization problems can also be dealt with by
graphical means. The general approach of drawing the axes with appropriate scales and
inserting lines representing the limitations is the same as for maximizing problems but the
following differences between maximizing and minimizing problems will be found.Normally
in a minimizing problem the limitations are of the grater than or equal to type (  ). So that,
the feasible region will be above the limitation lines (all or most of the lines).
Sensitivity Analysis of Graphical model
An LP model is a snapshot of a real situation in which the model parameters (objective and
constraint coefficients) assume static values. to enhance the applicability of Lp in practice, we
need to add a dynamic dimension that investigates the impact of making changes in the model
parameters (objective and constraint coefficients ) on the optimal solution. The procedure is
referred as sensitivity analysis because it studies the sensitivity of the optimal solution to
changes made in the model.

Here two cases of sensitivity analysis based on the graphical LP solution:

1) Changes in the objective coefficients, and
2) Changes in the right-hand side of the constraints.

Changes in the objective function’s coefficients

The general objective function in a two variable LP problem can be written as maximize or
minimize Z = C1X1 + C2X2

Changes in the coefficients C1 and C2 will change the slope of Z and, hence, possibly, the
optimal corner point.

Change in availability of resources.

In LP model, constraints, directly or indirectly, represent the usage of limited resources. In
this case, the right–hand side can be thought of as representing limits on the availability of the
resource. This section investigates the sensitivity of the optimum solution to making changes
in the amount of available resources.

Multiple solutions
A linear programming problem may have more than one optimal solution, as illustrated in the
following example.

Example:-
Consider the following LP problem
Maximize Z= 2.5x + 4.5y
subject to
x7
y3
5x + 9y  45
x, y  0

Redundant constraints
A particular constraint is said to be redundant if that constraint can be omitted without
changing the set of feasible solution to the LP problem

Consider the constraints

x + y  8 …………….. (1)
2x + 3y  12 ……………..(2) with x, y  0
BBAA 31013- Quantitative Methods for Business page 5