CHAPTER 4

LINEAR PROGRAMMING: THE GRAPHICAL METHOD

After working through this chapter you should be able to:

 Solve minimisation and maximisation linear programming problems
using the graphical method (Corner Point Method).
 Calculate the Range of optimality to see by how much the profit
contribution is allowed to fluctuate.

4.1 INTRODUCTION
Linear Programming is a widely used mathematical technique designed to help
managers in planning and decision making relative to resource allocation.

Requirements of a LP problem:
 All problems seek to minimise or maximise some quantity
 All problems have restrictions or constraints that limit the degree to which we
can pursue our goal
 There must be alternative courses of action to choose from
 The objective and constraints in LP problems must be expressed in terms of
linear equations or inequalities

THE GRAPHICAL METHOD (USING THE CORNER POINT METHOD):

In view of the current economic situation worldwide, much attention has been given in
recent years to the execution and completion of tasks in the most efficient and
economical way possible. In this section, we shall be interested in maximising and
minimising linear expressions of the form

b1 x1 + b2 x2 + b3 x3 + … + bn xn

where the b’s are the coefficients (constants) of the x’s and the x’s are variables. The
corner point method, one of the methods for the optimisation (maximisation and
minimisation) of the linear function subject to certain constraints will be studied in this
course.
 In Linear Programming there exists theoretical backing to the statement that an
optimal solution will lie at a corner point of the feasible region, where the feasible
region is the region where all constraints are satisfied. This holds true to any LP
problem.
The equation to be optimised will from now on be known as the objective function.

The goal of this section involves the scheduling of all tasks to obtain the optimal
production quantities and in so doing, optimising the objective function.

This procedure is useful only when there are two decision variables.

4.2 OBJECTIVE: MAXIMISE THE PROFIT BY USING THE CP METHOD

Example 1
Cellphones4U owns two small factories where they manufacture two types of cellular phones,
namely Habari and Mzuri. Each Habari cellular phone requires 10 manhours to manufacture
and 3 manhours in the testing department. Each Mzuri cellular phone requires 8 manhours to
manufacture and 4 manhours in the testing department. The manufacturing department has a
maximum of 800 manhours available per week, while the testing department has a maximum
of 360 manhours available per week. The factory has to manufacture at least 60 of the Mzuri
models each week. The profit on one Habari phone is R200 while the profit on one Mzuri phone
is R270.

How many of each type of phone must the company manufacture to maximise the profit?

The following steps need to be considered when setting up an LP formulation:

Step 1: Define x1 and x2: (x and y could also be used)
x1 = the number of Habari phones manufactured and
x2 = the number of Mzuri phones manufactured.

Step 2: Determine the objective function:

To maximise profit, P = 200x1 + 270x2
Step 3: Identify and set up constraints:
The following constraints therefore exist for the given day’s production:
x1 , x2  0
10 x1  8 x2  800
3x1  4 x2  360
x2  60
Note:
It is clear that a negative number of either models cannot be manufactured and
therefore both x1 and x2 can also assume values which are either positive or zero.

Step 4: Plot the constraints graphically as follows:

x1 , x2  0 Implies that the values for x1 and x2 cannot be negative, so we are sketching

our graph in the first quadrant of the Cartesian plane.

10 x1  8x2  800

Firstly, consider the inequality as an equality, and sketch the straight line:
10 x1  8x2  800
x1 intercept: Let x2 = 0, then x1 = 80 Point (80; 0)
x2 intercept: Let x1 = 0, then x2 = 100 Point (0; 100)

120 x2

100

80

60

40

20

x1

20 40 60 80 100 120

Secondly, indicate the region that satisfies the inequality:

10 x1  8x2  800 , hence the region satisfying the inequality is either above or below the
line. To confirm which of the two it is, substitute a point below the line into the
inequality. If the point satisfies the inequality, then the region satisfying the inequality
is below the line, else it is above the line. The easiest point to substitute is (0, 0).

10(0) + 8(0) = 0  800, therefore the graphical representation of the first constraint is:

120 x2

100

80

60

40

20

x1

20 40 60 80 100 120

Similarly, the second constraint can be represented on the same plane:

x1 , x2  0
10 x1  8 x2  800
3x1  4 x2  360

120 x2

100

80

60

40
Finally the last inequality can be added:
x1 , x2  0
10 x1  8 x2  800
3x1  4 x2  360
x2  60

120 x2

100

80

60

40

20

x1
20 40 60 80 100 120

Step 5: Identify the feasible region

The feasible region must satisfy all the constraints:

120 x2

100

80

60

40

20

x1
Step 6: Identify the corner points of the feasible region and calculate the profit
associated to each corner point:

120 x2

100
(b)
(c)
80

60
(a) (d)

40

20

x1
20 40 60 80 100 120 Therefore it can be
Corner points: concluded that profit is
(a): (0; 60): P = 200x1 + 270x2 = 0 + 270 (60) = 16 200 maximized at point (b), hence
(b): (0; 90): P = 200x1 + 270x2 = 0 + 270 (90) = 24 300 0 units of Habari and 90 units
(c): (20; 75): P = 200x1 + 270x2 = 200(20) + 270 (75) = 24 250 of Mzuri needs to be
(d): (32; 60): P = 200x1 + 270x2 = 200(32) + 270 (60) = 22 600 manufactured to maximize
profit.
Note: The coordinates of point (c) are found by setting the equations for the two lines
that cross at (c) equal to each other:
10 x1  8x2  800 and 3x1 + 4x2 = 360 (multiply the second equation

by 2):
10 x1  8x2  800
6x1 + 8x2 = 720 Subtract the second equation from the first:
4 x1  80
Substitute into one of the original equations:
 x1  20

10(20)  8x2  800

 x2 = 75

The coordinates of point (d) are found by setting the equations for the two lines that
cross at (d) equal to each other:
 x2 = 60 and 10 x1  8x2  800  10 x1  8(60)  800

 x1 = 32

Example 2
The Bradlows Furniture Company

The Bradlows furniture company produces inexpensive tables and chairs. The production
process for each is similar in that both require a certain number of hours of carpentry work and
a certain number of labour hours in the painting and varnishing department. Each table takes
four hours of carpentry work and two hours in the painting and varnishing shop. Each chair
requires three hours in the carpentry and one hour in painting and varnishing. During the current
production period, 240 hours of carpentry and 100 hours of painting and varnishing time are
available. Each table sold yields a profit of R7; each chair produced may be sold for a profit of
R5. This company would like to maximize their profits.
 4.3 OBJECTIVE: MINIMISE THE COST BY USING THE CP METHOD

Example 3
A travel agency has to transport a minimum of 1 200 passengers and 36 000kg of luggage
from one airport to another. Two types of aircraft are available. The SilverJet can transport a
maximum of 120 passengers and 2 000kg of luggage in one flight. The GoldenFlyer can
transport a maximum of 60 people and 3 000kg of luggage in one flight. The travel agency
may not use more than 16 aircraft. The rental of a SilverJet is R40 000 per flight and R48 000
per flight on a GoldenFlyer.

How many of each aircraft must the travel agency hire in order to mimimise cost?

Step 1: Define x1 and x2 :

Let x1 be the number of SilverJet aircraft
Let x2 be the number of GoldenFlyer aircraft
Step 2: Determine the objective function of the problem:
Minimise C  40000 x1  48000x2
Step 3: Identify and set up constraints:
x1 , x2  0
120 x1  60 x2  1200
2000 x1  3000 x2  36000
x1  x2  16
Step 4: Plot the constraints graphically as follows:
x1 , x2  0 Implies that the values for x1 and x2 cannot be negative, so we are sketching

our graph in the first quadrant of the Cartesian plane.

120 x1  60 x2  1200 : Intercepts (0; 20) and (10; 0)

2000 x1  3000 x2  36000 : Intercepts (0; 12) and (18; 0)

x1  x2  16 : Intercepts (0; 16) and (16; 0)

20 x2

15

10

x1

5 10 15 20
Step 6: Identify the feasible region

20 x2

15

(a)

10

(b)

5 (c)

x1

5 10 15 20

Step 7: Identify the corner points of the feasible region and calculate the profit
associated to each corner point:
(a): 120 x1  60 x2  1200 -------------(1)

and x1  x2  16 -------------(2)

Multiply eq (2) by 60: 60 x1  60 x2  960 ------------(3)

Subtract eq (3) from eq (1): 60 x1  240  x1  4

Substitute x1 = 4 into any of the equations above, then x2 = 12
Coordinates of point (a) is (4; 12)

(b): 120 x1  60 x2  1200 -------------(1)

and 2000 x1  3000 x2  36000 -------------(2)

Multiply eq (1) by 50: 6000 x1  3000 x2  60000 ------------(3)

Subtract eq (2) from eq (3): 4000 x1  24000  x1  6

Substitute x1 = 6 into any of the equations above, then x2 = 8
Coordinates of point (b) is (6; 8)
(c): 2000 x1  3000 x2  36000 ------------(1)

and x1  x2  16 ------------(2)

Multiply eq (2) by 2000: 2000 x1  2000 x2  32000 ------------(3)

Subtract eq (3) from eq (1): 1000 x2  4000  x2  4

Substitute x2 =4 into any of the equations above, then x1 = 12
Coordinates of point (c) is (12; 4)

Corner Points:
(4; 12): C = 40000 x1  48000 x2 = 40 000(4) + 48 000(12) = 736 000

(6; 8): C = 40000 x1  48000 x2 = 40 000(6) + 48 000(8) =624 000

(12; 4): C = 40000 x1  48000 x2 = 40 000(12) + 48 000(4) = 672 000

Therefore it can be concluded that cost is minimised when hiring 6 SilverJet aircraft
and 8 GoldenFlyers.

Example 4
Ace Auto manufacturers

Ace Auto manufactures luxury cars. The company believes it is most likely that their
customers are high – income people. In order to reach these groups, Ace Auto has embarked
on an ambitious tv advertising campaign and has decided to purchase one minute commercial
spots on two types of programs; comedy shows and cricket games shows. Seven million high
– income women and two million high - income men see each comedy commercial. Two
million high – income women and twelve million high – income men see each cricket
commercial. A one - minute comedy ad costs R50 000 and a one - minute cricket ad costs
R100 000. Ace Auto would like the commercials to be seen by at least 28 million high – income
women and at least 24 million high – income men, but at a minimum cost. How many
commercial spots must Ace Auto purchase on comedy shows and cricket games to minimize
cost?
EXERCISES ON SECTIONS 4.1 –4.3

1. A clothing designer has been invited to a fashion show where a variety of

suits and dresses will be shown. She will receive R80 for every suit shown and R60
for every dress shown. Designing a suit takes her 2 hours, while designing a dress
takes 4 hours. It will take her seamstress 4 hours to make a suit and 1 hour to make
a dress. She has 160 working hours available before the show, while her seamstress
has 180 working hours available. How many suits and how many dresses should the
designer make for the show to maximize her income?

2. A farmer wishes to minimize the cost of feeding his 50 camels weekly. There
are two types of feed available, Type A and Type B. Type A costs 60c per kg while
Type B costs 100c per kg. Each camel must receive at least 100 units of energy weekly
and at least 120 units of protein weekly. Each kg of Type A contains 2 units of energy
and 1 unit of protein, while Type B contains 1 unit of energy and 3 units of protein. Use
the graphical method to find the most affordable combination of the two types of feed
that the farmer should purchase.

Formulate the following linear programming problem (do not solve):

3. A manufacturer makes two products. The first product requires five hours for
processing, three hours for assembling and four hours for packaging. The
second product requires two hours for processing, twelve hours for
assembling and eight hours for packaging. The plant has forty hours available
for processing, sixty for assembling and forty eight for packaging. The profit
margin for the first product is R7 and R21 for the second product.

4. A manufacturer of kitchen units makes two types of units, A and B. It takes 2

days to assemble one unit of type A and 3 31 days to assemble one unit of

type B. This is done in a workshop that is available for only 20 days each
month. The paint shop can handle a maximum of 8 units per month. At least 2
units of type A must be produced each month.
 Units of type A yield a profit of R700 per unit while units of type B yield a profit
of R900 per unit. Maximise the profit.

Solve the following LP problems, using the corner-point method

5. Maximize profit: P  10x1  2x 2

Subject to:
2x 1  4x 2  160
4x 1  x 2  180
x1, x2 ≥ 0

6. Minimize cost: C  5x1  18x 2

Subject to:
2x 1  x 2  100
x 1  3x 2  120
x1, x2 ≥ 0

7. Maximize profit: P  12 x  10 y
Subject to:
x  3 y  15
3 x  y  13
y 1
x0

8. Minimize: z  40 x  60 y
Subject to:
x y 5
x  2y  4
x, y  0

9. Maximise profit, P = 4 x  5 y
Subject to:
 4 x1  8 x2  40
3 x1  x2  20
x1  5
x1 , x2  0

10. Minimise cost, C = 7 x1  3x2

Subject to:
4 x1  2 x2  28
x1  6 x2  29
x1  10
x2  0

11. Maximise profit, P = 20 x  15 y

Subject to:
3 x  6 y  96
4 x  2 y  56
y  15
x0
4.5 SENSITIVITY ANALYSIS

SA is a study of how the changes in the coefficients of a linear program affect the
optimal solution.
SA helps us determine when:
 A change in the coefficient of the objective function (profit / cost contribution)
and
 A change in the right hand side (RHS) value of a constraint (amount of hours
allocated)
will affect the optimal solution.

Introduction to Sensitivity Analysis

Because we live in an ever-changing environment, SA is an important tool in the life
of the decision maker in the real world. Prices of raw materials change, product
demand changes and stock prices change, to mention a few possible uncertainties. If
an LP model has been used in this environment, we can expect some of the
coefficients of the objective function to change over time. If so, then we would want to
determine how these changes would affect the optimal solution.

In example 2 on page 83, the corner points were :(0; 0), (0; 80), (30; 40), (50; 0) and
the profit function was P  7 x1  5x2 and the optimal solution found was (30; 40).

What would the optimal solution have been if the profit function was P  6 x1  4 x2 ?

ANSWER: (30; 40)

What would the optimal solution have been if the profit function was P  7 x1  7 x2 ?

ANSWER: (0; 80)

Hence with the profit function changing, the optimal solution could change.

The Range of optimality allows us to see by how much the profit contribution is
allowed to fluctuate, all else being constant and yet have the current production
quantities remain optimal.
Example 5
Consider example 2 in Chapter 4, page 83.
The Bradlows furniture company produces inexpensive tables and chairs. The production process for each is similar in that both require a
certain number of hours of carpentry work and a certain number of labour hours in the painting and varnishing department. Each table takes
four hours of carpentry work and two hours in the painting and varnishing shop. Each chair requires three hours in the carpentry and one hour
in painting and varnishing. During the current production period, 240 hours of carpentry and 100 hours of painting and varnishing time are
available. Each table sold yields a profit of R7; each chair produced may be sold for a profit of R5. This company would like to maximize their
profits.

The optimal solution found was (30; 40), meaning that the company should produce 30 tables
and 40 chairs.
By how much can the profit associated to the “tables” change and yet have the current
production quantities remain optimal?

Let us compute the range of optimality (ROO) for the Tables.

By how much can the profit associated to the “tables” change and yet have the current
production quantities remain optimal. The current production quantities are x1 = 30
and x2 = 40.

Graphical solution with slope of objective function

100 x2

90 2x1 + x2 = 100

80

70

60

50
Optimum
40

30
4x1 + 3x2 = 240
20

10
x1

10 20 30 40 50 60 70
A careful inspection of this graph shows that as long as the slope of the objective
function is between the slope of line 2x1 + x2 = 100 and the slope of line 4x1 + 3x2 =
240 the optimal point will remain constant. The intersection of the two constraints
that produce the optimal corner point serves as the upper and lower limits for the
range of optimality. At this optimal corner point, x1 = 30 and x2 = 40.

Changing an objective coefficient for x1 or x2 will cause the slope of the objective
function to change. In the figure we see that such changes cause the objective
function line to rotate around the optimal corner point.

Rotating the objective function line anti clockwise causes the slope to become less
negative, and the slope increases. The objective function line should be rotated anti
clockwise (slope increased) enough to coincide with line 4x1 + 3x2 = 240.

Any further anti clockwise rotation of the objective function line will cause the current
optimal corner point to be non-optimal. Hence the slope of line 4x1 + 3x2 = 240, which
4
is  , provides an upper limit for the slope of the objective function line.
3

Rotating the objective function line clockwise causes the slope to become more
negative, and the slope decreases. Any further clockwise rotation of the objective
function line will cause the current optimal corner point to be non-optimal. Hence the
slope of line 2x1 + x2 = 100, which is – 2 , provides a lower limit for the slope
of the objective function line.

Thus, the current optimal corner point will remain optimal solution as long as

4
– 2  slope of the objective function line  
3
Note: To find the slope of the line, make x2 the subject of the equation, then the slope
is the coefficient of x1.
4 240 4
4x1 + 3x2 = 240  x2   x1  . Thus the slope is 
3 3 3
2x1 + x2 = 100  x2  2 x1  100 . Thus the slope is – 2

The slope of the objective function line

Let us now consider the general form of the slope of the objective function.
Let:
P= the value of the objective function
A= the profit on tables
B = the profit on chairs

Using this notation, the objective function line can be written as

P = A x1 + Bx2

Writing this equation in slope-intercept form, we obtain

B x 2 = - A x1 + P
and
x 2 = (- A/ B) x1 + P / D

Thus we see that the slope of the objective function is given by – A / B.

4
– 2  - A/ B  
3

To compute the range of optimality (ROO) for the “tables” profit contribution, we
hold the profit contribution for the “chairs” fixed at its initial value B = 5.
4
– 2  - A/ 5  
3
From the left hand inequality, we have
– 2  - A/ 5
multiplying both sides by 5: – 10  - A
so, A  10

From the right hand inequality, we have

 4
- A/ 5  
3
multiplying both sides by 5: – A  – 6.67
so, A ≥ 6.67

Combining the calculated limits for A provides the following range of optimality for
the “tables” profit contribution:
6.67  A  10

The range of optimality for A tells management that, with other coefficients
unchanged, the profit contribution for the “tables” can be anywhere between
R6.67 and R10 and the production quantities of 30 tables and 40 chairs will
remain optimal. Note, however, that while the production quantities will not change,
the total profit contribution (value of objective function) will change due to the change
in profit contribution per table.

These computations can be repeated, holding the profit contribution for tables constant
at A = 7. The range of optimality for the “chairs” profit contribution can also then be
determined.

Example 6
Consider example 2 in Chapter 4, page 83.
Calculate the range of optimality for the “chairs”.

Special Cases
In cases where the rotation of the objective function line about an optimal extreme
point causes the objective function line to become vertical or horizontal, there will be
either no upper limit or no lower limit for the slope.

Consider example 2 in Chapter 4, page 83:

What would the optimal solution have been if the profit function was P  7 x1  7 x2 ?

ANSWER: (0; 80)

In this case, the optimal point does not lie where the two lines cross, hence there will
only be one limit (upper or lower) of the objective function when calculating the ROO
of either of the profit contributions.

We will not be considering these cases.

4.6 EXERCISES: SECION 4.5

1. (See exercise 1 on graphical method) A clothing designer has been invited to

a fashion show where a variety of suits and dresses will be shown. She will
receive R80 for every suit shown and R60 for every dress shown. Designing a
suit takes her 2 hours, while designing a dress takes 4 hours. It will take her
seamstress 4 hours to make a suit and 1 hour to make a dress. She has 160
working hours available before the show, while her seamstress has 180 working
hours available.

In order to maximize her income, she needs to make 40 suits and 20 dresses.
a) Find the ROO for the suits’ contribution to the profit.
b) Find the ROO for the dresses’ contribution to the profit.

2. In order to maximize profit, P  12x1  8x 2 , subject to the constraints:

2x 1  4x 2  80
6x 1  2x 2  90
x1, x2 ≥ 0
it was found that profit is maximised at the point (10; 15).
Calculate and determine the range of optimality for x1 profit contribution.

3. In order to maximize profit, P= 100x1 + 50x2, subject to the constraints:

2x1 + 3x2 ≤ 60
4x1 + x2 ≤ 40
x1, x2 ≥ 0
it was found that profit is maximised at the point (6; 16).
Calculate and determine the range of optimality for x2 profit contribution.
4. In order to maximize profit, P = 200x1 + 150x2 , subject to the constraints:
x1 + 4x2 ≤ 240
2x1 + x2 ≤ 200
x1, x2 ≥ 0
it was found that profit is maximised at the point (80; 40).
Calculate and determine the range of optimality for x1 profit contribution.