Chapter 5

Linear programming

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Learning objectives

After studying this chapter, you should be able to:

• Describe situations where the use of linear programming may be
appropriate
• Explain when the graphical method of linear programming can be
used
• Explain how graphical linear programming can be used to find
optimum output
levels for short-term product mix decisions
• Formulate a linear programming model using the simplex method
• Explain the term ‘shadow prices’ (or ‘opportunity costs’)
• Explain different uses for linear programming
• Explain the assumptions and limitations of linear programming.
Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Constraints
• Single resource constraints
• Chapter 4 covered basic limiting factor analysis to
determine the profit-maximising sales mix whereby a
single resource constraint is present.
• Multiple resource constraints
• When there is more than one scarce resource, it is
not quite so easy to establish the optimal usage plan.
The technique mentioned above cannot be applied
when there is more than one limiting factor. In such
a situation, linear programming is an effective tool to
determine the most efficient use of scarce resources.
Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Linear programming

• a mathematical technique
• applied where there is more than one constraint
• to find the best possible way to allocate scarce resources,
e.g. energy, machinery, materials, funds, labour
• to maximise profit or minimise cost

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
Optimal manufacturing plan

The manufacturing plan that would result in:

• the highest possible profit
• lowest possible cost
• making the best use of scarce resources

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Methods to do linear
programming
• the graphical method (only for problems
involving two products)
• the simplex method

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Linear programming
assumptions
• Relationships between variables are linear and there is an
optimal solution.
• The contribution per unit for each product and the usage of
resources per unit are the same, regardless of the quantity of
output manufactured and sold.
• The products and resources that are allocated are infinitely
divisible.
• There is one clearly defined objective function.
• Single-value estimates are used for the uncertain values.
• The situation remains static in all other respects.

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Illustrative example 5.1
Maskia (Pty) Ltd currently manufactures two types of specialised
machinery parts. The standard costs of the products are as follows:

Xabex Yowek

Selling price per unit R220 R236

Variable costs: R192 R204
Material (16 units @ R4 per unit); (8 units @ R4 per unit) R64 R32
Labour (12 hours @ R10 per hour); (16 hours @ R10 per hour) R120 R160
Variable overheads (8 machine hours @ R1 per hour); (12 hours
@ R1 per hour) R8 R12
Contribution R28 R32

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Illustrative example
(cont.)
During the next financial period, the availability of resources is expected
to be as follows:
• Material 3 360 units
• Labour 5 760 labour hours
• Machine capacity 2 760 machine hours
The owner of the organisation believes that the maximum sales potential
for the Xabex is 200 units, while there is no expected sales limitation on
the Yowek.

Required:
Give advice on the optimal manufacturing plan for Maskia (Pty) Ltd to
make best use of scarce resources, obtain the highest possible profit and
incur the lowest cost.
Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Illustrative example
(cont.)
Solution
Step 1:
• We use linear programming to find linear relationships between variables or factors
that affect decisions. As a first step we therefore need to identify decision variables.
• In our example it will be the following:
X = number of Xabex units to manufacture
Y = number of Yowek units to manufacture
• Now we can determine the equations that will explain the problem in algebraic
terms.
• Since this is the graphical method, it is necessary to decide which product will be
denoted on the x-axis and which product on the y-axis. In this example, we can put
Xabex on the x-axis and Yowek on the y-axis. We are also going to denote the
products as such in the algebraic equations.

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Illustrative example
(cont.)
Step 2:
• We can assume the goal of Maskia is to maximise its contribution. When
we determine the objective function, we need to take this into account.
• The Xabex gives a contribution of R28 per unit and the Yowek a
contribution of R32 per unit. If the objective of the company is to
maximise its contribution, the objective function would be stated as:
C ≥ 28X + 32Y.

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Illustrative example
(cont.) .
Step 3:
• After establishing the objective of the company by means of the objective
function,we can pay attention to possible constraints the company is facing. The
constraints are presented in the form of algebraic equations of the limitations in
resources the company faces and which limits its output.
• Maskia faces possible resource constraints in the form of available material, labour
hours and machine capacity. The availability of resources is given as a maximum of 3
360 units of material, 5 760 labour hours and 2 760 machine hours.
• Each Xabex uses 16 units of material and each Yowek uses 8 units of material. Each
Xabex uses 12 labour hours and each Yowek uses 16 labour hours. Each Xabex uses 8
machine hours and each Yowek uses 12 machine hours. The constraints can be stated
as follows:
Material: 16X + 8Y ≤ 3 360
Labour: 12X + 16Y ≤ 5 760
Machine capacity: 8X + 12Y ≤ 2 760

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Illustrative example
(cont.)
• Linear programming as a mathematical tool does not promise that the answer will
necessarily make sense. For example, it may happen that an answer indicates that a
negative number of a certain product should be manufactured. To prevent this from
happening, a non-negativity factor must be included in the algebraic equations to
state that one cannot manufacture and sell less than zero of a product.
• This will be denoted as 0 ≤ X ≤ 200, since only 200 units of Xabex can be sold, and Y ≥
0 because demand for Y is unlimited.

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 The optimal solution
An optimal solution to a linear programming problem is a feasible solution
with the largest objective function value in the case of a maximisation
problem. A linear programme may have multiple optimal solutions, but only
one optimal solution value. A linear programme s unfeasible if it has no
feasible solutions, that is, the feasible region is empty. A linear programme is
unbounded if the optimal solution is unbounded, that is, it is either 1 or 21.
Every linear programme is therefore:
• Unfeasible, or
• Unbounded, or
• A unique optimal solution value.
.
Finding the optimal solution
The optimal solution can be determined using a graph and supplementing
the graph by solving simultaneous equations.

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Illustrative example
(cont.)
Step 4:
When the algebraic equations are determined, it can be presented on a
graph.
All the constraints need to be plotted on a single graph but for the sake of
clarity we are first going to observe each constraint on its own graph.

Taking the first constraint, which is materials, we need to determine the

number of units of X that can be manufactured from available materials if no
units of Y are manufactured, and also the number of units of Y that can be
manufactured from available materials if no units of X are manufactured. This
can be determined by substituting a zero into the equation for the product of
which we assume no units will be manufactured. This will look as follows:

In the equation 16X + 8Y ≤ 3 360, if Y = 0, then 16X + 8(0) ≤ 3 360.

This means the number of units of X if no units of Y are manufactured would
be: 3 360 ÷ 16 ≤ 210 units.

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Illustrative example
(cont.)
In the same equation 16X + 8Y ≤ 3 360, if X = 0, then 16(0) + 8Y ≤ 3 360.
This means the number of units of Y if no units of X are manufactured would be
3 360 ÷ 8 ≤ 420 units.
On a graph, it will be denoted as follows:

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Illustrative example
(cont.)
The same will have to be done for the other constraints. For labour, it means that
the maximum number of units that can be manufactured of each product will be:
12X + 16Y ≤ 5 760
If Y = 0, then 12X + 16(0) ≤ 5 760, therefore making X ≤ 480.
If X = 0, then 12(0) + 16Y ≤ 5 760, therefore making Y ≤ 360.
On the graph, it will be denoted as follows:

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Illustrative example
(cont.)
For machine capacity, the maximum number of units that can be manufactured of
each product will be:
8X + 12Y ≤ 2 760
If Y = 0, then 8X + 12(0) ≤ 2 760, therefore making X ≤ 345.
If X = 0, then 8(0) + 12Y ≤ 2 760, therefore making Y ≤ 230.
On the graph it will be denoted as follows:

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Illustrative example
(cont.)
In addition, we need to show non-negativity constraints. The Y = 0 is shown by
not allowing the graph to go into negative figures (stopping the graph at zero,
where the -axis and the y-axis meet). The 0 = X = 200 is shown by drawing a line
on the x-axis at 200 units to show that production cannot go past that amount.

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Illustrative example
(cont.)
The shaded area in Figure 5.4 (ABCDE) is the area within which optimal
manufacturing would take place. The contribution-maximising (optimal)
manufacturing plan would be as close as possible to a corner to the right of the
shaded area.

Step 5:
To find the optimal plan, choose a random contribution value that falls within the
shaded area. We do not know the maximum value of the objective function;
however, we can draw an iso-contribution (or ‘profit’) line that shows all the
combinations of X and Y that provide the same total value for the objective
function. You can pick any total contribution figure, but a multiple of 28 and 32 is
easiest. Hence, the best way to accomplish this is to choose a value that will be
divisible by 28 and 32 (the contributions from the equation C ≥ 28X + 32Y).

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Illustrative example
(cont.)

An easy way to determine a figure to use would be to take 28 (contribution of X)

multiplied by 32 (contribution of Y) and multiplied by 5 (fixed). Let us use 4 480.
Substituted into the equation, this means:

If Y = 0, 4 480 ≥ 28X + 32(0), then X ≥ 160, and

If X = 0, 4 480 ≥ 28(0) + 32Y, then Y ≥ 140.

The fixed ‘5’ we multiply the total with is to make the values of X and Y large
enough so that they can be easily indicated on the graph. If you multiplied with
10, the values for X and Y would have been larger than the range of values on
the graph.

This means that we can draw the iso-contribution line for the equation
C ≥ 28X + 32Y on the graph as shown by the dashed line in Figure 5.5.

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
Illustrative example
(cont.)

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Illustrative example
(cont.)
The dashed line means that any combination of X and Y on that line will give a contribution
of R4 480. This dashed line can be extended until it touches the last corner of the shaded
area, which in this case would be point B. This is the optimal solution to the problem. At
this point, X is approximately 145, and Y is 135. This gives a total contribution of:
C = 28(145) + 32(135) = R8 380.

Step 6
Since a small drawing error in a graph can provide an inexact solution to the number of
units at the optimal manufacturing plan, a more accurate way would be to solve the
simultaneous equations for the constraints that are applicable at the point of the optimal
manufacturing plan. We can see that the applicable constraints at the point of the optimal
plan are materials and machine capacity. The reason for this is that point B on the graph is
where the materials and machine capacity constraint lines cross. It is clear that labour
hours is not a constraint in the feasible area of this example.

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Illustrative example
(cont.)
Therefore:
8X + 12Y = 2 760 [1]
16X + 8Y = 3 360 [2]

Multiply equation [1] with a figure of 2 to derive equation [3]. The reason we multiply by 2
is to get one of the variables to have the same value. In this case, multiplying by 2 ensures
that both equations have a 16X in it, which can be cancelled out:
16X + 24Y = 5 520 [3]
16X + 8Y = 3 360 [2]

16X cancels each other out. Now subtract equation [2] from equation [3]:
16Y = 2 160
Y = 135

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Illustrative example
(cont.)
Substitute the answer into equation [3]:
16X + 24(135) = 5 520
16X + 2 280
X + 142.5
This gives an exact contribution of C = 28(142) + 32(135) = R8 296.

Remember that a company cannot produce half a unit, therefore the 142.5 will have to be
rounded down to a full unit. You will not round up, because that means an additional unit
is manufactured, which will not be possible with the limited resources available.

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Shadow prices

• If one can reduce a constraint by paying a premium, one

must determine what the maximum additional amount
is that can be paid.
• To answer that, one must determine how the optimal
solution would change if an additional unit of a scarce
resource is added.
• This change in the optimal manufacturing plan is called
the marginal rate of substitution.
• The monetary value of an additional unit of a scarce
resource is called the opportunity cost or the shadow
price.
Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Shadow prices

Calculating shadow prices

The simplest way to calculate shadow prices for a critical
constraint is as follows:
Step 1: Take the equations of the straight lines that intersect at
the optimal point. Add one unit to the constraint concerned,
while leaving the other critical constraint unchanged.
Step 2: Use simultaneous equations to derive a new optimal
solution.
Step 3: Calculate the revised optimal contribution. The increase
is the shadow price for the constraint under consideration.

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Shadow prices
Implications of shadow prices
Management can use shadow prices as a measure of the maximum
premium that they would be willing to pay for one more unit of the
scarce resource. However, the shadow price should be considered
carefully. For example, the shadow price of material may be calculated
as R100 per kg. However, it may be possible to negotiate a lower
shadow price than this.

In addition, if more of the critical constraint is obtained, the constraint

line will move outwards, altering the shape of the feasible region. After
a certain point, there will be little point in buying more of the scarce
resource since any non-critical constraints will become critical.

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Slack variables

• They are variables that are added to all the

equations in a linear programming model.
• They compensate for resources that may still
be unused at the optimum manufacturing
plan.
• E.g. 16X + 8Y +S1 = 3 360

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
The simplex method of
linear programming

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 The simplex method

• When more than two products are manufactured

from scarce resources, the graphical method is not
useful.
• The simplex method is an alternative.
• It also has the ability to provide information such as
marginal rates of substitution and opportunity cost.
• The simplex method is useful as it can consider
more complex problems involving more than two
output variables.

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Simplex matrices
• Equations are given in a matrix as follows:
Quantity X Y Description
S1 = 3 360 -16 -8 materials
constraint
S2 = 5 760 -12 -16 labour
constraint
S3 = 2 760 -8 -12 machine
capacity
constraint
S4 = 200 -1 0 demand
constraint
C=0 28 32 contribution

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Solving the matrix

• After the first matrix has been prepared, the

product with the highest contribution is first
considered.
• This results in a new set of equations to be
presented in a new matrix.
• This process is followed until a final matrix can
be prepared in which the optimal solution
appears.

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
 Uses for linear programming

• relevant cost calculations

• selling different products
• maximum payment for scarce resources
• control
• capital budgeting
• sensitivity analysis

Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
Linear programming limitations
Objective functions and constraints cannot always be
expressed in linear form; it is an assumption that one
makes.
• Fractional values are permitted for the decision
variables, but many problems require that the
decision variables should be obtained in integer
values.
• The coefficient of basic variables cannot be
determined with certainty, only with probability.
• Where a problem consists of multiple objectives,
linear programming cannot provide a solution.
Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229
Linear programming limitations
(cont.)

• Linear programming does not take into

account the effect of time and uncertainty.
• Parameters appearing in a linear
programming model are assumed to be
constant, but in real life situations they are
frequently neither known nor constant.
• In case of large, complex and constrained
problems, calculating the problems can be a
huge task.
Cost and Management Accounting: Operations and Management – A southern African approach (3rd edition)
© Juta and Company Ltd 2021
ISBN 9781485131229