Mathematical Modelling: Linear Programming

Optimization is an important and fascinating area of management science and operations research. It
helps to do less work, but gain more.

Linear programming (LP) is a central topic in optimization. It provides a powerful tool in modeling many
applications. LP has attracted most of its attention in optimization during the last six decades for two
main reasons:

• Applicability: There are many realworld applications that can be modeled as linear
programming;
• Solvability: There are theoretically and practically efficient techniques for solving large-scale
problems.

In this tutorial, the basic elements of an LP will be introduced and some examples that can be modeled
as an LP will be presented. In the next tutorials, we will discuss solution techniques.

Each optimization problem consists of three elements:

• parameters: describe the numerical values used in mathematical representation of

relationships;
• decision variables: describe our choices that are under our control;
• objective function: describes a criterion that we wish to minimize (e.g., cost) or maximize (e.g.,
profit);
• constraints: describe the limitations that restrict our choices for decision variables.

Formally, we use the term “linear programming (LP)” to refer to an optimization problem in which the
objective function is linear and each constraint is a linear inequality or equality.

Let’s start with an example. The example will be described first in words, and then it will be translated
into a linear program.

Example Problem Statement:

A company makes two products (say, P and Q) using two machines (say, A and B). Each unit of P that
is produced requires 50 minutes processing time on machine A and 30 minutes processing time on
machine B. Each unit of Q that is produced requires 24 minutes processing time on machine A and 33
minutes processing time on machine B.
Machine A is going to be available for 40 hours and machine B is available for 35 hours. The profit per
unit of P is $25 and the profit per unit of Q is $30. Company policy is to determine the production
quantity of each product in such a way as to maximize the total profit given that the available resources
should not be exceeded.

Task: The aim is to formulate the problem of deciding how much of each product to make in the current
week as an LP.

We often start with identifying decision variables (i.e., what we want to determine among those things
which are under our control).

The company wants to determine the optimal product to make in the current week. So there are two
decision variables:

x: the number of units of P

y: the number of units of Q

We usually seek a criterion (or a measure) to compare alternative solutions. This yields the objective
function.

We want to maximize the total profit. The profit per each unit of product P is $25 and profit per each
unit of product Q is $30. Therefore, the total profit is 25x+30y if we produce x units of P and y units of
Q. This leads to the following objective function:

max 25x+30y

Note that:

1: The objective function is linear in terms of decision variables x and y (i.e., it is of the form ax
+ by, where a and b are constant).

2: We typically use the variable z to denote the value of the objective.

So the objective function can be stated as:

max z=25x+30y

In many practical problems, there are limitations (such as resource / physical / strategic / economical)
that restrict our decisions. We describe these limitations using mathematical constraints.

What are the constraints in our example?

The amount of time that machine A is available restricts the quantities to be manufactured. If we
produce x units of P and y units of Q, machine A should be used for 50x+24y minutes since each unit
of P requires 50 minutes processing time on machine A and each unit of Q requires 24 minutes
processing time on machine A. On the other hand, machine A is available for 40 hours or equivalently
for 2400 minutes.

This imposes the following constraint:

50x + 24y ≤ 2400

Similarly, the amount of time that machine B is available imposes the following constraint:

30x + 33y ≤ 2100

These constrains are linear inequalities since in each constraint the left-hand side of the inequality sign
is a linear function in terms of the decision variables x and y and the right hand side is constant.

Note: In most problems, the decision variables are required to be nonnegative, and this should be
explicitly included in the formulation. This is the case here. So you need to include the following two
nonnegativity constraints as well:

x ≥0 and y ≥0

So the constraints we are subject to (s.t.) are :

50x + 24y ≤ 2400 (machine A time)

30x + 33y ≤ 2100 (machine B time)

x≥0

y ≥ 0.

Now, write the LP by putting all the LP elements together.

Here is the LP:

max z= 25x + 30y

s.t. 50x + 24y ≤ 2400

30x + 33y ≤ 2100

x≥0

y≥0
Note: To be realistic, we would require integrality for the decision variables x and y. It will lead to an
integer program if we include integrality. Integer programs are harder to solve and we will consider
them in later lectures. For the moment, we leave out integrality and consider this LP in this tutorial.

A Manufacturing Example

Problem Statement: An operations manager is trying to determine a production plan for the next
week. There are three products (say, P, Q, and Q) to produce using four machines (say, A and B, C, and
D). Each of the four machines performs a unique process. There is one machine of each type, and each
machine is available for 2400 minutes per week. The unit processing times for each machine is given
in Table 1.

Table 1: Machine Data

The unit revenues and maximum sales for the week are indicated in Table 2. Storage from one week
to the next is not permitted. The operating expenses associated with the plant are $6000 per week,
regardless of how many components and products are made. The $6000 includes all expenses except
for material costs.

Table 2: Product Data

Task: Here we seek the “optimal” product mix-- that is, the amount of each product that should be
manufactured during the present week in order to maximize profits. Formulate this as an LP.
We are trying to select the optimal product mix, so we define three decision variables as follows:

p: number of units of product P to produce,

q: number of units of product Q to produce,

r: number of units of product R to produce.

Our objective is to maximize profit:

Profit = (90-45)p + (100-40)q + (70-20r – 6000 = 45p + 60q + 50r – 6000

Note: The operating costs are not a function of the variables in the problem. If we were to drop the
$6000 term from the profit function, we would still obtain the same optimal mix of products. Thus, the
objective function is

z = 45p + 60q + 50r

The amount of time a machine is available and the maximum sales potential for each product restrict
the quantities to be manufactured. Since we know the unit processing times for each machine, the
constraints can be written as linear inequalities as follows:

20p+10q +10r ≤ 2400 (Machine A)

12p+28q+16r ≤ 2400 (Machine B)

15p+6q+16r ≤ 2400 (Machine C )

10p+15q+0r ≤ 2400 (Machine D)

Observe that the unit for these constraints is minutes per week. Both sides of an inequality must be in
the same unit. The market limitations are written as simple upper bounds.

Market constraints:

P ≤ 100, Q ≤ 40, R ≤ 60

Logic indicates that we should also include nonnegativity restrictions on the variables .

Nonnegativity constraints:

P ≥ 0, Q ≥ 0, R ≥ 0
By combining the objective function and the constraints, we obtain the LP model as follows:

max z=45p+60q+50r

s.t. 20p+10q+10r ≤ 2400

12p+28q+16r ≤ 2400

15p+6q+16r ≤ 2400

10p+15q+0r ≤ 2400

0 ≤ p ≤ 100

0 ≤ q ≤ 40

0 ≤ r ≤ 60

The optimal solution to this LP is

p*=81.82, q*=16.36, r*=60

with the corresponding objective value z=$7664. To compute the profit for the week, we reduce this
value by $6000 for operating expenses and get $1664. By setting the production quantities to the
amount specified by the solution, we find machine usage as shown in the following table.