BM1710

LINEAR PROGRAMMING

Linear programming is a widely used mathematical technique designed to help operations managers plan
and make the decisions necessary to allocate resources.
Characteristics of Linear Programming Models
Before solving problems in operations with linear programming, there are seven (7) characteristics of all
linear programming models:
Objective function: Linear programming is an optimization process. A single objective function states
mathematically what is being maximized (e.g., profit or present value) or minimized (e.g., cost or scrap).
The objective function provides the scorecard on which the attractiveness of different solutions is judged.

Decision variables: Solving a linear programming model yields the optimal values for decision variables,
which represent choices that the decision-maker can control. For example, a decision variable could be
the number of units of a product to make next month or the number of units of inventory to hold next
month. Linear programming is based on the assumption that decision variables are continuous; they can
be fractional quantities and need not be whole numbers. Often, this assumption is realistic, as when the
decision variable is expressed in dollars, hours, or some other continuous measure. Even when the
decision variables represent non-divisible units, such as workers, tables, or trucks, we sometimes can
simply round the linear programming solution up or down to get a reasonable solution that does not violate
any constraints, or we can use a more advanced technique, called integer programming.

Constraints: Limitations that restrict the permissible choices for the decision variables are called
constraints. Each limitation can be expressed mathematically in one (1) of three (3) ways: a less-than-or-
equal-to (≤), an equal-to (=), or a greater-than-or-equal-to (≥) constraint. A ≤ constraint puts an upper
limit on some function of decision variables and most often is used with maximization problems. For
example, a ≤ constraint may specify the maximum number of customers who can be served or the
capacity limit of a machine. An = constraint means that the function must equal some value. For example,
100 (not 99 or 101) units of one (1) product must be made. An = constraint often is used for certain
mandatory relationships, such as the fact that ending inventory always equals beginning inventory plus
production minus sales. A ≥ constraint puts a lower limit on some function of decision variables. For
example, a ≥ constraint may specify that production of a product must exceed or equal demand.

Feasible region: Every linear programming problem must have one (1) or more constraints. Taken
together, the constraints define a feasible region, which represents all permissible combinations of the
decision variables. In some unusual situations, the problem is so tightly constrained that there is only one
(1) possible solution—or perhaps none. However, in the usual case, the feasibility region contains
infinitely many possible solutions, assuming that the feasible combinations of the decision variables can
be fractional values. The goal of the decision-maker is to find the best possible solution.

Parameters: The objective function and constraints are functions of decision variables and parameters.
A parameter, also known as a coefficient or given constant, is a value that the decision-maker cannot
control and that does not change when the solution is implemented. Each parameter is assumed to be
known with certainty. For example, a computer programmer may know that running a software program
will take 30 minutes—no more, no less.
Linearity: The objective function and constraint equations are assumed to be linear. Linearity implies
proportionality and additivity—there can be no products (e.g., 10𝑥𝑥1 𝑥𝑥2) or powers (e.g., 𝑥𝑥13 ) of decision
variables. Suppose that the profit gained by producing two (2) types of products (represented by decision
variables 𝑥𝑥1 and 𝑥𝑥2) is 2𝑥𝑥1 + 3𝑥𝑥2. Proportionality implies that one (1) unit of 𝑥𝑥1 contributes to $2 to profits
and two (2) units contributes $4, regardless of how much of 𝑥𝑥1 is produced.

12 Handout 1 *Property of STI

 [email protected] Page 1 of 6
 BM1710

Non-negativity: This means that the decision variables must be positive or zero. A firm that makes
spaghetti sauce, for example, cannot produce a negative number of jars. To be formally correct, a linear
programming formulation should show 𝑎𝑎 ≥ 0 constraint for each decision variable (Krajewski, Malhotra,
& Ritzman, 2016).

Formulating a Linear Programming Model

Case of Stratton Company

The Stratton Company produces two (2) basic types of plastic pipe. Three (3) resources are crucial to
the output of pipe: extrusion hours, packaging hours, and a special additive to the plastic raw material.
The following data represent next week’s situation. All data are expressed in units of 100 feet of pipe.
PRODUCT
Resource Type 1 Type 2 Resource Availability
Extrusion 4 hr 6 hr 48 hr
Packaging 2 hr 2 hr 18 hr
Additive mix 2 lb 1 lb 16 lb

The contribution to profits and overhead per 100 feet of pipe is $34 for type 1 and $40 for type 2.
Formulate a linear programming model to determine how much of each type of pipe should be produced
to maximize contribution to profits and to overhead, assuming that everything produced can be sold.
Step 1: Define the decision variables.
Let:
𝑥𝑥1 be the amount of type 1 pipe to be produced and sold next week measured in 100-foot increments
(e.g., 𝑥𝑥1 = 2 means 200 feet of type 1 pipe
𝑥𝑥2be the amount of type 2 pipe to be produced and sold next week, measured in 100-foot increments
Step 2: Write out the objective function.

The goal is to maximize the total contribution that the two (2) products make to profits and overhead.
Each unit of 𝑥𝑥1 yields $34, and each unit of 𝑥𝑥2 yields $40. For specific values of 𝑥𝑥1 and𝑥𝑥2, we find the
total profit by multiplying the number of units of each product produced by the profit per unit and adding
them. Thus, our objective function becomes:

𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀: $34𝑥𝑥1 + $40𝑥𝑥2

Step 3: Write out the constraints.
Each unit of 𝑥𝑥1 and 𝑥𝑥2 produced consumes some of the critical resources. In the extrusion department,
a unit of 𝑥𝑥1 requires four (4) hours and a unit of 𝑥𝑥2 requires six (6) hours. The total must not exceed the
48 hours of capacity available, so we use the ≤ sign. Thus, the first constraint is:

4𝑥𝑥1 + 6𝑥𝑥2 ≤ 48
Similarly, the constraints for the packaging and raw materials are:
2𝑥𝑥1 + 2𝑥𝑥2 ≤ 18 (𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝)
2𝑥𝑥1 + 𝑥𝑥2 ≤ 16 (𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎𝑎 𝑚𝑚𝑚𝑚𝑚𝑚)
The three (3) constraints restrict the choices of values for the decision variable because the values that
will be chosen for 𝑥𝑥1 and 𝑥𝑥2 must satisfy all of the constraints. Negative values for 𝑥𝑥1 and 𝑥𝑥2 do not make
sense. Thus, non-negativity restrictions should be added to the model:
𝑥𝑥1 ≥ 0 and 𝑥𝑥2 ≥ 0

12 Handout 1 *Property of STI

 [email protected] Page 2 of 6
 BM1710

The entire model with the complete definitions of variables:

𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀: $34𝑥𝑥1 + $40𝑥𝑥2
𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑡𝑡𝑡𝑡: 4𝑥𝑥1 + 6𝑥𝑥2 ≤ 48
2𝑥𝑥1 + 2𝑥𝑥2 ≤ 18
2𝑥𝑥1 + 𝑥𝑥2 ≤ 16
𝑥𝑥1 ≥ 0 and 𝑥𝑥2 ≥ 0
Where:
𝑥𝑥1 be the amount of type 1 pipe to be produced and sold next week measured in 100-foot
increments
𝑥𝑥2be the amount of type 2 pipe to be produced and sold next week, measured in 100-foot
increments
Graphic Analysis
Linear Programming Model for Stratton Company
𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀: $34𝑥𝑥1 + $40𝑥𝑥2
𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 𝑡𝑡𝑡𝑡: 4𝑥𝑥1 + 6𝑥𝑥2 ≤ 48
2𝑥𝑥1 + 2𝑥𝑥2 ≤ 18
2𝑥𝑥1 + 𝑥𝑥2 ≤ 16
𝑥𝑥1 ≥ 0 and 𝑥𝑥2 ≥ 0
Where:
𝑥𝑥1 be the amount of type 1 pipe to be produced and sold next week measured in 100-foot
increments
𝑥𝑥2be the amount of type 2 pipe to be produced and sold next week, measured in 100-foot
increments
Step 1: Plot the constraints.
To start plotting the constraint equations, disregard the inequality portion of the constraints (< or >).
Making each constraint an equality (=) transforms it into the equation for a straight line.
To find the 𝑥𝑥1 axis intercept, set 𝑥𝑥2 equal to 0 and solve the equation for 𝑥𝑥1.
For the 𝑥𝑥1 axis intercept, 𝑥𝑥2 = 0, thus,
4𝑥𝑥1 + 6(0) = 48
𝑥𝑥1 = 12
For the 𝑥𝑥2 axis intercept, 𝑥𝑥1 = 0, thus,
4(0) + 6𝑥𝑥2 = 48
𝑥𝑥2 = 8
Graph the points (12, 0) and (0, 8) with a straight line with 𝑥𝑥1 axis as the horizontal axis, and 𝑥𝑥2 axis as
the vertical axis.

12 Handout 1 *Property of STI

 [email protected] Page 3 of 6
 BM1710

Figure 1. Graph of the Extrusion Constraint

Source: Operations Management Processes and Supply Chains (11th ed.), 2016, p.433

Find the axis points for the other constraints.

Packaging process line: (9, 9)
2𝑥𝑥1 + 2𝑥𝑥2 = 18
2𝑥𝑥1 + 2(0) = 18
𝑥𝑥1 = 9
2(0) + 2𝑥𝑥2 = 18
𝑥𝑥2 = 9
Additive mix’s line: (8, 16)
2𝑥𝑥1 + 0 = 16
𝑥𝑥1 = 8
2(0) + 𝑥𝑥2 = 16
𝑥𝑥2 = 16

Figure 2. Graph of the Three (3) Constraints

Source: Operations Management Processes and Supply Chains, 2016, p.433

12 Handout 1 *Property of STI

 [email protected] Page 4 of 6
 BM1710

Step 2: Identify the feasible region.

1. For the = constraint, only the points on the line are feasible solutions.
2. For the ≤ constraint, the points on the line and the points below or to the left of line are feasible
solutions.
3. For the ≥ constraint, the points on the line and the points above or the right of the line are feasible
solutions.
Because the problem contains only ≤ constraints, and the parameters on the left-hand side of each
constraint are not negative, the feasible portions are to the left of and below each constraint. The feasible
region, shaded in the figure below, satisfies all three (3) constraints simultaneously.

Figure 3. Feasible Region for the Stratton Company Linear Programming Model
Source: Operations Management Processes and Supply Chains (11th ed.), 2016, p.435

Step 3: Plot the objective function line. (Corner-Point Solution Method)

The corner-point solution method involves looking at the profit at every corner point of the feasible region.
As seen in Figure 3, there are five (5) corner points in the feasible region:
Point A (0, 0) Profit $34(0) + $40(0) = 0
Point B (0, 8) Profit $34(0) + $40(8) = 320
Point E (8, 0) Profit $34(8) + $40(0) = 272
Step 4: Algebraic Method
To determine what profit would point C and D yield:
Corner point C lies on equations 4𝑥𝑥1 + 6𝑥𝑥2 ≤ 48 and 2𝑥𝑥1 + 2𝑥𝑥2 ≤ 18. Equate the two (2) equations and
multiply the second equation by 2.
4𝑥𝑥1 + 6𝑥𝑥2 = 48
−4𝑥𝑥1 − 4𝑥𝑥2 = −36
2𝑥𝑥2 = 12
𝑥𝑥2 =6
To find 𝑥𝑥1,
4𝑥𝑥1 + 6𝑥𝑥2 = 48
4𝑥𝑥1 + 6(6) = 48

12 Handout 1 *Property of STI

 [email protected] Page 5 of 6
 BM1710

4𝑥𝑥1 + 36 = 48
4𝑥𝑥1 = 12
𝑥𝑥1 =3
Thus corner point C is (3, 6).
Corner point D lies on equations 2𝑥𝑥1 + 𝑥𝑥2 ≤ 16 and 2𝑥𝑥1 + 2𝑥𝑥2 ≤ 18. Equate the two (2) equations:
2𝑥𝑥1 + 𝑥𝑥2 = 16
−2𝑥𝑥1 − 2𝑥𝑥2 = −18
−𝑥𝑥2 = −2
𝑥𝑥2 =2
To find 𝑥𝑥1,
2𝑥𝑥1 + 𝑥𝑥2 = 16
2𝑥𝑥1 + 2 = 16
2𝑥𝑥1 = 14
𝑥𝑥1 =7
Thus corner point D is (7, 2).
Going back to the corner point method,
Point C (3, 6) Profit $34(3) + $40(6) = 342
Point D (7, 2) Profit $34(7) + $40(2) = 318
In conclusion, the optimal solution is corner point C, which yields the highest profit of $342.

REFERENCES:

Heizer, J., Render, B., & Munson, C. (2017). Operations management: Sustainability and supply chain
Management (12th ed.). Boston: Pearson Education Inc.
Krajewski, L. J., Malhotra, M. K., & Ritzman, L. P. (2016). Operations management: Processes and supply
chains (11th ed.). Essex: Pearson Education Limited.

12 Handout 1 *Property of STI

 [email protected] Page 6 of 6