Chapter

LINEAR PROGRAMMING

Many major decisions faced by a manager of a business focus on the best

way to achieve the objectives of the firm, subject to the restrictions placed on the
manager by the operating environment. These restrictions can take the form of
limited resources, such as time, labor, energy, material, or money; or they can be
in the form of restrictive guidelines, such as a recipe for making cereal or
engineering specifications. One of the most frequent objectives of business firms
is to gain the most profit possible or, in other words, to maximize profit. The
objective of individual organizational units within a firm (such as a production or
packaging department) is often to minimize cost. When a manager attempts to
solve a general type of problem by seeking an objective that is subject to
restrictions, the management science technique called linear programming is
frequently used. It has two types, the maximization and minimization model.
There are three steps in applying the linear programming technique. First,
the problem must be identified as being solvable by linear programming. Second,
the unstructured problem must be formulated as a mathematical model. Third, the
model must be solved by using established mathematical techniques. The linear
programming technique derives its name from the fact that the functional
relationships in the mathematical model are linear, and the solution technique
consists of predetermined mathematical steps—that is, a program. In this chapter
we will concern ourselves with the formulation of the mathematical model that
represents the problem and then with solving this model by using a graph.
General Objectives:

At the end of the chapter, you should be able to:

1. appreciate the use and application of linear programming in decision-

making;
2. formulate model in solving business problems specially on the
maximization and minimization of the business resources; and
 3. appreciate the use of graph in decision-making problem.

Lesson 1: Maximization Model

A linear programming model consists of certain common components and

characteristics. The model components include decision variables, an objective
function, and model constraints, which consist of decision variables and
parameters. Decision variables are mathematical symbols that represent levels of
activity by the firm. For example, an electrical manufacturing firm desires to
produce radios, toasters, and clocks, where, and are symbols representing
unknown variable quantities of each item. The final values of, and, as determined
by the firm, constitute a decision (e.g., the equation radios is a decision by the firm
to produce 100 radios).

At the end of the lesson, you should be able to:

1. enumerate the model components of linear programming;
2. apply maximization model on decision-making problem;
3. formulate maximization linear programming; and
4. formulate mathematical model using a graph.

The objective function is a linear mathematical relationship that describes

the objective of the firm in terms of the decision variables. The objective function
always consists of either maximizing or minimizing some value (e.g., maximize the
profit or minimize the cost of producing radios).
The model constraints are also linear relationships of the decision variables;
they represent the restrictions placed on the firm by the operating environment.
The restrictions can be in the form of limited resources or restrictive guidelines.
For example, only 40 hours of labor may be available to produce radios during
production. The actual numeric values in the objective function and the constraints,
such as the 40 hours of available labor, are parameters.
The next section presents an example of how a linear programming model
is formulated. Although this example is simplified, it is realistic and represents the
type of problem to which linear programming can be applied. In the example, the
model components are distinctly identified and described. By carefully studying
this example, you can become familiar with the process of formulating linear
programming models.

Maximization Model
Beaver Creek Pottery Company is a small crafts operation run by a Native
American tribal council. The company employs skilled artisans to produce clay
bowls and mugs with authentic Native American designs and colors. The two
primary resources used by the company are special pottery clay and skilled labor.
Given these limited resources, the company desires to know how many bowls and
mugs to produce each day in order to maximize profit. This is generally referred to
as a product mix problem type. This scenario is illustrated in Figure 2.1.
 The two products have the following resource requirements for production
and profit per item produced (i.e., the model parameters):
Resource Requirements
Product Labor Clay Profit
(hr./unit) (lb./unit) (per unit)
Bowl 1 4 40
Mug 2 3 50

There are 40 hours of labor and 120 pounds of clay available each day for
production. We will formulate this problem as a linear programming model by
defining each component of the model separately and then combining the
components into a single model. The steps in this formulation process are
summarized as follows:

Summary of LP Model Formulation Steps

Step 1: Define the decision variables

How many bowls and mugs to produce
Step 2: Define the objective function
Maximize profit
Step 3: Define the constraints
The resources (clay and labor) available)

Decision Variables
The decision confronting management in this problem is how many bowls
and mugs to produce. The two decision variables represent the number of bowls
and mugs to be produced on a daily basis. The quantities to be produced can be
represented symbolically as
𝑥1 = 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑏𝑜𝑤𝑙𝑠 𝑡𝑜 𝑝𝑟𝑜𝑑𝑢𝑐𝑒
𝑥2 = 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑚𝑢𝑔𝑠 𝑡𝑜 𝑝𝑟𝑜𝑑𝑢𝑐𝑒
The Objective Function
The objective of the company is to maximize total profit. The company’s
profit is the sum of the individual profits gained from each bowl and mug. Profit
derived from bowls is determined by multiplying the unit profit of each bowl, $40,
by the number of bowls produced, 𝑥1 . Likewise, profit derived from mugs is derived
from the unit profit of a mug, $50, multiplied by the number of mugs
produced, 𝑥2 . Thus, total profit, which we will define symbolically as Z, can be
expressed mathematically as $40𝑥1 + $50𝑥2 . By placing the term maximize in
front of the profit function, we express the objective of the firm—to maximize total
profit:
𝑚𝑎𝑥𝑖𝑚𝑖𝑧𝑒 𝑍 = $40𝑥1 + $50𝑥2
where
Z = total profit per day
$40𝑥1 = profit from bowls
$50𝑥2 = profit from mugs
Model Constraints
In this problem, two resources are used for production—labor and clay—both of
which are limited. Production of bowls and mugs requires both labor and clay. For
each bowl produced, 1 hour of labor is required. Therefore, the labor used for the
production of bowls is 1𝑥1 hours. Similarly, each mug requires 2 hours of labor;
thus, the labor used to produce mugs every day is 2𝑥2 hours. The total labor used
by the company is the sum of the individual amounts of labor used for each
product:
1𝑥1 + 2𝑥2
However, the amount of labor represented by is limited to 40 hours per day;
thus, the complete labor constraint is
1𝑥1 + 2𝑥2 ≤ 40 ℎ𝑟
The “less than or equal to” (≤) inequality is employed instead of an equality
because the 40 hours of labor is a maximum limitation that can be used, not an
amount that must be used. This constraint allows the company some flexibility; the
company is not restricted to using exactly 40 hours but can use whatever amount
is necessary to maximize profit, up to and including 40 hours. This means that it is
possible to have idle, or excess, capacity (i.e., some of the 40 hours may not be
used).
The constraint for clay is formulated in the same way as the labor constraint.
Because each bowl requires 4 pounds of clay, the amount of clay used daily for
the production of bowls is 4𝑥1 pounds; and because each mug requires 3 pounds
of clay, the amount of clay used daily for mugs is 3𝑥2 . Given that the amount of
clay available for production each day is 120 pounds, the material constraint can
be formulated as
4𝑥1 + 3𝑥2 ≤ 120 𝑙𝑏
A final restriction is that the number of bowls and mugs produced must be
either zero or a positive value because it is impossible to produce negative items.
These restrictions are referred to as nonnegativity constraints and are expressed
mathematically as
𝑥1 ≥ 0, 𝑥2 ≥ 0
The complete linear programming model for this problem can now be
summarized as follows:
𝑚𝑎𝑥𝑖𝑚𝑖𝑧𝑒 𝑍 = $40𝑥1 + $50𝑥2
Subject to
1𝑥1 + 2𝑥2 ≤ 40
4𝑥1 + 3𝑥2 ≤ 120
𝑥1 , 𝑥2 ≥ 0
The solution of this model will result in numeric values for and that will
maximize total profit, Z. As one possible solution, consider 𝑥1 = 5 bowls and 𝑥2 =
10 mugs. First, we will substitute this hypothetical solution into each of the
constraints in order to make sure that the solution does not require more resources
than the constraints show are available:
1 (5) + 2(10) ≤ 40
25 ≤ 40
and
4(5) + 3(10) ≤ 120
 50 ≤ 120
Because neither of the constraints is violated by this hypothetical solution, we say
the solution is feasible (i.e., possible). Substituting these solution values in the
objective function gives Z = 40 (5) + 50 (10) = $700. However, for the time being,
we do not have any way of knowing whether $700 is the maximum profit.
Now consider a solution of 𝑥1 = 10 bowls and 𝑥2 = 20 mugs. This solution
results in a profit of
Z = $40 (10) + $50(20)
= 400 + 1,000
= $1,400
Although this is certainly a better solution in terms of profit, it is infeasible
(i.e., not possible) because it violates the resource constraint for labor:
1(10) + 2(20) ≤ 40
50 ≤ 40
The solution to this problem must maximize profit without violating the
constraints. The solution that achieves this objective is 𝑥1 = 24 bowls and 𝑥2 = 8
mugs, with a corresponding profit of $1,360. The determination of this solution is
shown using the graphical solution approach in the following section.
Graphical Solutions of Linear Programming Models
Following the formulation of a mathematical model, the next stage in the
application of linear programming to a decision-making problem is to find the
solution of the model. A common solution approach is to solve algebraically the
set of mathematical relationships that form the model either manually or using a
computer program, thus determining the values for the decision variables.
However, because the relationships are linear, some models and solutions can be
illustrated graphically.
The graphical method is realistically limited to models with only two decision
variables, which can be represented on a graph of two dimensions. Models with
three decision variables can be graphed in three dimensions, but the process is
quite cumbersome, and models of four or more decision variables cannot be
graphed at all.
 Although the graphical method is limited as a solution approach, it is very
useful at this point in our presentation of linear programming in that it gives a
picture of how a solution is derived. Graphs can provide a clearer understanding
of how the computer and mathematical solution approaches presented in
subsequent chapters work and, thus, a better understanding of the solutions.

Graphical Solution of a Maximization Model

The product mix model will be used to demonstrate the graphical
interpretation of a linear programming problem. Recall that the problem describes
Beaver Creek Pottery Company’s attempt to decide how many bowls and mugs to
produce daily, given limited amounts of labor and clay.
The complete linear programming model was formulated as
𝑚𝑎𝑥𝑖𝑚𝑖𝑧𝑒 𝑍 = $40𝑥1 + $50𝑥2
Subject to
1𝑥1 + 2𝑥2 ≤ 40 ℎ𝑟 𝑜𝑓 𝑙𝑎𝑏𝑜𝑟
4𝑥1 + 3𝑥2 ≤ 120 lb of clay
𝑥1 , 𝑥2 ≥ 0
where
𝑥1 = 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑏𝑜𝑤𝑙𝑠 𝑝𝑟𝑜𝑑𝑢𝑐𝑒𝑑
𝑥2 = 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑚𝑢𝑔𝑠 𝑝𝑟𝑜𝑑𝑢𝑐𝑒𝑑
Figure 2.2 is a set of coordinates for the decision variables and, on which the graph
of our model will be drawn. Note that only the positive quadrant is drawn (i.e., the
quadrant where 𝑥1 , 𝑎𝑛𝑑 𝑥2 will always be positive) because of the nonnegativity
constraints, 𝑥1 ≥ 0 𝑎𝑛𝑑 𝑥2 ≥ 0 .
 The first step in drawing the graph of the model is to plot the constraints on
the graph. This is done by treating both constraints as equations (or straight lines)
and plotting each line on the graph. Let’s consider the labor constraint line first:
𝑥1 + 2𝑥2 = 40
A simple procedure for plotting this line is to determine two points that are
on the line and then draw a straight line through the points. One point can be found
by letting and solving for 𝑥2 :
0 + 2𝑥2 = 40
𝑥2 = 20
Thus, one point is at the coordinates 𝑥1 = 0 𝑎𝑛𝑑 𝑥2 = 20 . A second point
can be found by letting 𝑥2 = 0 and solving for 𝑥1 :
𝑥1 + 2 (0) = 40
𝑥1 = 40
Now we have a second point, 𝑥1 = 40, 𝑥2 = 0. The line on the graph
representing this equation is drawn by connecting these two points, as shown in
Figure 2.3. However, this is only the graph of the constraint line and does not reflect
the entire constraint, which also includes the values that are less than or equal to
(≤) this line. The area representing the entire constraint is shown in Figure 2.4.
 To test the correctness of the constraint area, we check any two points—
one inside the constraint area and one outside. For example, check point A in
Figure 2.4, which is at the intersection of 𝑥1 = 10, 𝑎𝑛𝑑 𝑥2 = 10. Substituting these
values into the following labor constraint,
10 + 2 (10) ≤ 40
30 ≤ 40 ℎ𝑟
shows that point A is indeed within the constraint area, as these values
for 𝑥1 𝑎𝑛𝑑 𝑥2 yield a quantity that does not exceed the limit of 40 hours. Next, we
check point B at and 𝑥1 = 40 𝑎𝑛𝑑 𝑥2 =30:
40 + 2 (30) ≤ 40
 100 ≤ 40 ℎ𝑟

Point B is obviously outside the constraint area because the values for
𝑥1 𝑎𝑛𝑑 𝑥2 yield a quantity (100) that exceeds the limit of 40 hours.
We draw the line for the clay constraint the same way as the one for the
labor constraint— by finding two points on the constraint line and connecting them
with a straight line. First, let 𝑥1 = 0 𝑎𝑛𝑑 𝑠𝑜𝑙𝑣𝑒 𝑓𝑜𝑟 𝑥2 :
40 (0) + 3 𝑥2 = 120
𝑥2 = 40
Performing this operation results in a point, 𝑥1 = 0 , 𝑥2 = 40. Next, we
, 𝑥2 = 0 and then solve for 𝑥1 :
40 𝑥1 + 3(0) = 120
𝑥1 = 30

Combining the two individual graphs for both labor and clay (Figures 2.4 and 2.5)
produces a graph of the model constraints, as shown in Figure 2.6. The shaded
area in Figure 2.6 is the area that is common to both model constraints. Therefore,
 this is the only area on the graph that contains points (i.e., values for 𝑥1 𝑎𝑛𝑑 𝑥2 )
that will satisfy both constraints simultaneously. For example, consider the points
R, S, and T in Figure 2.7. Point R satisfies both constraints; thus, we say it is a
feasible solution point. Point S satisfies the clay constraint but exceeds the labor
constraint; thus, it is infeasible. Point T satisfies neither constraint; thus, it is also
infeasible.
The shaded area in Figure 2.7 is referred to as the feasible solution area
because all the points in this area satisfy both constraints. Some point within this
feasible solution area will result in maximum profit for Beaver Creek Pottery
Company. The next step in the graphical solution approach is to locate this point.

The feasible solution area is

an area on the graph that is
bounded by the constraint
equations
The Optimal Solution Point
The second step in the graphical solution method is to locate the point in
the feasible solution area that will result in the greatest total profit. To begin the
solution analysis, we first plot the objective function line for an arbitrarily selected
level of profit. For example, if we say profit, Z, is $800, the objective function is
$800 = 40 𝑥1 + 50 𝑥2
Plotting this line just as we plotted the constraint lines results in the graph
shown in Figure 2.8. Every point on this line is in the feasible solution area and will
result in a profit of $800 (i.e., every combination of 𝑥1 𝑎𝑛𝑑 𝑥2 on this line will give
a Z value of $800). However, let us see whether an even greater profit will still
provide a feasible solution. For example, consider profits of $1,200 and $1,600, as
shown in Figure 2.9.
A portion of the objective function line for a profit of $1,200 is outside the
feasible solution area, but part of the line remains within the feasible area.
Therefore, this profit line indicates that there are feasible solution points that give
a profit greater than $800. Now let us increase profit again, to $1,600. This profit
line, also shown in Figure 2.9, is completely outside the feasible solution area. The
fact that no points on this line are feasible indicates that a profit of $1,600 is not
possible.
 Because a profit of $1,600 is too great for the constraint limitations, as
shown in Figure 2.9, the question of the maximum profit value remains. We can
see from Figure 2.9 that profit increases as the objective function line moves away
from the origin (i.e., the point 𝑥1 = 0 , 𝑥2 = 0). Given this characteristic, the
maximum profit will be attained at the point where the objective function line is
farthest from the origin and is still touching a point in the feasible solution area.
This point is shown as point B in Figure 2.10.
To find point B, we place a straightedge parallel to the objective function
line $800 = 40 𝑥1 + 50 𝑥2 in Figure 2.10 and move it outward from the origin as far
as we can without losing contact with the feasible solution area. Point B is referred
to as the optimal (i.e., best) solution.

The Solution Values

The third step in the graphical solution approach is to solve for the values
of 𝑥1 𝑎𝑛𝑑 𝑥2 once the optimal solution point has been found. It is possible to
determine the 𝑥1 𝑎𝑛𝑑 𝑥2 coordinates of point B in Figure 2.10 directly from the
graph, as shown in Figure 2.11. The graphical coordinates corresponding to point
B in Figure 2.11 are 𝑥1 = 24 𝑎𝑛𝑑 𝑥2 = 8 . This is the optimal solution for the
decision
variables in the problem. However, unless an absolutely accurate graph is drawn,
it is frequently difficult to determine the correct solution directly from the graph. A
more exact approach is to determine the solution values mathematically once the
optimal point on the graph has been determined. The mathematical approach for
determining the solution is described in the following pages. First, however, we will
consider a few characteristics of the solution.
In Figure 2.10, as the objective function was increased, the last point it
touched in the feasible solution area was on the boundary of the feasible solution
area. The solution point is always on this boundary because the boundary contains
the points farthest from the origin (i.e., the points corresponding to the greatest
profit). This characteristic of linear programming problems reduces the number of
possible solution points considerably, from all points in the solution area to just
those points on the boundary. However, the number of possible solution points is
reduced even more by another characteristic of linear programming problems.
The solution point will be on the boundary of the feasible solution area and
at one of the corners of the boundary where two constraint lines intersect. (The
graphical axes, you will recall, are also constraints because 𝑥1 ≥ 0 𝑎𝑛𝑑 𝑥2 ≥ 0).
These corners (points A, B, and C in Figure 2.11) are protrusions, or extremes, in
the feasible solution area; they are called extreme points. It has been proven
mathematically that the optimal solution in a linear programming model will always
occur at an extreme point. Therefore, in our sample problem the possible solution
points are limited to the three extreme points, A, B, and C. The optimal extreme
point is the extreme point the objective function touches last as it leaves the
feasible solution area, as shown in Figure 2.10.
From the graph shown in Figure 2.10, we know that the optimal solution
point is B. Because point B is formed by the intersection of two constraint lines, as
shown in Figure 2.11, these two lines are equal at point B. Thus, the values of and
at that intersection can be found by solving the two equations simultaneously.
First, we convert both equations to functions of 𝑥1:
𝑥1 + 2𝑥2 = 40
𝑥1 = 40 − 2𝑥2

and
 4𝑥1 + 3𝑥2 = 120
4𝑥1 = 120 − 3𝑥2
𝑥1 = 30 − 3𝑥2 /4
Now, we let 𝑥1 in the first equation equal 𝑥1 in the second equation
40 − 2𝑥2 = 30 − 3𝑥2 /4
and solve for 𝑥2 :
5𝑥2 /4 = 10
𝑥2 = 8
Substituting 𝑥2 = 8 into either one of the original equations given a value
for 𝑥1 :
𝑥1 = 40 − 2𝑥2
𝑥1 = 40 − 2(8)
𝑥1 = 24
Thus, the optimal solution at point B in Figure 2.11 is 𝑥1 = 24 𝑎𝑛𝑑 𝑥2 = 8.
Substituting these values into the objective function gives the maximum profit,
Z = 40 𝑥1 + 50 𝑥2
Z = 40(24) + 50(8)
Z = $1,360
In terms of the original problem, the solution indicates that if the pottery
company produces 24 bowls and 8 mugs, it will receive $1,360, the maximum daily
profit possible (given the resource constraints).
Given that the optimal solution will be at one of the extreme corner points,
A, B, or C, we can also find the solution by testing each of the three points to see
which results in the greatest profit, rather than by graphing the objective function
and seeing which point it last touches as it moves out of the feasible solution area.
Figure 2.12 shows the solution values for all three points, A, B, and C, and the
amount of profit, Z, at each point.
As indicated in the discussion of Figure 2.10, point B is the optimal solution
point because it is the last point the objective function touches before it leaves the
solution area. In other words, the objective function determines which extreme
point is optimal. This is because the objective function designates the profit that
will accrue from each combination of 𝑥1 𝑎𝑛𝑑 𝑥2 values at the extreme points. If
the objective function had had different coefficients (i.e., different 𝑥1 𝑎𝑛𝑑 𝑥2 profit
values), one of the extreme points other than x B might have been optimal.

Let’s assume for a moment that the profit for a bowl is $70 instead of $40,
and the profit for a mug is $20 instead of $50. These values result in a new
objective function, Z = 70 𝑥1 + 20 𝑥2 . If the model constraints for labor or clay are
not changed, the feasible solution area remains the same, as shown in Figure 2.13.
However, the location of the objective function in Figure 2.13 is different from that
of the original objective function in Figure 2.10. The reason for this change is that
the new profit coefficients give the linear objective function a new slope.
 The slope can be determined by transforming the objective function into the
general equation for a straight line, y = a + bx , where y is the dependent variable,
a is the y intercept, b is the slope, and x is the independent variable. For our sample
objective function, 𝑥2 is the dependent variable corresponding to y (i.e., it is on the
vertical axis), and 𝑥1 is the independent variable. Thus, the objective function can
be transformed into the general equation of a line as follows:
Z = 70 𝑥1 + 20 𝑥2
20 𝑥2 = Z − 70 𝑥1

This transformation identifies the slope of the new objective function as

−7/2 (the minus sign indicates that the line slopes downward). In contrast, the
slope of the original objective function was −4/5 .
If we move this new objective function out through the feasible solution area,
the last extreme point it touches is point C. Simultaneously solving the constraint
lines at point C results in the following solution:
𝑥1 = 30
4𝑥1 + 3𝑥2 = 120
and
𝑥2 = 40 − ( 4𝑥1 /3)
𝑥2 = 40 − 4(30)/3)
𝑥2 = 0
Thus, the optimal solution at point C in Figure 2.13 is 𝑥1 = 30 bowls, 𝑥2 =
0 mugs, and 𝑍 = $2,100 profit. Altering the objective function coefficients results in
a new solution.
This brief points. First, the optimal extreme point is determined by the
objective function, and an extreme point on one axis of the graph is as likely to be
the optimal solution as is an extreme point on a different axis. Second, the solution
is sensitive to the values of the coefficients in the objective function. If the objective
function coefficients are changed, as in our example, the solution may change.
Likewise, if the constraint coefficients are changed, the solution space and solution
points may change also. This information can be of consequence to the decision
maker trying to determine how much of a product to produce. Sensitivity
analysis— the use of linear programming to evaluate the effects of changes in
model parameters.
It should be noted that some problems do not have a single extreme point
solution. For example, when the objective function line parallels one of the
constraint lines, an entire line segment is bounded by two adjacent corner points
that are optimal; there is no single extreme point on the objective function line. In
this situation there are multiple optimal solutions. This and other irregular types
of solution outcomes in linear programming are discussed at the end of this chapter

Slack Variable
Once the optimal solution was found at point B in Figure 2.12, simultaneous
equations were solved to determine the values of 𝑥1 𝑎𝑛𝑑 𝑥2 . Recall that the
solution occurs at an extreme point where constraint equation lines intersect with
each other or with the axis. Thus, the model constraints are considered as
equations (=) rather than ≤ or ≥inequalities.
There is a standard procedure for transforming ≤ inequality constraints into
equations. This transformation is achieved by adding a new variable, called a
slack variable, to each constraint.
For the pottery company example, the model constraints are
𝑥1 + 2𝑥2 ≤ 40 ℎ𝑟. 𝑜𝑓 𝑙𝑎𝑏𝑜𝑟
𝑥1 + 3𝑥2 ≤ 120 𝑙𝑏. 𝑜𝑓 𝑐𝑙𝑎𝑦
The addition of a unique slack variable, 𝑠1, to the labor constraint and 𝑠2 to
the constraint for clay results in the following equations:
𝑥1 + 2𝑥2 + 𝑠1 = 40 ℎ𝑟. 𝑜𝑓 𝑙𝑎𝑏𝑜𝑟
4 𝑥1 + 3𝑥2 + 𝑠2 = 120𝑙𝑏. 𝑜𝑓 𝑐𝑙𝑎𝑦
 The slack variables in these equations, 𝑠1 𝑎𝑛𝑑 𝑠2 , will take on any value
necessary to make the left-hand side of the equation equal to the right-hand side.
For example, consider a hypothetical solution of 𝑠1 = 5 𝑎𝑛𝑑 𝑠2 = 10. Substituting
these values into the foregoing equations yields
𝑥1 + 2𝑥2 + 𝑠1 = 40 ℎ𝑟. 𝑜𝑓 𝑙𝑎𝑏𝑜𝑟
5 + 2(10) + 𝑠1 = 40 ℎ𝑟. 𝑜𝑓 𝑙𝑎𝑏𝑜𝑟
𝑠1 = 15 ℎ𝑟. 𝑜𝑓 𝑙𝑎𝑏𝑜𝑟
and
4𝑥1 + 3𝑥2 + 𝑠2 = 120 𝑙𝑏. 𝑜𝑓 𝑐𝑙𝑎𝑦
5(5) + 3(10) + 𝑠2 = 120 𝑙𝑏. 𝑜𝑓 𝑐𝑙𝑎𝑦
𝑠2 = 70 𝑙𝑏. 𝑜𝑓 𝑐𝑙𝑎𝑦
In this example, 𝑥1 = 5 bowls and 𝑥2 = 10 mugs represent a solution that
does not make use of the total available amount of labor and clay. In the labor
constraint, 5 bowls and 10 mugs require only 25 hours of labor. This leaves 15
hours that are not used. Thus, 𝑠1 represents the amount of unused labor, or slack.
In the clay constraint, 5 bowls and 10 mugs require only 50 pounds of clay.
This leaves 70 pounds of clay unused. Thus, 𝑠2 represents the amount of unused
clay. In general, slack variables represent the amount of unused resources.
The ultimate instance of unused resources occurs at the origin, where 𝑥1 =
0 𝑎𝑛𝑑 𝑥2 = 0. Substituting these values into the equations yields
𝑥1 + 2𝑥2 + 𝑠1 = 40
0 + 2(0) + 𝑠1 = 40
𝑠1 = 400 ℎ𝑟. 𝑜𝑓 𝑙𝑎𝑏𝑜𝑟
and
4𝑥1 + 3𝑥2 + 𝑠2 = 120
4(0) + 3(0) + 𝑠2 = 120
𝑠2 = 120 𝑙𝑏. 𝑜𝑓 𝑐𝑙𝑎𝑦
 Because no production takes place at the origin, all the resources are
unused; thus, the slack variables equal the total available amounts of each
resource: 𝑠1 = 40 hours of labor and 𝑠2 = 120 pounds of clay.
What is the effect of these new slack variables on the objective function?
The objective function for our example represents the profit gained from the
production of bowls and mugs,
𝑍 = 40𝑥1 + 50𝑥2
The coefficient 40 is the contribution to profit of each bowl; 50 is the
contribution to profit of each mug. What, then, do the slack variables and
contribute? They contribute nothing to profit because they represent unused
resources. Profit is made only after the resources are put to use in making bowls
and mugs. Using slack variables, we can write the objective function as
𝑚𝑎𝑥𝑖𝑚𝑖𝑧𝑒 (𝑍) = 40𝑥1 + 50𝑥2 + 0𝑠1 + 0 𝑠2
As in the case of decision variables (𝑥1 𝑎𝑛𝑑 𝑥2 ), slack variables can have
only nonnegative values because negative resources are not possible. Therefore,
for this model formulation 𝑥1 , 𝑥2 , 𝑠1 𝑎𝑛𝑑 𝑠2 ≥ 0.
The complete linear programming model can be written in what is referred
to as standard form with slack variables as follows:
𝑚𝑎𝑥𝑖𝑚𝑖𝑧𝑒 (𝑍) = 40𝑥1 + 50𝑥2 + 0𝑠1 + 0 𝑠2
Subject to
𝑥1 + 2𝑥2 + 0𝑠1 = 40
4𝑥1 + 3𝑥2 + 𝑠2 = 120
𝑥1 , 𝑥2 , 𝑠1 𝑎𝑛𝑑 𝑠2 ≥ 0.
The solution values, including the slack at each solution point, are
summarized as follows:
 Figure 2.14 shows the graphical solution of this example, with slack
variables included at each solution point

Summary of the Graphical Solution Steps

The steps for solving a graphical linear programming model are
summarized here:
1. Plot the model constraints as equations on the graph; then, considering
the inequalities of the constraints, indicate the feasible solution area.
2. Plot the objective function; then, move this line out from the origin to
locate the optimal solution point.
3. Solve simultaneous equations at the solution point to find the optimal
solution values.
Or
1. Solve simultaneous equations at each corner point to find the solution
values at each point.
2. Substitute these values into the objective function to find the set of
values that results in the maximum Z value
Lesson 2: Minimization Model

As mentioned at the beginning of this chapter, there are two types of linear
programming problems: maximization problems (like the Beaver Creek Pottery
Company example) and minimization problems. A minimization problem is
formulated the same basic way as a maximization problem, except for a few minor
differences. The following sample problem will demonstrate the formulation of a
minimization model.
At the end of the lesson, you should be able to:
1. enumerate the model components of linear programming;
2. apply minimization model on decision-making problem;
3. formulate minimization linear programming; and
4. formulate mathematical model using a graph.

A farmer is preparing to plant a crop and needs to fertilize a field. There are two
brands of fertilizer to choose from, Super-gro and Crop-quick. Each brand yields a
specific amount of nitrogen and phosphate per bag, as follows:

The farmer’s field requires at least 16 pounds of nitrogen and at least 24

pounds of phosphate. Super-gro costs $6 per bag, and Crop-quick costs $3. The
farmer wants to know how many bags of each brand to purchase in order to
minimize the total cost of fertilizing. This scenario is illustrated in Figure 2.15.
The steps in the linear programming model formulation process are
summarized as follows:
Decision Variables
The farmer’s objective is to minimize the total cost of fertilizing. The total
cost is the sum of the individual costs of each type of fertilizer purchased. The
objective function that represents total cost is expressed as
𝑚𝑎𝑥𝑖𝑚𝑖𝑧𝑒 (𝑍) = 6𝑥1 + 3𝑥2
where
6𝑥1 = 𝑐𝑜𝑠𝑡 𝑜𝑓 𝑏𝑎𝑔𝑠 𝑜𝑓 𝑆𝑢𝑝𝑒𝑟 − 𝑔𝑟𝑜
3𝑥2 = 𝑐𝑜𝑠𝑡 𝑜𝑓 𝑏𝑎𝑔𝑠 𝑜𝑓 𝐶𝑟𝑜𝑝 − 𝑞𝑢𝑖𝑐𝑘
Model Constraints
The requirements for nitrogen and phosphate represent the constraints of
the model. Each bag of fertilizer contributes a number of pounds of nitrogen and
phosphate to the field. The constraint for nitrogen is
2𝑥1 + 4𝑥2 ≥ 16 lb.
where
2𝑥1 = 𝑡ℎ𝑒 𝑛𝑖𝑡𝑟𝑜𝑔𝑒𝑛 𝑐𝑜𝑛𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 (𝑙𝑏) 𝑝𝑒𝑟 𝑏𝑎𝑔 𝑜𝑓 𝑆𝑢𝑝𝑒𝑟 − 𝑔𝑟𝑜
4𝑥2 = 𝑡ℎ𝑒 𝑛𝑖𝑡𝑟𝑜𝑔𝑒𝑛 𝑐𝑜𝑛𝑡𝑟𝑖𝑏𝑢𝑡𝑖𝑜𝑛 (𝑙𝑏) 𝑝𝑒𝑟 𝑏𝑎𝑔 𝑜𝑓 𝐶𝑟𝑜𝑝 − 𝑞𝑢𝑖𝑐𝑘

Rather than a ≤ (less than or equal to) inequality, as used in the Beaver
Creek Pottery Company model, this constraint requires a ≥ (greater than or equal
to) inequality. This is because the nitrogen content for the field is a minimum
requirement specifying that at least 16 pounds of nitrogen be deposited on the
farmer’s field. If a minimum cost solution results in more than 16 pounds of nitrogen
on the field, that is acceptable; however, the amount cannot be less than 16
pounds.
The constraint for phosphate is constructed like the constraint for nitrogen:
4𝑥1 + 3𝑥2 ≥ 24 lb.
With this example, we have shown two of the three types of linear
programming model constraints, ≤ and ≥.The third type is an exact equality, = .This
type specifies that a constraint requirement must be exact. For example, if the
farmer had said that the phosphate requirement for the field was exactly 24
pounds, the constraint would have been
4𝑥1 + 3𝑥2 ≥ 24 lb.
As in our maximization model, there are also nonnegativity constraints in
this problem to indicate that negative bags of fertilizer cannot be purchased:
𝑥1 , 𝑥2 ≥ 0
The complete model formulation for this minimization problem is
𝑚𝑎𝑥𝑖𝑚𝑖𝑧𝑒 (𝑍) = 6𝑥1 + 3𝑥2
Subject to
 2𝑥1 + 4𝑥2 ≥ 16 lb. of nitrogen
4𝑥1 + 3𝑥2 ≥ 24 lb. of phosphate
𝑥1 , 𝑥2 ≥ 0

Graphical Solution of a Minimization Model

We follow the same basic steps in the graphical solution of a minimization
model as in a maximization model. The fertilizer example will be used to
demonstrate the graphical solution of a minimization model.
 The first step is to graph the equations of the two model constraints, as
shown in Figure 2.16. Next, the feasible solution area is chosen, to reflect the ≥
inequalities in the constraints, as shown in Figure 2.17
.

After the feasible solution area has been determined, the second step in the
graphical solution approach is to locate the optimal point. Recall that in a
maximization problem, the optimal solution is on the boundary of the feasible
solution area that contains the point(s) farthest from the origin. The optimal solution
point in a minimization problem is also on the boundary of the feasible solution
area; however, the boundary contains the point(s) closest to the origin (zero being
the lowest cost possible).
As in a maximization problem, the optimal solution is located at one of the
extreme points of the boundary. In this case, the corner points represent
extremities in the boundary of the feasible solution area that are closest to the
origin. Figure 2.18 shows the three corner points—A, B, and C—and the objective
function line.
As the objective function edges toward the origin, the last point it touches in
the
feasible solution area is A. In other words, point A is the closest the objective
function can get to the origin without encompassing infeasible points. Thus, it
corresponds to the lowest cost that can be attained

The final step in the graphical solution approach is to solve for the values of
and at point A. Because point A is on the 𝑥2 axis, 𝑥1 = 0; ; thus,4(0) + 3𝑥2 = 24
𝑥2 = 8
Given that the optimal solution is 𝑥1 = 0, 𝑥2 = 8, the minimum cost, Z, is
Z = 6𝑥1 + 3𝑥2
Z = 6(0) + 3(8)
Z = $24
This means the farmer should not purchase any Super-gro but, instead,
should purchase eight bags of Crop-quick, at a total cost of $24.
Surplus Variables
Greater-than or equal-to constraints cannot be converted to equations by
adding slack variables, as with constraints. Recall our fertilizer model, formulated
as
𝑚𝑎𝑥𝑖𝑚𝑖𝑧𝑒 (𝑍) = 6𝑥1 + 3𝑥2
Subject to
2𝑥1 + 4𝑥2 ≥ 16 lb. of nitrogen
4𝑥1 + 3𝑥2 ≥ 24 lb. of phosphate
𝑥1 , 𝑥2 ≥ 0
where
𝑥1 = 𝑏𝑎𝑔𝑠 𝑜𝑓 𝑆𝑢𝑝𝑒𝑟 − 𝑔𝑟𝑜 𝑓𝑒𝑟𝑡𝑖𝑙𝑖𝑧𝑒𝑟
𝑥2 = 𝑏𝑎𝑔𝑠 𝑜𝑓 𝐶𝑟𝑜𝑝 − 𝑞𝑢𝑖𝑥𝑘 𝑓𝑒𝑟𝑡𝑖𝑙𝑖𝑧𝑒𝑟
𝑍 = 𝑓𝑎𝑟𝑚𝑒𝑟 ′𝑠𝑡𝑜𝑡𝑎𝑙 𝑐𝑜𝑠𝑡 𝑜𝑓 𝑝𝑢𝑟𝑐ℎ𝑎𝑠𝑖𝑛𝑔 𝑓𝑒𝑟𝑡𝑖𝑙𝑖𝑧𝑒𝑟
Because this problem has constraints as opposed to the constraints of the
Beaver Creek Pottery Company maximization example, the constraints are
converted to equations a little differently. Instead of adding a slack variable as we
did with a constraint, we subtract a surplus variable. Whereas a slack variable is
added and reflects unused resources, a surplus variable is subtracted and reflects
the excess above a minimum resource requirement level. Like a slack variable, a
surplus variable is represented symbolically by and must be nonnegative. For the
nitrogen constraint, the subtraction of a surplus variable gives
2𝑥1 + 4𝑥2 − 𝑠1 = 16
The surplus variable transforms the nitrogen constraint into an equation.
As an example, consider the hypothetical solution
𝑥1 = 0
𝑥2 = 10
Substituting these values into the previous equation yields
2(0) + 4(10) − 𝑠1 = 16
− 𝑠1 = 16 - 40
 − 𝑠1 = 24 lb. of nitrogen
In this equation, can be interpreted as the extra amount of nitrogen above
the minimum requirement of 16 pounds that would be obtained by purchasing 10
bags of Crop-quick fertilizer. In a similar manner, the constraint for phosphate is
converted to an equation by subtracting a surplus variable, 𝑠2 :
4𝑥1 + 3𝑥2 − 𝑠2 = 24
As is the case with slack variables, surplus variables contribute nothing to
the overall cost of a model. For example, putting additional nitrogen or phosphate
on the field will not affect the farmer’s cost; the only thing affecting cost is the
number of bags of fertilizer purchased. As such, the standard form of this linear
programming model is summarized as
𝑚𝑎𝑥𝑖𝑚𝑖𝑧𝑒 (𝑍) = 6𝑥1 + 3𝑥2 + 0𝑠1+0 𝑠2
Subject to
2𝑥1 + 4𝑥2 − 𝑠1 = 16
4𝑥1 + 3𝑥2 − 𝑠2 = 24
𝑥1 , 𝑥2 , 𝑠1 , 𝑠2 ≥ 0
Figure 2.19 shows the graphical solutions for our example, with surplus variables
included at each solution point.
Exercises
1. Moore’s Meatpacking Company produces a hot dog mixture in 1,000-pound
batches. The mixture contains two ingredients—chicken and beef. The cost
per pound of each of these ingredients is as follow
Ingredient Cost/lb
Chicken Php75
Beef Php125

Each batch has the following recipe requirements:

a. At least 500 pounds of chicken
b. At least 200 pounds of beef
The ratio of chicken to beef must be at least 2 to 1. The company wants to
know the optimal mixture of ingredients that will minimize cost. Formulate a
linear programming model for this problem
2. Solve the following linear programming model graphically:
𝑀𝑎𝑧𝑖𝑚𝑖𝑧𝑒 𝑍 = 4𝑥1 + 5𝑥2
Subject to
𝑥1 + 2𝑥2 ≤ 10
6𝑥1 + 6𝑥2 ≤ 36
𝑥1 ≤ 4
𝑥1 , 𝑥2 ≥ 0