Chapter 02 Part 01

An Introduction to Linear
Programming

Presenter:
Vincent Harold Teng
Organization:
CEMDS, CvSU Main
I. Introduction to Linear
Programming
I. Introduction
●
Linear programming is a management science
technique used in many countries and by many different
types of organization – manufacturing companies,
service organizations, government agencies, not-for
profit organizations.
●
Managers are often in a situation where they are looking
to make the best decision possible but where the
options they face limit – or constraint – what they are
actually able to do.
Typical Applications of Linear Programming.
●
A manufacturer of mobile phones is looking at the production schedule for the next few
months and trying to decide how many each phone model to manufacturer in order to meet
demand forecasts for different sales regions. Clearly, the production schedule will need to
ensure that enough phones are produced to meet projected demand in each sales region but
at the same time minimize production costs.
●
A financial analysts must select an investment portfolio for a high-value client from a variety
of investment alternatives. The analyst would like to establish the portfolio that maximizes the
return on investment. At the same time, though, the analyst also has to satisfy certain client
requirements in terms of the mix of investments and the risk associated with each.
●
A marketing manager wants to determine how best to allocate a fixed advertising budget
among alternative advertising media, such as Internet, television, newspaper, and magazine.
The manager would like to determine the media mix that maximizes advertising effectiveness
but within the budget constraints faced.
●
A regional health center collects and stores blood donatioins that are used in the region’s
hospitals. The transportation manager has to ensure that each hospital receives the supply of
blood it needs to treat patients and at the same time minimize the time taken to transport
blood supplies from the center to the various hospitals.
Typical Applications of Linear Programming.
●
A close scrutiny reveals one basic property they all have in common.
– The manufacturer wanted to minimize costs
– The financial analyst want to maximize return on investment
– The marketing manager wanted to maximize advertising effectiveness.
– The company wanted to minimize total transportation time.
●
In a linear programming problems, the maximization or minimization
of some quantity is the objective.
●
All linear problems also have a second property; restrictions or
constraints that limits the degree to which the objective can be
pursued.
1. Maximization Problem
●
Gulf-golf is a small family-run company in the United Arab Emirates. It
manufactures golfing equipment aimed at the increasing demand, primarily
from tourists coming from golfing holidays in the Gulf region. The company
has decided to manufacture two new products; a medium-priced golf bag
and a more up-market, expensive golf bag. The company’s distributor is
enthusiastic about the new product line and has agreed to buy all the golf
bags the company produces over the next three months.
●
After a thorough investigation of the steps involved in manufacturing a golf
bag, management determined that each golf bag produced will require the
following four operations:
– Cutting and dyeing of materials needed
– Sewing
– Finishing
– Inspection and Packing
1. Maximization Problem
NUMBER OF HOURS REQUIRED TO PRODUCE
Department Standard Bag Deluxe Bag
Cutting and Dyeing 7/10 1
Sewing 1/2 5/6
Finishing 1 2/3
Inspection and Packaging 1/10 1/4

Production is constrained by the limitted number of hours available in each

department. After studying departmental workload projections, the director of
manufacturing estimates that 630 hours for cutting and dyeing, 600 hours of sewing,
708 hours for finishing and 135 hours for inspection and packaging will be available
for the production of golf bags during the next three months.

The accounting department analyzed the production data, assigned all relevant
variable costs and arrived at prices for both bags that will result in a proft contribution
of $10 every standard bag and $9 for every deluxe bag produced.
a. Problem Formulation
●
Problem formulation or modelling is the process of translating the
verbal statement of a problem into a mathematical statement.
Formulating models is an art that can only be mastered with
practice and experience. But even though every problem has
unique features, most problems also have common features. As a
result, some general guidelines for model formulation can be
helpful, especially for beginner.
●
Understand the Problem Thoroughly – Read the problem
description quickly to get a feel for what is involved. You can take
notes that will help focus on the key issues and facts. We can take
our Gulfgolf as an example.
Describing the Objective and Describe Each
Constraint
●
The objective here is to maximize the total contribution to profit.
●
Four constraint relates to the number of hours of manufacturing time
available; they restrict the number of standard bags and the number of
deluxe bags that cna be producd.
– Constraint 1: Number of hours of cutting and dyeing time used must be less
than or equal to the number of hours of cutting and dyeing time available.
– Constraint 2: Number of hours of sewing time used must be less than or equal
to the number of hours sewing time available.
– Constraint 3: Number of hours of finishing time used must be less than or
equal to the number of hours of finishing time available.
– Constraint 4: Number of hours of inspection and packaging time used must be
less than or equal to the number of hours of inspection and packaging
available.
Define Decision Variables
●
A decision variable is one where management can control its
value and therefore decide what value the variable should
take. The controllable values are (1) the number of standard
bags produced (2) the number of deluxe bags produced: Let:
S = number of Standard Bags
D = number of Deluxe Bags

In the linear programming terminology. S and D are referred to

as decision variables.
Write the Objective in Terms of the Decision
Variables
●
The company’s profit contribution comes from two sources: (1) the
profit contribution made by producing S standard bags, (2) the profit
contribution made by producing D deluxe bags. If the company
makes $10 for every standard bag, the company will make $10S if S
standard bags are produced. Also, if the company makes $9 for
every deluxe bag, the company will make $9D if deluxe bags are
produced, so we have the following:
Total Profit Contribution = 10S + 9D
●
Because the objective – maximize the total profit contribution – is a
function of the decision variable S and D, we refer to 10S + 9D as the
objective function. Using ‘Max’ as abreviation we can write the
objective as
Max 10S + 9D
Write the Constraints in Terms of the Decision
Variables – Constraint 1
●
Every standard bag GulfGolf produces will use 7/10 hour cutting and
dyeing time; therefore, the total number of hours of cutting and dyeing
time used in the manufacture of S standard bags is 7/10S. In addition
because every deluxe bag produced uses one hour of cutting and
dyeing time, the production of D deluxe bags will use 1D hours of
cutting and dyeing time. So the total cutting and dyeing time required
for the production of S standards bags and D deluxe bags is given by:
7/10S + 1D ≤ 630
●
The director of GulfGolf has at most 630 hours of cutting and dyeing
time available. Therefore, the production combination we select must
satisfy the requirement:
Write the Constraints in Terms of the Decision
Variables – Constraint 2
●
From the table provided earlier, we see that every
standard bag manufactured will require ½ hour for
sewing, and every deluxe bag will require 5/6 hours for
sewing. Because 600 hours of sewing time are available
it follows that:
1/2S + 5/6D ≤ 600
Write the Constraints in Terms of the Decision
Variables – Constraint 3
●
Every standard manufactured will require one hour for
finishing, and every deluxe bag will require ⅔ hour for
finishing. Within 708 hours of finishing time available, it
follows that:
1S + ⅔D ≤ 708
Write the Constraints in Terms of the Decision
Variables – Constraint 4
●
Every standard bag manufactured will require 1/10 hour for
inspection and packaging, and every deluxe bag will require ¼
hour for inspection and packaging. Because 135 hours of
inspection and packaging time are available, it follows that:
1/10S + 1/4D ≤ 135
●
We have now specified the mathematical relationships for the
constraints associated with the four departments. Since
companys cannot produce a negative number of bags, we must
add to the constraints the following:
S, D ≥ 0
Mathematical Statement of GulfGolf Problem

●
The mathematical statement model or mathematical
formulation of the GulfGolf problem can now be placed
in a set of mathematical relationships referred to as
mathematical model. We can now have:
Mathematical Statement of GulfGolf Problem
●
Our job now is to find the product mix that satisfies the constraints and, at the same
time, yields the maximum possible value for objective function. Once these values are
created, we will have found the optimal solution to the problem.
●
The mathematical model for the problem is a linear programming model, or linear
program. The problem has the objective and constraints that, as we said earlier, are
common properties of all linear programs. The special feature that makes it a linear
program is that the objective function and all constraint functions (the left-hand sides of
the constraint inequalities) are linear functions of the decision variables.
●
Mathematical functions in which each variable appears in a separate term and is raised
to the first power are called linear functions. The objective function (10S + 9D) is
linear because each decision variable appears in a separate term and has an exponent
of 1. The amount of production time required in the cutting and dyeing department
(0.7S + 1D) is also a linear function of the decision variables for the same resaon.
Some Notes...
●
The three key assumptions necessary for linear programming model to be
appropriate are proportionality, additivity, and divisibility. Propotionaly means
that the contribution to the objective function and the amount of resources used
in each constraint are proportional to the value of each decision variable.
Additivity means that the value of the objective function and the total resources
used can be found by summing the objective function contribution and the
resources used for all decision variables. Divisiblity means that decision
variables are continuous. The dividibility assumption plus the nonnegativity
constraints mean that decision variables can take on any value greater than or
equal to zero.
●
Management scientists formulate and solve a variety of mathematical models
that contain an objective function and a set of constraints. Models of this type
are referred to as mathematical programming models. Linear programming
models are one type of mathematical programming model in that objective
function and all constraint functions are linear.
Graphical Solution Procedure
●
A linear programming problem involving only two decision variables can be
solved using a graphical solution procedure. Clearly, in the real world LP
problems will have many more decision variables (and constraints) and
cannot be solved graphically. However, the graphical solutions introduces
some important principles of LP solution that we need to understand.
●
The graph in figure 2.1 will have many values of S on the horizontal axis
and values of D on the vertical axis. Any point along the graph can be
identified by the S and D values, which indicate the position of the point
along the horizontal and vertical axes, respectively.
●
Because every point (S,D) corresponds to possible solution, every point on
the graph is called a solution point. The solution point where S=0 and D=0
is referred to as origin.
●
The result earlier was 0.7S + 1D ≤ 630
Figure 2.1
Graphical Solution Procedure
●
To show all solution points that satisfy this relationship, we start by graphing
the solution points satisfying the constraint as an equality. That is, the
points where 0.7S + 1D = 630. Because the graph of this equation is a line,
it can be obtained by identifying any two points that satisfy the equation and
then drawing a line through those points. Setting S=0 and solving for D, we
see that the point (S=0, D = 630) satisfies the equation. To find a second
point of satisfying this equation, we set D = 0 and solve for S. By doing so,
we obtain 0.7S + 1(0) = 630, or S = 900. Thus, a second point satisfying the
equation is (S=900, D=0). Given these two points, we can now graph the
line corresponding ot the equation. This line, which will be called the cutting
and dyeing constraint line, shown in Figure 2.2 We label the line C and D to
indicate that it represents cutting and dyeing constraint line.
Figure 2.2
Graphical Solution Procedure
●
Recall that inequality representing the cutting and dyeing constraint is 0.7S + 1D ≤
630. Because all points on the line satisfy the constraint, we know any point on this
line must satisfy the constraint. But to locate the solution points satisfying 0.7S +
1D ≤ 630 we can consider two solution points for instance if we let (S = 200, D =
200) and (S= 600, D = 500). The first solution line is below the constraint, while the
other one is above the constraint. Which of the solutions will satisfy the cutting and
dyeing constraint? We can get
0.7(200) + 1(200) = 340
●
This satisfies the constraint since we know that the result should be less than or
equal to 630. For the second solution line we find the solution to be
07(600) + 1(500) = 920
●
Thus the solution provided for the second line is rejected due to it being more than
630 and is not feasible.
●
In this case S=600, D=500 is not feasible, then all other points on the same side is
not feasible as well. Since S=200, D=200 met the requirements, then all solutions
on the point of that side is feasible.
Figure 2.3
Figure 2.4
Graphical Solution Procedure
●
We can do the same for other constraints. This can be shown in figure
2.4.
●
The four separate graphs shows the feasible solution points for each of
the four constraints simultaneously. To find these solution points, we
can draw all four constraints in one graph and observe the region
containing the points that satisfy all the constraints simultaneously.
●
The combined graph is shown in figure 2.5, The shaded region
includes every solution point that satisfies all the constraints
simultaneously. Solutions that satisfy all the constraint are termed
feasible solutions, and the shaded region is called the feasible
solution region or within the feasible region is a feasible solution point.
Figure 2.5
Graphical Solution Procedure
●
One approach in finding the optimal solution would be to calculate
the objective function for each feasible solution; the optimal solution
would then be the one with the largest value. The difficulty with this
approach is that there are a huge number of feasible solutions so
this trial-and-error procedure cannot be used to identify the optimal
solution.
●
Rather than trying to calculate the profit contribution for every
feasible solution, we select one arbitrary value for profit contribution
and identify all feasible solutions (S,D) that yield this selected value.
For example, if we choose $ 1,800 we can have the model 10S +
9D = 1,800
Figure 2.6
Graphical Solution Procedure
●
This expression is simply the equation of a line, So, all feasible solution
points (S,D) yielding a profit contribution of $1,800 must be on this line.
The procedure for graphing the profit or objectives line is the same. Letting
S = 0, D must be 200, by letting D= 0, S must be 180 which are also on
the line. Drawing the line through these two points identifies all the
solutions that have a profit contribution of $1,800.
●
However, we all know, the $1800 was an arbitrary choice. Since we have a
maximization LP problem, we are seeking the maximum possible profit.
We can also consider $3,600 and $5,400.
●
Figure 2.8 has the following: (1) the profit lines are parallel to each other,
and (2) higher profit lines are obtained as we move farther from the origin.
Figure 2.7
Figure 2.8
Graphical Solution Procedure
●
Because the profit lines are parallel and higher profit lines are
farther from the origin, we can obtain solutions that give
increasingly larger values for the objective function by continuing
to move the profit line farther from the origin in such a fashion
that it remains parallel to the other profit lines. However, at some
point we will find that any further outward movement will place
the profit line completely outside the feasible region. Because
solutions outside the feasible region are unacceptable, the point
in the feasible region that lies on the highest profit line is the
optimal solution to the linear program.
Figure 2.9
Graphical Solution Procedure
●
The optimal values of the decision variables are the S and D values at
the optimal solution. Depending on the accuracy of the graph, you may
not be able to determine the exact S and D values. Referring to the
graph in Fig. 2.9, the best we can do is to conclude that the optimal
production combination consists of approximately 550 standard bags
(S) and approximately 250 Deluxe Bags (D).
●
A close look will give us a solution point of 0.7S + 1D = 630 and the
finishing constraint line as 1S + 0.6667D = 708
●
Thus the optimal values of the decision variables S and D must satisfy
both Equations in 2.7 and 2.8 simultaneously. Solving for S gives us
0.7S = 630 – 1D or S = 900 – 1.4286D.
Graphical Solution Procedure
Graphical Solution Procedure
●
The exact location of the optimal solution point is S=540 and D=252. Hence, the
optimal production quantities are 540 standard bags and 252 deluxe bags, with a
resulting profit contribution of 10(540) + 9(252) = $7668.
●
We can now advise the management that the maximum possible profit
contribution that can be achieved is $7668 and that it is achieved by producing
540 standard golf bags and one that would have been very difficult to obtain any
other way.
●
For linear programming problem with two decision variables, the exact values of
the decision variables can be determined by first using the graphical solution
procedure to identify the optimal solution point and then solving the simultaneous
constraint equations associated with it.
●
There are two constraints we have used to confirm the optimal solution it is these
two constraints we have used to confirm the optimal solution are referred to as
binding constraints. (07S +1D=630) and Non-binding constraints like Sewing and
for Inspection and Packaging, which do not affect the optimal solution.
A Note on Graphing Lines
●
An important aspect of the graphical method is the ability to graph
lines showing the constraints and the objective function of the
linear program. The procedure we use for graphing the equation
of a line is to find any two points satisfying the equation and then
draw the line through the two points. For GulfGolf constraints, the
two points were easily found by first setting S=0 and solving the
constraint equation for D. Then we set D =0 and solved for S. For
the cutting and dyeing constraint line we have 0.7S+1D=630 this
procedure is identified the two points (S=0, D=630) and (S=900,
D=0). The cutting and dyeing constraint line was then graphed
drawing a line through these two points.
A Note on Graphing Lines
●
All constraint and objective function line in two-variable linear programs can
be graphed if two points on the line can be identified. However, finding the two
points on the line is not always easy as shown in the GulfGolf problem.
●
Suppose a company manufactures two models of a small handheld computer:
the Assistant (A) and the Professional (P). Management needs 50 units of the
Professional model for its own salesforce, and expects sales of the
Professional to be at most one-half of the sales of the assistant. A constraint
requirement will end up with P – 50 ≤ 0.5A OR 2P – 100 ≤ A OR 2P – A ≤ 100.
●
Using the equality form and setting P=0, we find the point (P = 0, A = -100) is
on the constraint line. Setting A=0, we find a second point (P=50, A=0) on the
constraint line.
A Note on Graphing Lines
●
If we have drawn only the non-negative (P ≥ 0, A ≥0) portion of the
graph, the first point (P=0, A = -100) cannot be plotted because A= -
100 is not on the graph.
●
Whenever we have two points on the line, but one or both of the
points cannot be plotted in the non-negative portion of the graph,
the simplest approach is to enlarge the graph. In the example
provided, the point (P = 0, A = -100) can be plotted by extending the
graph to include the negative A axis. Once both points satisfying the
constraint equation have been located, the line can be drawn. The
constraint line and feasible solutions for constraint 2P – A ≤ 100 is
shown in Figure 2.10
Figure 2.10
Summary of the Graphical Solution Procedure
Maximization Projects
●
The steps are summarized as follows:
– Draw a graph of the feasible solutions for all of the constraints.
– Determine the feasible region by identifying the solutions that satisfy
all constraints simulataneously.
– Choose an arbitrary (but convenient) value for the objective function.
– Draw a line on the graph showing the values of the decision variables
that will give this value for the objective function.
– Using a ruler or straightedge, move the objective function line as far
from the origin possible until any further movement would take line
out of the feasible region altogether.
– The feasible solution on this objective function is the optimal solution.
– Confirm the solution point mathematically simultaneous equations.
Slack Variables

●
In addition to optimal solution and its associated profit
contribution, Gulfgolf management will want information
about the production time requirements for each
production operation.
●
How many hours do we actually need to produce the
optimal solution quantities? We can substitute what we
found earlier (S=540, D=252) into the constraints.
●
We can see the next slide.
Slack Variable

Constraint Hours Required Hours Available Unused Hours

Cutting and Dyeing 7/10(540)+1(252)=630 630 0
Sewing 1/2(540) + 5/6(252)=480 600 120
Finishing 1(540) + 2/3(252)=708 708 0
Inspection & 1/10(540) + 1/4(252) = 135 18
Packaging 117

Slack is evident in the sewing and inspection packaging.

Often,, variables called slack variables, are added to the formulation of a linear
programming problem to represent the slack, or idle capacity. Unused contribution
makes no contribution to profit.
Additional Notes:
●
The standard uniform distribution of a linear programming model, the objective
function coefficients for slack variables are zero. This zero coefficient implies that
slack variables, which represent unused resources, do not affect the value of
objective function. However, in some applications, unused resources can be sold
and contribute to profit. In such cases, the corresponding slack variable becomes
decision variables representing the amount of unused resources to be sold. For
each of these variables, a nonzero coefficient in the objective function would
reflect the profit associated with selling a unit of the corresponding resource.
●
Redundant constraints do not affect the feasible region; as a result, they can be
removed from a linear programming model without affecting the optimal solution.
However, if the linear programming model is to be resolved later, changes in
some of the data might make a previously redundant constraint a binding
constraint. Thus, we recommend keeping all constraints in the linear programming
model even though at some point in time one or more of the constraints may be
redundant.
Extreme Points and Optimal Solution
●
Suppose that the profit contribution for the company’s standard
golf bag is reduced from $10 to $5 per bag, while the profit
contribution for the deluxe golf bag and all the constraints remain
unchanged. The complete linear programming model of this new
problem is identical to the mathematical model in Section 2.1,
except for the revised objective function: Max 5S +9D
●
Figure 2.12 shows the graphical solution of this new problem with
the revised objective function. Note that without any change in the
constraints, the feasible region does not change. However, the
profit lines have been altered to reflect the new objective function.
Figure 2.12
Extreme Points and Optimal Solution
●
By moving the profit line in a parallel manner toward higher profit
values, we find the optimal solution in Figure 2.12. the values of the
decision variables at this point are S = 300 and D = 420. The
reduced profit contribution for the standard bag caused a change in
the optimal solution. In fact, as you may have suspected, we are
cutting back the production of the lower-standard bags and
increasing the production of the higher-profit deluxe bags.
●
Notice that the optimal solutions occur at one of the vertices are
referred to as the extreme points of the feasible region. This have
five extreme points. The optimal solution to a linear program can be
found at an extreme point of the feasible region.
Figure 2.13