LINEAR

PROGRAMMING
MODELS




Submitted to: Mr. Bashir Hayat


Submitted by: Group 1


Group Members: Syeda Afsheen Sardar, Sana Hayat, Laiba Maqbool
Section : C
Semester : 6
BBA (Hons)
QTM

2013

2 | P a g e

Linear Programming Models
Linear Programming Models



Contents:
1.Linear Programming Page 3
2. Modeling Page 3
3. Key Elements Page 3
3.1. Decision Variables Page 4
3.2. Objective Page 4
3.3. Constrains Page 5
4. Linear Programming Models Page 5
4.1. Capital Budgeting Model Page 5
4.1.1 Example Page 6
4.2. The Diet-Mix Models Page 11
4.2.1. Example Page 11
4.3. Blending Models Page 13
4.3.1. Example Page 14




3 | P a g e

Linear Programming Models

Linear Programming:
Linear programming is mathematical technique for maximizing or minimizing a
linear function of several variables, such as output or cost. Linear programming is
not only confined to the field of mathematics. It is also applied to other fields such
as human resource management i.e. how many employees should a company hire,
marketing i.e. how much cost should be put in advertising a new product or is it
suitable for a company to enter in new market, resource and development i.e. how
much resources should be allocated and that how much workers should be hired in
the factory, production and manufacture, investment i.e. how much amount should
be invested in different sectors to get maximum profit. Linear programming can
also be applied to the field of sports, health and many other fields.

Modeling:
Linear programming models are found in almost every field of business. For
applying linear programming to these various fields we begin by modeling
problems related to these fields. Modeling a problem using linear programming
involves writing it in the language of linear programming.


Key Elements:
However, in order to transform a problem in the form of model, three key elements
are required:
4 | P a g e

Linear Programming Models

1. Decision Variables

2. Objective


3. Constrains

Decision Variables:
The decision variables represent (unknown) decisions to be made. We represent
these unknown values by x1 and x2 respectively. For example we can say that let
x1 be the number of notebooks produced and x2 be the number of desktops
produced.

Objective:
Every linear program model has an objective. This objective can be either to
minimize a cost or to maximize a certain profit. For example an objective in a
model can be to invest in 6 different industries to get maximum profit.



5 | P a g e

Linear Programming Models
Constrains:
Similarly, every linear program model also has certain constraints within which the
decisions have to be made. For example a student has a budget of Rs.10, 000
within which he has to spend his entire month in hostel. Now this 10,000 is a
constrain within which the student has to spend his entire month in hostel.

Linear Programming Models:
There are different linear programming models; however the following three will
be discussed below i.e.

Capital Budgeting Model
Blending Model
Diet Mix Model

Capital Budgeting Model:
Capital budgeting decisions involves the allocation of limited investment funds
among a set of competing investment alternatives. Usually, the alternatives
available in any given time period are each characterized by an investment cost as
well as some estimated benefits. The investment costs involved in the investment
can be the amount used to purchase a house or in case of buying a car from bank
the investment cost can be insurance fee, the fee of car as well as other documents,
now the determination of the investment costs is relatively easy because we know
how much of the amount we hae invested for a transaction.
6 | P a g e

Linear Programming Models
However, determining the profits can be a little difficult, especially when the
projects in which an individual wants to invest are characterized by less tangible
returns for example projects having social benefits. Typically, the problem which
usually occurs is to select a set of alternatives which would maximize overall
benefits subject to the budgetary constraints and the other constrains which might
affect the choice of projects. Thus, in order to deal with this problem we create
capital budgeting models.

Example:
Suppose there is a company which has a total budget of $ 1 billion to invest in
different areas so that maximum profit can be gained. Now the management team
comprising of the economists as well as staff from finance department review
about 50 fields, among which they chose 6 for investment. Each of the six projects
has been evaluated in relation to potential profits expected over the next 10 years.
These estimated benefits are shown in the table below.
.
Project Project
Classification
Net Benefit per $
Invested
Requested Level
of Funding,
$ Millions
1 Solar 4.4 220
2 Cotton 3.8 180
3 Synthetic Fuels 4.1 250
4 Coal 3.5 150
5 Nuclear 5.1 400
6 Vehicles 3.2 120

The value $ 4.4 represents that for every dollar investment in the field of solar will
result in a net profit of 4.4 million dollars over the next 10 years. Similarly 3.8
7 | P a g e

Linear Programming Models
shows that for every dollar investment in the cotton will result in a net profit of 3.8
million dollars over the next 10 years. 4.1 shows that every dollar investment in
synthetic fuels will give an average profit of about 4.1 million dollars over the next
10 years. 3.5 shows that if we invest one dollar in the field of coal so the average
benefit that we get over 10 years would be 3.5 million dollars. Similarly 5.1 shows
that a dollar investment in the field of nuclear will result in an average profit of 5.1
million dollars. While the value of 3.2 shows that a dollar invested in the field of
vehicles could result in an average profit of about 3.2 million dollars over 10 years.


Now we know that we always have a decision variable so suppose that xj be equal
to the number of dollars invested in project j. So our decision variables for this
problem would be:
x1= number of dollars invested in the field of solar
x2= number of dollars invested in field of cotton
x3= number of dollars invested in field of synthetic fuels
x4= number of dollars invested in field of coal
x5= number of dollars invested in field of nuclear
x6= number of dollars invested in field of vehicles
Similarly we know that capital budgeting deals with a problem to invest in a set of
alternatives in such a way that profit is maximized, so for this purpose we require
an objective function. As objective function comprises of a combination of total
investments made, so it can be written as:
8 | P a g e

Linear Programming Models
Z = 4.4x1 + 3.8x2 + 4.1x3 + 3.5x4 + 5.1x5 + 3.2x6
Where Z represents maximum level of profit.
Similarly there are also certain constraints present as discusses above. The
requested level of funding in the table above shows the maximum amount which
can be invested in any project. The agency can invest any amount up to the
indicated maximum level of investment amount for a given project. So the
constrains for investing in each field are:
x1 220
This constrain represents that the maximum amount which can be invested in the
field of solar in 220 million dollars. The agency cannot spend more than this
amount in this field.
x2 180
This constrain represents that the maximum amount which can be invested in the
field of cotton is 180 million dollars. The agency cannot spend more than this
amount in this field.
x3 250
This shows that the maximum amount which can be invested in the field of
synthetic fuels is 250 million dollars. The agency cannot spend more than this
amount in this field.




9 | P a g e

Linear Programming Models

x4 150
Similarly it shows that the maximum amount which can be invested in the field of
coal is 150 million dollars. The agency cannot spend more than this amount in this
field.
The same interpretation would be done for the remaining two constraints i.e.
x5 400
x6 120
Now the maximum level of funding which is done by the agency is $1 billion
dollars i.e. at most the agency can invest up to 1 billion dollars, it cannot go
beyond this amount, its total investment made in the all the projects should be with
in the budgeted amount i.e. 1 billion dollars. This constrain can be written as
x1 + x2 + x3 + x4 + x5 + x6 1000
Correspondingly, the president has mandated that the nuclear project should be
funded to at least 50 percent of the requested amount which is to be funded for this
project. So this constrain can be written as
x5 50% (400)
x5 0.5 (400)
x5 200
Thus the minimum level of investment which has to be made for the field of
nuclear should be 200 i.e. at least 200 million dollars should be invested in this
field, it should not be less than it.
10 | P a g e

Linear Programming Models
Similarly, the agencys administration has a strong interest in the solar as well as
cotton projects and has requested that the combined amount invested in the solar as
well as cotton project should be at least $300 million. This constrain can be written
as
x1 + x2 300
So the complete formulation of this entire model can be written as
Maximize, Z = 4.4x1 + 3.8x2 + 4.1x3 + 3.5x4 + 5.1x5 + 3.2x6
Subject to, x1 + x2 + x3 + x4 + x5 + x6 1000
x1 200
x2 180
x3 250
x4 150
x5 400
x6 120

x5 200
x1 + x2 300
x1 , x2 , x3 , x4 , x5 , x6 0
The last constrain shows that the no negativity constraint has been included in the
formulation. This ensures that investment which is to be done for each project
should be greater than zero.
11 | P a g e

Linear Programming Models
The Diet-Mix Models:
The classic diet mix problem involves determining the items which should be
included in a meal so as to;
1. Minimize the cost of the meal
2. Satisfying certain nutritional requirements.
The nutritional requirements usually take the form of daily vitamin requirements;
restrictions encouraging variety in the meal e.g. do not serve each person 10
pounds of boiled potatoes and restrictions which consider taste and logical
companion foods. The following example illustrates a simple diet mix problem.
Example:
A dietitian is planning the menu for the evening meal at a university dining hall.
Three main items will be served, all having different nutritional content. The
dietitian is interested in providing at least the minimum daily requirement of each
of the three vitamins in this one meal. The table below shows the vitamin content
per ounce of each type of food, the cost per ounce of each food, the minimum
dailyrequirements (mdr) for the three vitamins. Any combination of the three foods
may be selected as long as the total serving size is at least 9 ounces.
Vitamins
Food 1 2 3 Cost per oz,$
1 50mg 20mg 10mg 0.10
2 30mg 10mg 50mg 0.15
3 20mg 30mg 20mg 0.12
Minimum
daily
requirements
290mg 200mg 210mg

12 | P a g e

Linear Programming Models
This problem is to determine the number of ounces of each food to be included in
the meal. The objective is to minimize the cost of each meal subject to satisfying
minimum daily requirements of the three vitamins as well as the restriction on
minimum serving size.
To formulate the linear programming model for this problem, let xj equal the
number of ounces included of food j. The objective function should represent the
total cost of the meal. Stated in dollars, the total cost equals the sum of the costs of
the three items, or
Z=0.10x1+0.15x2=0.12x3
Since we are interested in providing at least the minimum daily requirement for
each of the three vitamins, there will be three greater than or equal to constraints.
The constraint for each vitamin will have the form
Milligrams of vitamin intake mdr
Milligrams from food 1+milligrams from food2+milligrams from food 3>/mdr
The constraints are, respectively,
50x1+30x2+20x3290 (vitamin1)
20x1+10x2+30x3200 (vitamin2)
10x1+50x2+20x3210 (vitamin3)
The restriction that the serving size be at least 9 ounces is stated as
X1+x2+x39 (minimum serving size)

The complete formulation of the problem is as follows:
13 | P a g e

Linear Programming Models
Minimize subject to; z=0.10x1+0.15x2+0.12x3
50x1+30x2+20x3 290
20x1+10x2+30x3 200
10x1+50x2+20x3 210
X1+x2+x3 9
X1, x2, x3 0
Note that the no negativity constraint has been included in the formulation. This
ensures that negative quantities of any of the foods will not be recommended.

Blending Models
Linear programming has found wide application in an area referred to as blending
models. Blending problems refer to situations in which a number of components
(or commodities) are mixed together to yield one or more products. The problem is
to determine how much of each commodity should be purchased and blended with
the rest. Typically, different commodities are to be purchased. Each commodity has
known characteristics and costs. Blending models have been used in blending of
petroleum products, feed mixes for agricultural use, fertilizers and grass seeds,
spirits, teas and coffees, and so forth. The objective with such models often is to
minimize the cost of the blend. Typical constraints include lot size requirements for
each blend, technological (or recipe) requirements, and limited availability of
component ingredients.
The following example illustrates a blending model:
14 | P a g e

Linear Programming Models
EXAMPLE: A small refinery is about to blend four petroleum products into three
final blends of gasoline. Although the blending formulas are not precise, there are
some restrictions which must be adhered into the blending process.
Theses blending restrictions are as follow:
1. Component 2 should constitute no more than 40 percent of the volume of
blend 1.
2. Component 3 should constitute at least 25 percent of the volume of blend 2.
3. Component 1 should be exactly 30 percent of blend 3.
4. Component 2 and 4, together, should constitute at least 60 percent of the
volume of blend 1.
Further data says that there is limited availability of components 2 and 3,
1,500,000 liters and 1,000,000 liters, respectively. The Production manager wants
to blend a total of 5,000,000 liters. Of this total at least 2,000,000 liters of blend 1
should be produced. The wholesale price per liter from the sale of each final blend
is $0.26, $0.22, and $0.20, respectively. The input component cost $0.15, $0.18,
$0.12, and $0.14 per liter, respectively.


SOLUTION:
In formulating this problem, we will use the double-scripted variable Xij to
represent the number of liters of component i used in final blend j. A major
assumption in this model is that there is no volume loss in the blending process.
15 | P a g e

Linear Programming Models
That is, if 3 liters of component products are combined, the result is a final blend
of exactly 3 liters.

The following figure gives the understanding of the blending relationships:











X11
X12
X13
X21
X22
X23
X31
X32
X33
X41
X42
X43
16 | P a g e

Linear Programming Models
The objective function has the form:
Total profit contribution= total revenue from 3 blends total cost of 4 components
=total revenue from blend 1+ total revenue from blend 2+ total revenue from blend
3 cost of component 1 cost of component 2 cost of component 3 cost of
component 4
=$0.26 (no. of liters of blend 1) + $0.22 (no. of liters of blend 2) + $0.20 (no. of
liters of blend 3) $0.15 (no. of liters of component 1) $0.18 (no. of liters of
component 2) $0.12 (no. of liters of component 3) $0.14 (no. of liters of
component 4)
Examine the expressions in parentheses in the following equations and verify that
the objective function is
Z= 0.26(X11+X21+X31+X41) + 0.22(X12+X22+X32+X42)
+0.20(X13+X23+X33+X43) 0.15(X11+X12+X13) 0.18(X21+X22+X23)
0.12(X31+X32+X33) 0.14(X41+X42+X43)
Which can be simplified by combining like terms to yield:
Z= 0.11X11 +0.07X12 +0.05X13 +0.08X21 +0.04X22 +0.02X23 +0.14X31
+0.10X32 +0.08X33 +0.12X41 +0.08X42 +0.06X43 (total profit)

Regarding structural constraints,
1. the total production run must equal 5,000,000 liters, or
X11+X12+X13+X21+X22+X23+X31+X32+X33+X41+X42+X43= 5,000,000
2. Recipe restriction (1) is represented by the inequality
17 | P a g e

Linear Programming Models
Amount of component 2 used in blend 1 40% of amount of final blend 1
Or X21 0.40 (X11+X21+X31+X41)
This constraint simplifies to the form
0.4 X11 + 0.6 X21 0.4 X31 0.4 X41 0

3. Recipe restriction (2) is stated as
Amount of component 3 used in blend 2 25% of amount of final blend 2
Or X32 0.25 (X12+X22+X32+X42)
This constraint simplifies to the form
0.25 X12 0.25 X22 + 0.75 X32 0.25 X42 0
4. Restriction 3
X13 = 0.30 (X13+X23+X33+X43)
Or 0.7X13 0.3X23 0.3X33 0.3X43 =0
5. Restriction 4
X21 + X41 0.6(X11+X21+X31+X41)
Or 0.6X11 +0.4X21 0.6X31 +0.4X41 0


The limited availabilities of components 2 and 3 are represented by the inequalities
below:
18 | P a g e

Linear Programming Models
X21+X22+X23 1,500,000
X31+X32+X33 1,000,000
Finally, the minimum production requirement for final blend 1 is represented by
X11+X21+X31+X41 2,000,000

The complete formulation of this blending model is
Maximize:
Z= 0.11X11 +0.07X12 +0.05X13 +0.08X21 +0.04X22 +0.02X23 +0.14X31
+0.10X32 +0.08X33 +0.12X41 +0.08X42 +0.06X43
Subject to:
X11+X12+X13+X21+X22+X23+X31+X32+X33+X41+X42+X43= 5,000,000
0.4 X11 + 0.6 X21 0.4 X31 0.4 X41 0
0.25 X12 0.25 X22 + 0.75 X32 0.25 X42 0
0.7X13 0.3X23 0.3X33 0.3X43 =0
0.6X11 +0.4X21 0.6X31 +0.4X41 0
X21+X22+X23 1,500,000
X31+X32+X33 1,000,000
X11+X21+X31+X41 2,000,000
X11, X12, X13, X21, X22, X23, X31, X32, X33, X41, X42, X43 0