CATHOLIC UNIVERSITY OF SANTA MARIA

FACULTY OF PHYSICAL SCIENCES AND ENGINEERING

FORMAL
Professional School of Industrial Engineering

Consultation Material

Operational Research I

Made by:
Ing. Efraín Rafael Murillo Quispe

AREQUIPA–PERU

2019
Operations Research I 2

Introduction
Optimization, - in Engineering -, is synonymous with reduction of
costs, times, and other resources; therefore it is also
greater productivity, improvement in the quality of goods and
services; and in this way increase the competitive level of
the company.
The Optimization Techniques is based on the
Operations Research, Statistics, and Computer Science,
mainly.
This set of tools will allow management to have
a good support, for better decision making.
Many projects become very attractive after a
reengineering of its processes, fundamentally due to the–
optimization of its operations.
Finalmente, no debemos perder el foco principal de la
application of these techniques: -total satisfaction of the
client-, emphasizing that the modern organization has
focusing all your efforts on the customer.

Ing. Efraín Murillo Msc.

Operations Research I 3

1. Professional Profile
The Teacher presents the following professional profile:

Principal Professor of the National University of San Agustín

Arequipa (UNSA).

Assistant Professor at the Catholic University of Santa María

Arequipa (UCSM)

Industrial Engineer with a second specialty in Systems Engineering from the University
National University of San Agustin, Master in Industrial Engineering with a specialization in Management
Production of the National University of San Agustín in collaboration with the University
Federal University of Santa Catarina in Brazil. Doctorate in Production Engineering from the University
National University of San Agustín in agreement with the Federal University of Santa Catarina in Brazil.
Specialized consultant in the development of Optimization Projects through the use of
mathematical modeling and simulation. Teacher at the National University of San Agustín and
from the Catholic University of Santa María in Industrial Engineering. General Coordinator of the
Second Specialization in Systems Engineering from the Faculty of Production Engineering
and Services of the UNSA. Responsible for the promotion for the creation of the Center of
Investigación en Modelación y Simulación en la UNSA y en la UCSM. Ponente en congresos
national and international. Professional experience in optimization in companies
Centromin-Peru and Southern Peru Brewery Co.
Email:emurillo@[Link]
efrainmurillo@[Link]
Website:[Link]

THESIS:
FEASIBILITY STUDY FOR THE INDUSTRIALIZATION OF COCHINEAL. Thesis to obtain
the Title of Industrial Engineer. Arequipa 1994.
BINARY PROGRAMMING MODEL TO OPTIMIZE PROGRAMMING OF
BUSES ON A PASSENGER URBAN TRANSPORT ROUTE IN AREQUIPA, Thesis
to obtain the Master's degree in Industrial Engineering. Arequipa 2005.
DECISION SUPPORT SYSTEM BASED ON MATHEMATICAL MODELING FOR
THE OPTIMIZATION OF VEHICLE REPLACEMENT POLICIES IN COMPANIES OF
URBAN PASSENGER TRANSPORT OF AREQUIPA, Thesis to obtain the Title of Second
Specialty in Systems Engineering. Arequipa 2016.
SOLID WASTE MANAGEMENT MODEL BASED ON BINARY PROGRAMMING AND
SIMULATION TO REDUCE LEAD TIME, CASE: MUNICIPALITY OF JOSÉ LUIS
BUSTAMANTE AND RIVERO, Thesis to obtain the Degree of Doctor in Production Engineering.

Ing. Efraín Murillo Msc.

Operations Research I 4

Classification of IO Models, according to the nature of the data:

Source:[Link]-operaciones.com2009

Ing. Efraín Murillo Msc.

Operations Research I 5

Index
Page
1. Formulation of linear programming problems 6
[Link] programming: Graphic approach 20
[Link] analysis: Graphical approach 46
[Link] analysis: Computer solution 56
[Link] analysis 69
[Link] Method 75
7. Case Study 80
8. Problem bank 87

Ing. Efraín Murillo Msc.

Operations Research I 6

1 Formulation of Problems of
Linear Programming
. Taken from ACCOUNTS, Mario. Operations Research, Peru: UNA Editorial Group,
2009.

1.1 LINEAR PROGRAMMING MODELS FOR PRODUCT MIX DECISIONS

It is often necessary to determine how to allocate various scarce resources such as labor, materials
first, and the capital, to several alternatives that compete for these resources. The final decision is based on the
availability of these resources and in achieving a global objective for the organization. For example in
production problems, the mix of products to be manufactured is ultimately based on a corporate objective
global like the maximization of profits or the minimization of total production costs. The models of
Linear programming is often used to help make such decisions. Consider the problem faced
by the management of the company ABC.
Example 1.1 The ABC PRODUCT MIXTURE PROBLEM
ABC manufactures three rubber products: Airtex (Foam material), Extendex (Elastic material), and Resistex.
(Rigid material). The three products require the same three chemical polymers and a base. The amount of
ingredient loss used per pound of final product is shown in table 1.1

TABLE 1.1 Ingredients used in the production of Airtex, Extender and Resistex
INGREDIENTS (oz/lb of products)
POLYMER A POLYMER B POLYMER C BASE
Airtex 4 2 4 6
Extender 3 2 2 9
Resistex 6 3 5 2
ABC is committed to producing at least 1000 pounds of Airtex, 500 pounds of Extendex, and 400 pounds of
Resitex for next week, but the company management knows that it can sell more of each one of the
products.
The current inventories of the ingredients are 5600 pounds of polymer A, 425 pounds of polymer B, 650 pounds
of polymer C and 1100 pounds of the base. Each pound of Airtex generates a profit of $7 for the company, each
A pound of Extendex has a profit of $7 and each pound of Resistex has a profit of $6. As the manager of
production department, you need to determine an optimal production plan for this week.
1.1.1 Identification of the decision variables
Following the steps of problem formulation, first identify the decision variables. Ask yourself what
what can be controlled and the information that constitutes a production plan, this should lead to identifying the
following variables:
A = the number of pounds of Airtex to be produced this week
E = the number of pounds of Extender produced this week
R = the number of pounds of Resistex to produce this week
1.1.2 Identification of the objective function
For ABC, the logical objective is to determine how much of each product to produce to maximize total profit.
By applying the decomposition technique, one arrives at:
Ganancia total = ganancia de Airtex + ganancia de Extender + ganancia de Resistex

Ing. Efraín Murillo Msc.

Operations Research I 7

Since each pound of Airtex produces a profit of $7, A pounds of Airtex produce $7A. Similarly,
Extender and Resistex contribute $7E and $6R, respectively, to the total profit. In terms of the
decision variables and profit data, the objective function is
Maximize 7A + 7E + 6R
1.1.3 Identification of the restrictions
Applying the clustering technique should lead you to identify the following three groups of constraints:
1. Resource restrictions to ensure that no more than the three polymers and the base that are used.
available.
2. Demand restrictions to ensure that the company's commitments are met.
3. Logical restrictions to specify that all production quantities are non-negative.
RESOURCE RESTRICTIONS
This group consists of four constraints: one for each of the three polymers and one for the base. For the
limited availability of 500 pounds of polymer A:
Amount used of polymer A 500 pounds
The use of decomposition leads to:
Amount of polymer A used = (amount used to produce A pounds of Airtex) +
(amount used to produce E pounds of Extender) +
(amount used to produce R pounds of Resistex)
To determine the amount of polymer A used in the manufacturing of each product, work with an example.
specific. For example, set A = 100, E = 300 and R = 200. According to the data in table 1.1:
Amount of polymer A used in Airtex = 4(100) = 400
Amount of polymer A used in Extender = 3(300) = 900
Amount of polymer A used in Resistex = 6(200) = 1200
So, in terms of the decision variables, one could think that the appropriate constraint for polymer A
es:
4A + 3E + 6R 500
Sin embargo, esta restricciónnoes correcta. La razón es que lasunidadesen la expresión de la izquierda están
weights (see table 1.1), but the units on the right are in pounds. This discrepancy can be corrected.
converting units from one side to the other. For example, when converting 500 pounds
available from polymer A to 8000 ounces (1 pound is equal to 16 ounces) the following restriction is obtained:
4A + 3E + 6R 8,000 (polymer A)
Following a similar logic for the three remaining resource outcomes in these constraints:
2A + 2E + 3R 6,800 (polymer B
4A + 2E + 5R 10 400 (polymer C)
6A + 9E + 2R 17,600 (base)
DEMAND RESTRICTIONS
This group consists of three constraints: one for the minimum requirement on the quantity of each of the
three products. These restrictions are:
A 1 000 (Airtex)
E 500 (Extender)
R 400 (Resistex)
LOGICAL RESTRICTIONS
Since all production quantities must be non-negative, the following restrictions are needed.
logics:
A, E, R 0
1.1.4 Complete formulation and solution of the ABC product mix problem.

Ing. Efraín Murillo Msc.

Operations Research I 8

As the production department manager, you put all the pieces together, resulting in the following model.
mathematical formulation of the linear programming problem of ABC:

Maximize 7A + 7E + 6R
Depending on
RESOURCE RESTRICTIONS
4A + 3E + 6R 8,000 (polymer A)
2A + 2E + 3R 6,800 (polymer B
4A + 2E + 5R 10 400 (polymer C)
6A + 9E + 2R 17,600 (base)
DEMAND RESTRICTIONS
A 1 000 (Airtex)
E 500 (Extender)
R 400 (Resistex)
LOGICAL RESTRICTIONS
A, E, R 0
The optimal solution to this problem, calculated using any linear programming software package is
A = 1,000.00
E = 533.33
R = 400.00
With an objective function value of 13,133.33. In other words, the optimal weekly plan is to produce 1000 pounds.
from Airtex, 533.33 pounds of Extender and 400 pounds of Resistex, resulting in a net profit of
$13133.33. You may want to verify this solution with your own linear programming software.

1.2 LINEAR PROGRAMMING MODELS FOR MANUFACTURING DECISIONS OR

BUY
In many production problems, a company may not have enough resources to meet
an unexpectedly large demand for one or more products. In such cases, the company may supplement
its production capacity by buying some of the products from external suppliers. The central point in
These situations involve the decision by the administrators on the quantity of each product to be produced.
against the quantity that needs to be purchased from abroad. A linear programming model is often useful in the
making such decisions, as illustrated by the following example.
EXAMPLE 1.2 THE PROBLEM OF MAKING OR BUYING FROM ACEROS AREQUIPA. Aceros Arequipa
producing three sizes of pipes: A, B, and C, which are sold for $10, $12, and $9 per foot, respectively. For each
The foot of tube A requires 0.5 minutes of processing time on a particular type of machine.
modeling. Each foot of tube B requires 0.45 minutes and each foot of tube C requires 0.6 minutes. After the
production, each foot of pipe, regardless of type, requires 1 ounce of welding material. The total cost is estimated
at $3, $4, and $4 per foot for pipes A, B, and C, respectively.
For the following week, Aceros Arequipa has received exceptionally large orders totaling 2000.
feet of pipe A, 4000 feet of pipe B and 5000 feet of pipe C. Since only 40 hours of time is available
machine this week and there are only 5500 ounces of welding material in inventory, the department of
production will not be able to meet this demand, which requires a total of 97 hours of machine time and 11000
ounces of welding material. This high level of demand is not expected to continue. Instead of expanding the
production facilities capacity, the management of Aceros Arequipa is considering the purchase of
some of these tubes to suppliers in Japan at a delivery cost of $6 per foot of tube A, $6 per foot of tube
B at $7 per foot of pipe C. These various data are summarized in Table 1.2. As the manager of the department of
production, recommendations have been requested regarding the quantity of production of each type of tube and
the amount of purchase to Japan to satisfy the demand and maximize the profits of the company.

Ing. Efraín Murillo Msc.

Operations Research I 9

TABLE 1.2 Data for the make or buy problem of Aceros Arequipa
PRICE TIME OF MATERIAL COST OF COST OF
DEMAND
FOR SALE MACHINE FOR WELDING PRODUCTION BUY
(feet)
($/pie) (min/pie) (onz/pie) ($/pie) ($/pie)
A 10 2,000 0.50 1 3 6
B 1 4,000 0.45 1 4 6
C 9 5,000 0.60 1 4 7
Available quantity 40 hrs. 5,500 ounces

1.2.1 Identification of the decision variables

In this problem, you have the freedom to choose how many feet of each type of pipe to produce and how many feet to buy.
Japan. This results in the following six decision variables:
AP = number of feet of type A tube to produce
BP = number of feet of type B pipe to be produced
CP = number of feet of type C pipe to produce
AJ = number of feet of type A tube to be purchased from Japan
BJ = número de pies de tubo de tipo B por comprar a Japón
CJ = número de pies de tubo de tipo C por comprar a Japón
1.2.2 Identification of the objective function
As established in the problem description, the overall objective is to maximize total profits. If
we apply the decomposition we obtain:
Total earnings = (production profits) +
(profits from products purchased from Japan)

If we apply decomposition to production earnings, we have:

Production profits = (profits from producing type A tube) +
(profits from producing type B tube) +
profits from producing type C tube
Each of these gains, in turn, is calculated as income minus the cost per foot. For example, since the
Type A pipes are sold at $10 per foot but their production costs $3, the net profit is $7 per foot. Therefore,
The profit from producing type A tube feet is 7AP. A similar calculation for tubes of types B and C.
it results in:
Production profits = 7AP + 8BP + 5CP
Applying a decomposition and logic similar to products purchased from Japan, we have:
Profits from products purchased from Japan = 4AJ + 6BJ + 2CJ
As expected, each foot of tube produced results in a higher profit than each foot of tube.
purchased from the external supplier. The combination of these two profit components results in the following
global objective function:
Maximize 7AP + 8BP + 5CP + 4AJ + 6BJ + 2CJ
1.2.3 Identification of the restrictions
The application of the clustering technique should lead you to identify the following three groups of
restrictions:
1. Resource constraints to ensure that the amount of machine time and material for
solder used does not exceed the available supplies.
2. Demand constraints to ensure that the demand for each type of tube is met.
3. Logical restrictions.
RESOURCE RESTRICTIONS

Ing. Efraín Murillo Msc.

Operations Research I 10

Producing these tubes requires two resources: machine time and welding material. Since these resources are
limited require two restrictions to ensure that the available supplies are not exceeded. The
the available time restriction is:
Machine time should not exceed 40 hours.
The application of decomposition leads to:
Total machine time used = (machine time used to produce type A tube) +
(machine time used to produce type B tube) +
(machine time used to produce type C tube)
Remember from table 1.2 that each foot of tube A requires 0.5 minutes of machine time. Therefore, for
To produce AP feet, 0.5 AP minutes is required. Likewise, each foot of tube B requires 0.5 minutes and each
Pipe fitting C takes 0.6 minutes. The constraint is:
0.5AP + 0.45BP + 0.6CP 40
However, note that the amount on the left side is expressed in minutes, while that on the right is
express in hours. One way to correct this inconsistency is to convert 40 hours into 40 * 60 = 2400 minutes:
0.5AP + 0.45BP + 0.6CP 2400 (machine time)
Returning to the availability of welding material, the associated restriction is:
The total welding material must not exceed 5500 ounces.
Applying decomposition and remembering that every foot of pipe, regardless of type, requires 1 ounce of material.
for welding, this resource restriction is:
AP + BP + CP 5500 (soldering material)
DEMAND RESTRICTIONS
This group is made up of three constraints, one for the demand associated with each type of pipe. For the
Pipe A:
Total number of feet of type A pipe = 2000 feet
Applying decomposition:
Total number of feet (number of type A feet produced) +
of type A tube (number of type A feet purchased from Japan)
=AP+AJ
Consequently, the demand constraint for type A tube is:
AP + AJ = 2000 (demand of type A)
A similar logic results in the following demand constraints for pipes of type B and C:
BP + BJ = 4000 (demand of type B)
CP + CJ = 5000 (demand of type C)
LOGICAL RESTRICTIONS
The only logical restriction in this problem is that all variables must be non-negative.
1.2.4 Complete formulation and solution of the manufacturing or purchasing problem of Aceros Arequipa
Once all the pieces are put together, it results in the following linear programming model for the
Aceros Arequipa problem:
maximize 7AP + 8BP + 5CP + 4AJ + 6BJ + 2CJ
Depending on
DEMAND RESTRICTIONS
AP +AJ = 2000 (demand of type A)
BP + BJ = 4000 (demand of type B)
CP + CJ = 5000 (demand of type C)

Ing. Efraín Murillo Msc.

Operations Research I 11

RESOURCE RESTRICTIONS
0.5AP + 0.45BP + 0.6CP 2400 (machine time)
AP+ BP + CP 5500 (soldering material)
LOGICAL RESTRICTIONS
AP, BP, CP, AJ, BJ, CJ 0
The optimal solution to this problem, obtained with a linear programming software package, is:
AP = 2 000.000
BP = 0.000
CP = 2 333.000
AJ= 0.000
BJ = 4 000.000
CJ = 2 666.667
With a net profit of $55,000. In other words, Aceros Arequipa should produce 2,000 feet of pipe.
type A and 2333.333 feet of type C pipe and import 4000 feet of type B pipe and 2666.667 feet of type C pipe from
Japan. You may want to verify this solution with your own linear programming software.

1.3 LINEAR PROGRAMMING MODELS FOR DIET PROBLEMS

Linear programming models can also be applied in diet planning. In particular, given a
number of food alternatives, each of which produces a known amount of a nutrient
necessary, you want to determine how much of each type of food to include in a diet to ensure a
minimum nutrient requirements, while achieving an overall goal. The following example illustrates
a problem of this type.
EXAMPLE 1.3 THE DIET PROBLEM OF XYZ GENERAL HOSPITAL. The Department of
Nutrition of XYZ General Hospital prepares 30 dinner menus, one for each day of the month. A meal
consisting of spaghetti, turkey, scalloped potatoes, spinach, and apple pie. As director of
Department of Nutrition, you have determined that this meal must provide 63,000 milligrams (mg)
of protein, 10 mg of iron, 15 mg of niacin, 1 mg of thiamine, and 50 mg of vitamin C. Every 100 grams of
this food provides the amount of each nutrient and fats indicated in table 1.3
TABLE 1.3 Nutrients provided by different meals
NUTRIENTE (mg/100 g)
PROTEINAS HIERRO TIACINA TIAMINA VITAMINA C GRASA
Spaghetti 5,000 1.1 1.5 0.18 0.0 5,000
Peacock 29,300 1.8 5.4 0.06 0.0 5,000
Fathers 5,300 0.5 0.9 0.06 10.0 7,900
Spinach 3,000 2.2 0.5 0.07 28.0 300
Apple pie 4,000 1.2 0.6 0.15 3.0 14,300
To avoid too much of one type of food, no more than 300 grams of spaghetti should be included in it.
300 grams of turkey, 200 grams of potatoes, 100 grams of spinach, and 100 grams of apple pie.
director of the nutrition department, you wish to determine the composition of a meal that satisfies the
nutritional requirements and provides the minimum amount of fats.
1.3.1 Identification of decision variables
In this problem, you can control the amount of each of the five foods to include in the meal.
which leads to defining the following five variables:
SPAG = the number of 100 grams of spaghetti to include
PAVO = the number of 100 grams of turkey to include
PAPA = the number of 100 grams of potatoes to include
SPIN = el número de 100 gramos de espinaca que incluir
MANZ = the number of 100 grams of apple pie to include
For convenience, it has been chosen that the units of the variables are given in hundreds of grams because those are
the units used in table 1.3.

Ing. Efraín Murillo Msc.

Operations Research I 12

1.3.2 Identification of the objective function

As established in the problem description, the overall objective is to minimize the total fat content.
from the diet. Applying the decomposition results to the following:
Total fat content = (fat contributed by the spaghetti) +
(fat contributed by the turkey) +
(fat provided by the potatoes) +
(fat contributed by spinach) +
fat provided by the apple pie
If you use the data from the last column of table 1.3 and work with a specific example, you should be able to identify the
next global goal:
Minimize 5000SPAG+ 5000TURKEY+ 7900POTATO+ 300PIN+ 14300APPLE
1.3.3 Identification of constraints
The application of the clustering technique leads you to the following three groups of constraints:
Nutrient restrictions to ensure that food provides the minimum amount of each nutrient.
1. Limit restrictions to ensure that too much of one type of food is not included.
for example, ask a patient to eat 1000 grams of spinach
2. Logical constraints to ensure that all variables are non-negative.
Nutrient Requirements
This group consists of five constraints, one to ensure the minimum quantity of each of the five.
nutrients. Consider the protein requirement:
Total amount of protein in the food 63,000 mg
Applying decomposition:
Total amount of protein in the food = (amount of protein in the spaghetti) +
(amount of turkey protein) +
(amount of protein in potatoes) +
(amount of proteins in spinach)
(amount of protein in apple pie)
Refer to the first column of table 1.3. Every 100 grams of spaghetti contains 5000 mg of protein.
Therefore, SPAG one hundred grams of this food provides 5000 SPAG mg of protein to the food.
In a similar way, using the remaining data from the first column of table 1.3 results in the following.
restriction for proteins:
5000 SPAG + 29300 PAVO + 5300 PAPA + 3000 SPIN + 4000 MANZ 63,000 (proteins)
Although the units of the variables are expressed in hundreds of grams, the units on both sides of the
previous restriction is in milligrams.
Using the following four columns of data from table 1.3, we obtain the following similar constraints.
for each of the following four nutrients:
1.1 SPAG + 1.8 PAVO + 0.5 PAPA + 2.2 SPIN + 1.2 MANZ 10 (iron)
1.4 SPAG + 5.4 PAVO + 0.9 PAPA + 0.5 SPIN + 0.6 MANZ 15 (niacin)
0.18 SPAG + 0.06 PAVO + 0.06 PAPA + 0.07 SPIN + 0.15 MANZ 1 (thiamine)
10 DADDY + 28 SPIN + 3 MEN 50 (vitamin C)
LIMIT RESTRICTIONS
These restrictions limit the maximum amount of each type of food in the meal. Keeping in mind that the
The units of the variables are in hundreds of grams, the following limit constraints arise:
SPAG 3
PEACOCK 3
DAD 2
SPIN 1

Ing. Efraín Murillo Msc.

Operations Research I 13

MANZ 1
LOGICAL RESTRICTIONS
The only logical restriction in this problem is that all variables are non-negative.
1.3.4 Complete formulation and solution of the diet problem of XYZ General Hospital
All this information results in the following linear programming model for the problem of
General Hospital XYZ:
Minimize
5000SPAG+ 5000PAVO+ 7900PAPA+ 300SPIN+ 14300MANZ
Depending on
NUTRIENT REQUIREMENT
5000 SPAG + 29300 PAVO + 5300 PAPA + 3000 SPIN + 4000 MANZ 63,000 (proteins)
1.1 SPAG + 1.8 PAVO + 0.5 PAPA + 2.2 SPIN + 1.2 MANZ 10 (iron)
1.4SPAG+ 5.4 PAVO+ 0.9 PAPA+ 0.5 SPIN+ 0.6 MANZ 15 (niacin)
0.18SPAG+ 0.06 PAVO+ 0.06 PAPA+ 0.07 SPIN+ 0.15 MANZ thiamine
10DAD+ 28 SPIN+ 3MANZ 50 (vitamin C)

LIMIT RESTRICTIONS
SPAG 3
PEACOCK 3
DAD 2
SPIN 1
MANZ 1
LOGICAL RESTRICTIONS
["SPAG","PAVO","PAPA","SPIN","MANZ"] 0
The optimal solution to this problem, obtained with a linear programming software package, is:
SPAG = 3.000
PAVO = 2.833
PAPA = 2 000
SPIN = 1.000
MANZ = 0.667
With a fat content of 54,800 milligrams. In other words, the food should consist of 300 grams.
of spaghetti, 283.3 grams of turkey, 200 grams of potatoes, 100 grams of spinach, and 66.7 grams of cake
apple. You may want to verify this solution with your own linear programming software.

1.4 LINEAR PROGRAMMING MODELS FOR PORTFOLIO MANAGEMENT

VALUES
The decision in this type of problem is to determine which investments to select. This problem requires
make a decision 'no/yes' that will result in a whole programming model with 0-1 variables. How will it be
Now, an investment decision may also require determining how much to invest in each alternative.
available. A linear programming model can often be formulated for a problem of this nature.
The overall goal of an investor is to achieve the highest possible return. However, a high return has
a price: the risk. An investor must balance return against risk. It can often be formulated
a linear programming model to design an investment strategy that achieves the maximum return, to
same time that meets certain risk requirements. Consider the problem faced by the partners
General information of AFP JR.

EXAMPLE 1.4 THE INVESTMENT PROBLEM OF AFP JR. To the portfolio manager of AFP JR, he has
request to invest $1,000,000 from a large pension fund. The research department of Investments has

Ing. Efraín Murillo Msc.

Operations Research I 14

six mutual funds with variable investment strategies were identified, resulting in different returns
potential and associated risks, as summarized in table 1.4.
TABLE 1.4. Risk and expected rate of return of six investment funds
FUND
1 2 3 4 5 6
Asset ($/share) 45 76 110 17 23 22
Expected return (%) 30 20 15 12 10 7
Risk category ["High","High","High","Medium","Medium","Low"]

One way to control risk is to limit the amount of money invested in various funds. For that purpose, the
AFP JR management has specified the following guidelines:
The total amount invested in high-risk funds should be between 50 and 75% of the portfolio.
2. The total amount invested in medium-risk funds should be between 20 and 30% of the portfolio.
3. The total amount invested in low-risk funds must be at least 5% of the portfolio.
A second way to control risk is to diversify, that is, to spread the risk by investing in many
different alternatives. The management of AFP JR has specified that the amount invested in high funds
risk 1, 2 and 3 must be at the rate of [Link] respectively. The amount invested in the medium funds
Risk 4 and 5 must be 1:2.
With these guidelines, which portfolio should you, as a portfolio manager, recommend to maximize the expected rate of return?
return?
1.3.3 Identification of the decision variables
In this problem, you control how much to invest in each of the six mutual funds, thereby creating six.
decision variables. As always, you must specify the units associated with each variable. For example,
For fund 1, I could define any of the following variables:
F1the number of shares of the fund 1 to buy
F1the number of dollars to invest in fund 1
F1the fraction of the agenda for investing in fund 1
Each option leads to a different but equivalent mathematical model. Here, the last option is used. So
for each of the remaining funds, define:
F2the fraction of the portfolio to invest in fund 2
F3the fraction of the portfolio to invest in fund 3
F4the fraction of the portfolio to invest in fund 4
F5the fraction of the portfolio to invest in fund 5
F6the fraction of the portfolio to invest in fund 6
1.4.2 Identification of the objective function.
As stated in the problem description, the overall objective is to maximize the expected rate of
performance, that is,
Expected rate of return amount invested = total expected return
If we apply decomposition to the numerator, we obtain
Expected total return = (expected return of fund 1) +
(expected return of fund 2) +
(expected return of fund 3) +
(expected return of the fund 4) +
(expected return of the fund 5) +
(expected return of the fund 6)
To determine the expected return of fund 1, work with a specific example in which 10% of the
the portfolio is invested in fund 1, that is, F1= 0.10. In this case, 0.10 * 1,000,000 = $100,000 from the portfolio is
invest in fund 1. According to the data in table 1.4, this money is expected to return 30% or
0.30*1000000 = $30,000. Therefore, in terms of F1,
Rendimiento esperado del fondo 1 = (amount invested in fund 1) *
fund 1 yield rate

Ing. Efraín Murillo Msc.

Operations Research I 15

= (F1300000
= 300,000 F1
Using a similar logic for the remaining five funds, we arrive at
Expected total return 300,000 F1+ 200,000 F2+ 150,000 F3+
120 000 F4+ 100 000 F5+ 70 000 F6
Dividing this by the total investment of $1,000,000 we obtain the rate of return and therefore the following
objective function:
Maximize 0.30 F1+ 0.2 F2+ 0.15 F3+ 0.12 F4+ 0.10 F5+ 0.07 F6
1.4.3 Identification of the restrictions
By applying the clustering technique, you should be able to identify the following three groups of constraints:
1. Investment limitations to control the amount invested in each of the three categories of
risk.
2. Diversification restrictions to extend investment within each risk category.
3. Logical restrictions.
INVESTMENT RESTRICTIONS
This group consists of three subgroups of restrictions, one for each risk category, namely:
The total amount invested in high-risk funds should be between 50 and 75% of the portfolio. 1, F2
y F3they represent the fraction of the portfolio to be invested in high-risk funds, the fraction of the portfolio
total investment in high-risk funds is F1+ F2+ F3These restrictions are:
F1+ F2+ F3 0.50 (minimum at high risk)
F1+ F2+ F3 0.75 (maximum at high risk)
2. The total amount invested in medium-risk funds should be between 20 and 30% of the portfolio.
F4y F5represent the fraction of the portfolio to be invested in medium-risk funds, the fraction of the
the total portfolio invested in medium risk funds is F4+ F5These restrictions are:
F4+ F5 0.20 (minimum in medium risk)
F4+ F5 0.30 (maximum in medium risk)
3. The total amount invested in low-risk funds must be at least 5% of the portfolio. As F 6it is the
fraction of the portfolio invested in low-risk funds, this restriction is:
F6 0.05 (minimum in low risk)
DIVERSIFICATION RESTRICTIONS
This group of restrictions is used to control risk by ensuring that the amount invested in the funds
belonging to a given risk category falls within the specified rate, in the following manner:
The amount invested in high-risk funds 1, 2, and 3 must be in the ratio [Link]. This restriction
specify that the amount invested in fund 2 is double the amount invested in fund 1:
F2= 2F1
If we change the order so that all the variables are on the right, we obtain:
-2F1+ F2= 0 (proportion of F1a F2)
Similarly, the amount invested in fund 3 must be three times the amount invested in fund 1:
F3= 3F1
-3F1+ F3= 0 (proportion of F1a F3)
The amount invested in medium-risk funds 4 and 5 must be in the proportion of 1:2, that is,
the amount invested in fund 5 must be double that of fund 4:
F5= 2F4
If we change the order so that all the variables are on the right, we obtain:

Engineer Efraín Murillo Msc.

Operations Research I 16

-2F4+ F5= 0 (proportion of F)4a F5)

LOGICAL RESTRICTIONS
It is clear that a set of logical constraints is that each variable must be non-negative. Likewise, as it is
It is possible to buy fractional shares of a mutual fund; these variables are allowed to have any value.
fractional, which results in a linear programming problem, moreover, another logical constraint is required
to ensure that the total portfolio of precisely $1,000,000 is invested. As the decision variables
they represent the fraction of this portfolio for investing in various funds, this restriction is:
The total fraction of $1,000,000 invested must equal 1.
ó
F1+ F2+ F3+ F4+ F5+ F61.0 (total agenda)
1.4.4 Complete formulation and solution of the AFP JR investment problem.
Below is the complete linear programming model for the general partners of AFP JR:
Maximize
0.30 F1+ 0.2 F2+ 0.15 F3+ 0.12 F4+ 0.10 F5+ 0.07 F6
Depending on
INVESTMENT LIMITATION RESTRICTIONS
F 1+ F 2+ F3 0.50 (minimum in high risk)
F 1+ F 2+ F3 0.75 (maximum in high risk)
F 4+ F5 0.20 (minimum in moderate risk)
F 4+ F5 0.30 (maximum in medium risk)
F6 0.05 (minimum in low risk)
DIVERSIFICATION RESTRICTIONS
-2F1+ F2 =0 (proportion of F1a F2)
-3F1+ F3 =0 (proportion of F1a F3)
-2F4+ F5= 0 (proportion of F4a F5)
LOGICAL RESTRICTIONS
F 1+ F 2+ F 3+ F 4+ F5+ F6= 1.0 (total wallet)
F 1, F 2, F 3, F4, F5, F6 0
The optimal solution to this problem that any linear programming software package produces is:
F1= 0.1250
F2= 0.2500
F30.3750
F4= 0.0667
F5 0.1333
F6= 0.0500
With a return rate of 0.168583. In other words, the amount of money invested in each of the
six funds is:
Amount in the fund 1 = 0.1250 * 1 000 000 = $ 125 000
Quantity in the fund 2 = 0.2500 * 1,000,000 = $250,000
Quantity in the fund 3 = 0.3750 * 1,000,000 = $375,000
Amount in the fund 4 = 0.0667 * 1 000 000 = $ 66,700
Amount in the fund 5 = 0.1333 * 1 000 000 = $ 133 300
Amount in the fund 6 = 0.0500 * 1 000 000 = $ 50,000
Total investment $1,000,000
With an expected return rate of 16.86% (or $168,600). You may want to verify this solution with your
own linear programming software.
Remember that the decision variables are defined as the fraction of the portfolio to invest, rather than the amount.
of dollars. This approach has a clear advantage; if the amount of dollars in the portfolio changes, an event

Ing. Efraín Murillo Msc.

Operations Research I 17

likely, the current model remains unchanged. It simply needs to multiply the fractions obtained in the
previous solution due to the new size of the portfolio to determine the new amounts to invest in each one
of the six funds.

Problems for formulation:

1.-A furniture production company has two different types of wood; it has 1500 board feet of
type A and 1000 of type B (type B wood is imported), also has 800 man-hours available to carry out the
work. The estimated weekly demand is as follows: at least 40 tables, 130 chairs, 30 desks.
and no more than 10 bookshelves. The quantities of wood A and B, the man-hours required for the making of each
unit of article and the unit utilities are indicated in the following table:
Wood Hours Demand Utilities
Article A B Dear Man per unit
Table 5 2 3 no less than 40 $ 12
Chair 1 3 2 no less than 130 5
Desk 9 4 5 no less than 30 15
Bookshelf 12 1 10 no more than 10 10
Weekly availability 1500 1000 800
Formulate a Linear Programming Mathematical Model to determine how much should be produced weekly.
the company of each item, so that the profits are maximized?
2.-A company produces three types of products: A, B, and C. The man-hours required are H1 and H2 in the
departments 1 and 2 respectively for product A. One unit of B requires H3 and H1 hours in the
departments 2 and 3 respectively and a product C requires H3, H2, and H1 hours in departments 1, 2 and
3 respectively. The hourly wage is 'J' in any department. It is desired to manufacture up to a maximum of
FT units for the three products; a maximum of F1 units between products A and B and a minimum of F2
units of C. If products A, B, and C are sold at V1, V2, and V3 per unit respectively and the
Availability of hours are HA, HB and HC for departments 1, 2 and 3 respectively. Formulate a model.
Linear Programming for this problem.
3.- Three classifications of workers (w1, w2, and w3) can be used in three different jobs (j1, j2, and j3), according to
a salary scale. Each worker has a different cost (in $) for each job, which appears in the table
next:
Jobs Workers
Worker j1 j2 j3 available
w1 4.0 4.0 4.6 10
w2 3.6 3.5 4.4 15
w3 3.7 4.4 4.6 10
Workers
required 5 20 10
Formulate a Mathematical Model to determine the optimal assignment of workers to various jobs.
in order to minimize the cost.

XYZ has three power generation plants that supply the energy required to four
cities. Each plant can supply the following amounts of kilowatt-hours (kwh) of electrical energy: the
Plant 1, 35 million; plant 2, 50 million; plant 3, 40 million. The maximum energy demands in
these cities, which are presented at the same time (2 p.m.), are the following (in kwh): city 1, 40
millions; city 2, 20 million; city 3, 30 million; city 4, 30 million. The costs to send 1
million kWh of energy from a plant to a city depends on the distance the energy has to travel (see
Table 1). Formulate a Linear Programming model that allows determining the supply plan that minimizes
the cost to meet the maximum energy demand of each city.

Ing. Efraín Murillo Msc.

Operations Research I 18

Table 1
TOWARDS
Ciudad 1 Ciudad 2 Ciudad 3 Ciudad 4
FROM ($) ($) ($) ($)
Plant 1 8 6 10 9
Floor 2 9 12 13 7
Plant 3 14 9 16 5

5.- The Industrial Steel Company S.A. (IPASA) whose main plant is located in the city of Chincha (P1)
produce shock absorbers for trucks. IPASA also has production plants in Chimbote (P2) and in
Arequipa (P3).
TheproductsaresoldtofivewholesalersD1,D2,D3,D4,andD5atS/10perunit.
Sales projections indicate that next month the following quantities should be delivered:
a D1 13000
D2 3500
D3 4500
D4 10000
D5 8000
The maximum production capacities are:
P1 15000
P2 12000
P3 10000
The unit production cost at each plant is:
P1 $3
P2 $5
P3 $2
The unit transportation cost from each plant to each distributor is:
D1 D2 D3 D4 D5
P1 2 1 3 2 1
P2 3 2 4 1 1
P3 2 2 3 1 3
Formulate a Mathematical Model to determine the optimal production of each plant and the number of
shock absorbers that each plant must send to each distributor.

6.- The company XXX produces 2 items: car covers and awnings. For next week, it has the
followingsupplies:
400 man-hours
1200 meters2of linen
2700 meters2of canvas
600 machine hours
The selling price of car covers is 150 soles and of awnings is 500 soles.
If producing a car cover requires 1 man-hour, 6 meters2of linen and 1 machine hour, and to produce
a shelter requires 2 man-hours, 18 meters2of canvas and 1.5 hours-machine.
a) Define the decision variables of the problem.
b) Formulate the corresponding linear programming mathematical model.
c) Solve by the graphical method and determine the production plan that maximizes the revenue of the
Company. What is the maximum profit achieved?

Ing. Efraín Murillo Msc.

Operations Research I 19

7.-A farmer can produce tomatoes and lettuce. To produce one ton of tomatoes, he requires 0.15 hectares, 3
Kg of fertilizer and 30 days, and for one ton of lettuce requires 0.20 hectares, 2 kg of fertilizer and 32
sessions. Determine the optimal production schedule and the final situation of the inputs, if:
- 2 hectares of agricultural land are available, and
- 60 Kgs. of fertilizer.
- A budget of 25000 soles to pay wages.
In addition, the daily wage is 15 soles, a ton of tomatoes is sold for 900 soles, and a ton of lettuce for 1800 soles.

8.- A refinery produces 2 types of gasoline (1 and 2) from 2 types of crude oil (A and B); the prices of
sale, the raw material for producing gasoline, as well as the costs are:

Maximum Minimum Price

Quantity Cost % of A for sale
Available oil unit gasoline unit required
A 100TM 6 1 60 8
B 200TM 3 2 30 5
Determine a Linear Programming Program that maximizes profits. (suggestion: define the variable
as Xij: amount of crude oil type i purchased to produce gasoline type j
9.-Lupita is worried about her overweight and the cost of daily food; she knows that to lose weight, she must
consume daily at most 1350 calories, but requires a minimum of 500 mg of vitamin A, 350 mg of
Calcium, 200 mg of proteins and 150 mg of minerals. Assuming that it only has food available
shown in table B and that each food supplies the nutritional amounts shown in the same table,
Formulate the Linear Programming Model that determines the daily diet with the lowest cost that Lupita must follow.
Table B
Food Portion Vitamin A Football Mineral Proteins Cost Calories
Milk 1 cup 105 mg. 75mg. 50mg. 35mg. S/. 1.0 80
Egg 2 units 75 80 50 15 $0.7 50
Spinach 1 serving 100 125 78 $0.5
Pig 2 chops 25 10 55 $4.0 175
Fish 2 pieces 150 50 100 50 S/. 2.0 150
Pastel 2 slices 30 5 8 S/. 1.5 200

Eng. Efraín Murillo Msc.

Operations Research I 20

2 Linear Programming: Approach

Graph
Taken from MATHUR, Solow. Operations Research, the Art of Decision Making. Mexico:
Prentice Hall Hispanoamericana, 1996.

Since real linear programming problems can be very large and complex, they are employed
computation methods to solve them. The algorithms of these methods are generally based on the technique
called the Simplex method, and are analyzed later. Managers are more interested in solutions than in
the mathematical aspects of the methods. The solutions to linear programming problems are
important because it allows seeing many details of the decision problem. Therefore, this section presents
some ideas on the algebraic and graphical solutions of linear programming and the useful concepts that them
managers can obtain from these solutions.
In most cases, it is not possible to graphically solve linear programming problems, due to our
inability to visualize more than three spatial dimensions. However, it is advisable to see how it can be
to solve a simple problem graphically, since larger dimension problems have solutions
similar.

Situation. A manufacturer is trying to decide on the production quantities for two items:
tables and chairs. There are 96 units of material and 72 hours of labor available. Each table requires 12
material units and 6 hours of labor. On the other hand, the chairs use 8 units of material each and
it requires 12 hours of labor per chair. The contribution margin is the same for tables as for
Chairs: $5 per unit. The manufacturer promised to build at least two tables.
Step 1: Problem Formulation:
What is the objective? It is to maximize the contribution to profit. Each unit of tables or chairs produced
will contribute $5 to the profit. Then the objective function can be written as:
Maximize: Z = 5X1+ 5X2
where X1produced tables
X2 = sillas producidas
What are the restrictions or limitations of the problem? There are three restrictions. First, the material is
limited to 96 units. Each table uses 12 units of material and each chair uses 8 units. The first
restriction is, then,
12X1+ 8X2≤96
The second constraint is the total labor hours. A table takes 6 hours, a chair 12 hours and
It has a total of 72 hours. Thus
6X1+12X2≤72
There is one more limitation. The manufacturer promised to produce at least two tables. This can be expressed
like: X1≥2. Finally, the non-negativity constraints are: X1≥O, X2≥O
Putting together the entire model, we have:
Maximize: Z = 5X1+ 5X2
Restrictions: 12X1+ 8X2≤96 ,
6X1+12X2≤72
X1>=2
X1≥0, X2greater than or equal to zero

Ing. Efraín Murillo Msc.

Operations Research I 21

Step 2: Graph of the constraints

The next step in the graphical method is to draw all the constraints on a graph. This can be done in
any order. For convenience, we will start with the non-negativity constraints. These are shown in the
Figure 2-1. In this graph, a solution would be represented by a point with X coordinates.1(tables) and X2(chairs).
The coordinates would represent the quantities of each item that need to be produced.
The upper right quadrant is called the feasible region since it is the only quadrant where the constraints can be.
solutions. The other three quadrants are not feasible, as they would require the production of negative quantities
of tables or of chairs or both.

Figure 2-1
Non-negativity restrictions
The following restriction is X1≥2. The simplest way to draw resource constraints is in two
pasos: (1) convertir una desigualdad en una ecuación y graficar la ecuación y (2) sombrear el área apropiada
above and below the line resulting from step 1. Turning an equality into an equation here means ignoring
the part of 'greater than' or 'less than' of the restriction.
Thus, in the example, X1 >=2, becomes X1 = 2. This equation is graphed in figure 2-2. Any point
on line X1= 2 satisfies the equation. However, the restriction is broader, since any point X12
it will also fulfill it. This includes all the points that are to the right of the line X1= 2. Then, the region
feasible includes all values of X1that are on the right wing of line X1= 2.

Figure 2-2
Restriction X1≥2

Ing. Efraín Murillo Msc.

Operations Research I 22

The limitation on labor hours is the following constraint. As before, it is first converted into
an equation: 6X1+12X2=72. This line can be graphed if two points on it are found. The pair of
The easiest points to locate are the intersections with the X axes.1y X2To find the intersection with
the X axis2it is done X1= 0. The equation then reduces to:

12X2= 72
X2= 6
As a result, the point (0,6) is obtained.
The intersection with the X axis1is doing X2= 0. Thus,

6X1= 72
X1= 12
As a result, the point (12,0) is obtained.
These two points and the line that connects them are shown in figure 2-3. Any point that is above or below
this line will comply with the restriction.

Figure 2-3
Labor restriction
To determine the feasible region (shaded area), the origin test (0,0) is performed, which is substituted in
the originating inequality:

6X1+12X2<= 72
6(0) + 12(0) = 0
0≤72, the answer is yes, which means that the origin is included in the shadow.
If the answer had been no, the shadow would have to be contrary to the
origen.
Any point above this line will require more than 72 hours of labor and is not acceptable. In the figure
This restriction is combined with the previous one. In the feasible region, both restrictions are satisfied.
The last restriction is that of material. Following the previous procedure, the first step is to find the
intersections for equality. These are X1= 0, X2= 12 and X1= 8, X2= 0. The two points are located in the
graph; the line is drawn, and since the constraint is of the less than or equal to type, the area below it is shaded.
the line. The result is shown in figure 2-5. Any solution that is on the border or within the area
darkened will comply with all restrictions. The objective function will now be used to select the
optimal solution.

Ing. Efraín Murillo Msc.

Operations Research I 23

Figure 2-4
The first two restrictions

Figure 2-5
All combined restrictions
Step 3: obtaining the optimal solution: trial and error
The trial and error method for finding the optimal solution takes advantage of a special property of the
PL problems. Consider figure 2-6 which shows the constraints of the problem. The extreme boundary of the
The feasible region is formed by the lines connecting the four points A, B, C, and D. The extreme boundary is of
special interest because:
At least one of the intersection points on the extreme boundary will be an optimal solution.

Eng. Efraín Murillo Msc.

Operations Research I 24

Figure 2-6
Extreme boundary of the feasible region
With this special property, the list of possible optimal solutions can be reduced to only four points: A,
B, C and D. The point among them that gives the highest value of the objective function will be the optimal solution.
So, the task is to find the coordinates of these four points, substitute the values into the objective function.
and select the point that gives the highest value of Z.
Regarding point A, it is known that its coordinates are (8,0). Point B is the intersection of two constraints.
If the two corresponding equations are solved simultaneously, the result will be the coordinates of their
intersection. This operation is then carried out.
Material constraint equation 12X1+ 8X2= 96
Labor constraint equation (2) 6X1+ 12X2= 72
Multiply equation (2) by 2 (3) 12X, + 24X, = 144
Subtract equation (1) from (3) (4) 16X2= 48
X2= 3
Substitute in the equation (1) 12X1+ 8(3) = 96
12X1= 72
X16
Point C is also the intersection of two constraint lines. Solving the equations again
simultaneous
Minimum production constraint equation (1) X1= 2
Labor constraint equation (2) 6X1+ 12X2= 72
Replace (1) in (2) 6(2) + 12X2= 72
12X2= 60
X2= 5
Point D has coordinates (2,0).
With the coordinates of each of the four points, one can now proceed to substitute each possible solution.
in the objective function. The calculation of Z for each of these points gives:
Point A, (8,0): Z = 5(8) + 5(0) = $40
Point B, (6,3): Z = 5(6) + 5(3) = $45
Point C, (2,5): Z = 5(2) + 5(5) = $35
Point D, (2,0): Z = 5(2) + 5(0) = $10
The highest value is $45, thus point B is optimal: to produce six tables and three chairs.
Obtaining the optimal solution: Indifference lines
There is another method to find the optimal solution by graphing the objective function on the same graph as the
restrictions. This method will be considered.

Eng. Efraín Murillo Msc.

Operations Research I 25

The objective function in this problem is Z = 5X1+ 5X2Since the maximum feasible value of is still unknown
Z cannot be optimized for the objective function. However, it is possible to assume some values for Z and
graph the resulting lines. Figure 2-7 shows the lines for Z = 25 and Z = 50. Lines of this type
They are called indifference lines because any point on a given line yields the same total gain. Note that
that the perpendicular distance from the origin to the line increases as the value of Z increases. Also, all the lines of
Indifference curves are parallel to each other. These graphical properties can be used to solve the problem.
Figure 2-8 illustrates all the constraints and the two assumed indifference curves. In the graph, it can be
it can be observed that the indifference line for Z = 50 is completely outside the feasible region. For Z = 25,
part of the line falls within the feasible region.

Figure 2-7
Lines of Indifference

Figure 2-8
Lines of constraints and indifference.
Therefore, there is some combination of X1y X2that satisfies all the constraints and yields a total profit of
$25. Upon inspection, it can be observed that there are higher profits that are feasible. Imagining that the line
from indifference Z = 25 moves towards the line Z = 50, of the properties of the graph that were noted
First, the optimal point will be on the indifference curve furthest from the origin but still touching the region.
feasible. This is shown in figure 2-9.

Ing. Efraín Murillo Msc.

Operations Research I 26

With the optimal point graphically located, the only task left is to find the coordinates of the point.
Note that the optimal point is at the intersection of the constraint lines for materials and labor hours.
work. It was previously found that the coordinates of this point are (6,3). The substitution of this point into the function
maximum profit objective:
Z 5(6) 5(3) $45

Figure 2-9
Location of the optimal point.
Summary of the graphic method
To graphically solve linear programming problems:
Express the data of the problem as an objective function and constraints.
2. Graph each constraint
3. Locate the optimal solution
a) By trial and error, or
b) With indifference lines

PRACTICE EXERCISE 2-2

Graphically solve the following problem:
Maximize: Z 10X4Y
Restrictions:5X 10Y 50
10X6Y 60
X 0,Y 0

USE OF THE GRAPHIC METHOD FOR MINIMIZATION

This section considers a linear programming problem in which the objective is to minimize costs. The solution of
The minimization problem follows the same procedure as that of maximization problems. The only

Ing. Efraín Murillo Msc.

Operations Research I 27

the difference is that now the lowest possible value is desired for the objective function. Suppose that one has the
next problem.
Diet problem
Un comprador está tratando de seleccionar la combinación más barata de dos alimentos, que debe cumplir con
certain daily vitamin needs. The vitamin requirements are at least 40 units of
vitamin W, 50 units of vitamin X and 49 units of vitamin Y. Each ounce of food A provides 4
units of vitamin W, 10 units of vitamin X, and 7 units of vitamin Y; each ounce of food B
provide 10 units of W, 5 units of X and 7 units of Y. Food A costs 5 cents/ounce and the
Food B costs 8 cents/ounce.
Step 1: problem formulation
La meta de este problema es encontrar la manera menos costosa para satisfacer las necesidades vitamínicas. Las
Two available alternatives are food A and B. Mathematically, the objective function is
Minimize: From 5A 8B
The restrictions are the minimum requirements of the three vitamins.
These are shown next:
Restrictions: 4A 10B 40 vitamin W
10A5B 50 vitamin X
7A 7B 49 vitamin Y
A 0,B no negativity
Step 2: graph of the constraints
The procedure for graphing is the same as was used before: (1) graph each constraint equation; (2)
graph the appropriate area. For the first restriction, the equation is 4A 10B 40 . Las dos intersecciones con
the axes are (0,4) y (10,0) This line is shown in figure 2-10. The constraint requires 40 units or more.
vitamin W. Any point that is above the constraint line will be feasible and all points that are
Those below this line will be unacceptable. Figure 2-11 shows the feasible region.
B
12

10

6
4A +10B = 40
(0, 4)
4

2
(10, 0)
0 2 4 6 8 10 12 A

FIGURE 2-10
Line for the equation 4A10B 40
Then the constraint for vitamin X is graphed. Equation 10A 5B 50 has intersections with the axes
in (0,10) y (5,0) Figure 2-12 illustrates the restrictions for vitamins W and X. Note that the
Solutions that remain in the areas are not feasible, as they would be below one of the lines.
restriction.
By adding the third constraint, this second step is completed, as shown in figure 2-13.

Ing. Efraín Murillo Msc.

Operations Research I 28

Figure 2-11
Restriction of vitamin W

Figure 2-12
Restrictions for vitamins W and X

Figure 2-13
All restrictions

Ing. Efraín Murillo Msc.

Operations Research I 29

Step 3: locating the optimal solution

Figure 2-14 shows the extreme frontier plus two indifference lines, those of Z= 40 cents and Z= 60.
cents. The extreme boundary is formed by the points a, b, c, d, since these are the points of
feasible intersections closest to the origin.
Graphically, the goal of minimizing the value Z means adjusting an indifference line close to the origin.
as possible. In figure 2-14, it can be observed that there are many possible solutions for Z = 60, but
none for Z = 40. Imagined moving the line Z = 60 towards the origin, the last point of contact with the
extreme border will be the point. Then, the point is the intersection of two lines:
(1) 4A 10B 40
(2) 7A 7B 49
Solving this system of equations:
Multiply equation (1) by 7 (3) 28A 70B 280
Multiply equation (2) by 4 (4) 28A 28B 196
Subtract (4) from (3) 42B 84
B=2
Substitute in equation (1) 4A 10(2) 40
A 5
The least expensive solution is 5 ounces of food A and 2 ounces of food B. The cost of this combination is
Z - 5A + 8B - 25 + 16 - 41 cents

If the trial and error method is used to locate the

optimal solution, the coordinates must be found
from points a, b, and d. It should be calculated later.
the value of the objective function for each point.
The results of this are shown below
procedure:
Results of trial and error

Punto Coordinates Z=5A+8B

a A= 10,B=0 50
b A = 5, B = 2 41less
c A = 9,B = 4 47
d A = 0,B = 10 80
Figure 2-14
Indifference curve for Z = 40 and Z = 60

PRACTICEEXERCISE2–3
Graphically solve the following problem:
Minimize: Z = 5X + 1.5Y
Restrictions: 2X + 5Y ≥ 10
5X + 2Y ≥ 1.5Y
X ≥ 0, Y ≥ 0

Ing. Efraín Murillo Msc.

Operations Research I 30

2.3 SPECIAL CASES

LINEARPROGRAMSWITHSPECIALGEOMETRICPROPERTIES
In the previous section, we saw how to graphically solve a linear programming problem with two variables. In
In the examples, the optimal solution occurs at an extreme point of the feasible region. In this section, you will see examples of
linear programs for which this is not the case. It will also know what causes these exceptions and how
interpret them.

2.3.1 Infeasible linear programs

Example 2.3.1: IJK produces two solvents, CS-01 and CS-02. The plant operates 40 hours a week and employs 5
full-time workers and 2 part-time workers, who work 15 hours a week to operate the 7 machines
they mix certain chemicals to produce each solvent. This workforce provides 230 hours of work.
available in the mixing department. The products, once mixed, are refined in the department of
purification, which currently has 7 purifiers and employs 6 full-time workers and 1 part-time worker
part-time, that works 10 hours a week. This job provides up to 250 hours of available work in the
Purification department. The hours required in the mixing and purification departments to produce a thousand
the gallons of each of the solvents are listed in the following table:
CS-01 CS-02
Mixed 2 1
Purification 1 2
IJK has an almost limitless supply of the raw material it needs to produce the two solvents. It can sell
any amount of CS-01, but the demand for the more specialized product, CS-02, is limited to a maximum of
120,000 gallons per week. The accounting department estimates a profit margin of $0.30 per gallon.
CS-01 and $0.50 per gallon of CS-02. Since all employees are salaried and therefore, they are paid the same
amount regardless of how many hours they work, these salaries and the costs of the machines are considered fixed and not
they are included in the calculation of the profit margin. As the manager of the production planning department of IJK,
You are asked to determine the optimal weekly manufacturing plan for IJK.
The linear program is as follows:
Maximize 3x 1+ 5x2
Depending on: 2 x1+ x2≤ 230 (1)
x1+ 2x2≤250 (2)
x 2≤120 3
x1 ≥ 0 (4)
x 2≥ 0 (5)
The feasible region corresponding to those five constraints is described in figure 2.15.

Figure 2.15 Feasible region for the previous example

Ing. Efraín Murillo Msc.

Operations Research I 31

Example 2.3.2: To examine a type of exception, your sales manager informs you that he wants to sign a contract to
long term to supply 150,000 gallons of CS-01 each week. You are asked again to determine the plan for
optimal weekly manufacturing for IJK
The linear program is as follows: AN INFEASIBLE LINEAR PROGRAM
Maximize 3x 1+ 5x2
Depending on: 2 x1+ x2≤ 230 (1)
x1+ 2x2≤250 (2)
x 2≤120 (3)
x1 ≥ 0 (4)
x 2≥ 0 (5)
x1 >=150 (6)
The result of adding the new constraint (6) is illustrated in figure 2.16. As you can see, there are no values of x.1
y x2that satisfy the new constraint (6) and all previous constraints. This means that, with their
current resources, IJKnopuede cannot satisfy a contractual agreement to supply 150,000 gallons of CS-01 for
week. If the proposed contract is signed, IJK will need to obtain additional resources to increase the
production capabilities. The executive management will therefore have to make a strategic decision about the
value of this investment.
The linear program from example 2.3.2 was called an infeasible linear program, which means that no
the value of the variables satisfies all the constraints simultaneously, that is, there is no region
feasible. For such problems, it is recommended to reformulate them.
Perhaps the most common reason for a linear program to be infeasible is an error in one of the
[Link],itmayhavebeenaccidentallywrittensimultaneously
a restriction less than or equal to as a contraction greater than or equal. Alternatively, it may have been
an error occurred while capturing the data for a linear program in a computer program. For example,
the restriction x1+ x2≤230 could have been accidentally introduced into the computer as xl+2x≤130,
2 y
this error could result in an infeasible linear program. If you discover that a linear program is infeasible,
First, carefully check the formulation of the problem, and then ensure that the data has been
correctly introduced.

Figure 2.16 Graphic solution for example 2.3.2

Even if there are no errors in the formulation of the problem or in the way of capturing the data on the computer,

Eng. Efraín Murillo Msc.

Operations Research I 32

a linear program can still be infeasible. These infeasible problems occur in practice when the
restrictions are too restrictive. Such restrictions can arise in numerous ways. For example,
the requirement to produce at least 150 thousand gallons of CS-01 in restriction (6) of example 2.3.2 imposes a
limitation that cannot be handled by the resources available in the mixing departments
purification of IJK. As another example, the government can impose such severe restrictions on the amount of
pollutants emitted by a steel plant that make it impossible for any production plan to be feasible
satisfy those requirements and achieve a minimum level of profits.

2.3.2 Linear programs with unlimited solutions

Consider another example of a linear program whose solution does not occur at an extreme point of the region.
feasible. Let's assume that the inequality signs of the constraints (1) and (2) from example 2.3.2 are reversed.
accidentally during the formulation of the problem. The resulting linear program is given in example 2.3.3.
EXAMPLE 2.3.3 AN UNBOUNDED LINEAR PROGRAM
Maximize 3x 1+ 5x2
Depending on: 2 x1+ x2≥230 (1)
x1+ 2x2≥250 (2)
x 2≤120 (3)
x1 ≥ 0 (4)
x 2≥ 0 (5)
The feasible region of this linear program appears in Figure 2.17, along with the line of the objective function at
through point x1=100yx2=100.Accordingtotheproceduredescribedinsection2.2,thislinearprogramis
Resolve graphically by moving the line of the objective function parallel to itself in the direction of improvement.
indicated by the arrow on that line in figure 2.17, until the line is about to leave the feasible region. In
In this example, the line of the objective function will never move away because the rightmost end of the feasible region does not have
limit. Theoretically, this would mean that IJK can achieve infinite profit, which is clearly
impossible.
It is said that such a problem is an unlimited linear program, which means that the objective function can
improve indefinitely, that is, there are feasible values of the variables that can make the value of the
objective function as large as desired in the case of maximization (or as small as desired in the case of
minimization). An unlimited linear program is graphically identified when the line of the objective function can
to move parallel to itself in the direction of improvement without ever leaving the feasible region.

Figure 2.17 Graphical solution for example 2.3.3

The general cause of a linear program is an error in the formulation of the problem. For example, it may have been
an accidental restriction has been omitted. Likewise, an error may have been made when capturing the data in a
computer program. If you discover that a linear program is unbounded, first check carefully the
problem formulation, and then make sure that the data has been captured correctly. It is important

Ing. Efraín Murillo Msc.

Operations Research I 33

Observe that there is a difference between a limited feasible region and a limited linear program. How is this
As illustrated in figure 2.18, it is possible for the feasible region to be unbounded, but for the linear program not to be.
unlimited.

Minimize 3x1 2x2

Depending on 2x 1 x2 230(1)
x1 2x2 250 (2)
x2 120 (3)
x1,x2 0

Figure 2.18 An unlimited feasible region but not an unlimited linear program

KEY FEATURES

So far, you have seen three mutually exclusive types of linear programming problems.
to know, those who are:
[Link]: That is, they have an optimal solution.
2. Infeasible: That is, there are no values of the variables that satisfy all the constraints.
simultaneously.
3. Unlimited: That is, there are feasible values of the variables that make the objective function so
as big or as small as desired.
Every linear program is optimal, infeasible, or unbounded.

2.3.3 Linear programs with redundant constraints

When formulating a linear programming problem, one may encounter a constraint that is not necessary, in
The sense that the feasible region is exactly the same whether it includes this constraint or not.
Such a restriction is a redundant restriction. As an illustration, suppose that the following restriction is
add to example 2.3.1, the IJK problem:

The feasible region of the resulting linear program is shown in figure 2.19. Note that the constraint (6) does not
affects the feasible region. The feasible region in figure 2.19 with the constraint (6) is exactly the same as the
from figure 2.15, which does not include this restriction. The restriction (6) is therefore redundant.
You may not be aware that a constraint is redundant when formulating a linear program, but this does not...
necessarily concerns him. The optimal solution does not change if a redundant constraint is included. The only
the effect is that it increases the size of the linear program, which, in turn, can lead to a program of

Ing. Efraín Murillo Msc.

Operations Research I 34

computer takes extra time to solve the problem. The general rule is: do not worry about the
redundant restrictions; if you think you need a restriction, include it in the formulation.

Figure 2.19 A linear program with a redundant constraint

2.3.4 Linear programs with alternative or multiple optimal solutions

Some linear programs have more than one optimal solution. Each optimal solution is called a solution.
optimal alternative. Having optimal alternative solutions means only that there are different values
feasible for the variables that produce the same best value of the objective function. All solutions
optimal ones are the same in that, by definition, they are the best. However, you may prefer
one of these optimal alternative solutions over the others for some secondary reason, perhaps because one
solution is easier to implement than the others.
To illustrate the above, let us assume that the profit margins of example 2.3.1 are modified to $0.20 per
gallon of CS-01 and $0.40 per gallon of CS-02. The new linear program is given in example 2.3.4.

EXAMPLE 2.3.4 A LINEAR PROGRAM WITH AN ALTERNATIVE OPTIMAL SOLUTION

Maximize 2x 1+ 4x2
Depending on: 2 x1+ x2≤ 230 (1)
x1+ 2x2<=250 (2)
x 2<=120 (3)
x1 greater than or equal to zero (4)
x 2greater than or equal to 0 (5)

Ing. Efraín Murillo Msc.

Operations Research I 35

The graphical solution to this problem is illustrated in figure 2.20. There you can see that the two endpoints C and D
are optimal, in fact all the points on the line segment that connects these two endpoints are
optimal. Geometrically, the reason why this problem has alternative optimal solutions is due to
that the line of the objective function is parallel to one of the constraints that passes through one of the solutions
optima. In example 2.3.4, the objective function line is parallel to constraint (2).

Figure 2.20 A linear program with alternative optimal solutions

Ing. Efraín Murillo Msc.

Operations Research I 36

Ing. Efraín Murillo Msc.

Operations Research I 37

Practical Solution of Models with the Graphical Method

CASE 1. MODELS WITH SINGLE OPTIMAL SOLUTION.

The model is formulated by an investment consulting firm to develop a client's portfolio. The
variables X1 and X2 represent the quantity of Type 1 and Type 2 shares to buy in order to meet the objective
established to maximize the annual return on that investment or stock purchase. The total amount available
to invest is $80,000. The risk is a relative measure of the two alternative investments.
Type 1 action is a riskier investment. By limiting the total risk for the portfolio, the firm
the investor avoids placing excessive amounts of the portfolio in potentially high-return investments but
high risk. It also limits the amount of higher risk actions.

Max 3X1+ 5X2 Annual return in $

Subject to:
25 X1 + 50 X2 <= $80,000 available funds
0.5X1 + 0.25 X2 <= 700 maximum risk
1X1 <= 1.000
Type 1 actions
X1, X2 >= 0

a) Graphing the constraints:

Restriction 1: When X1 = 0, then X2 = 1.600; When X2 = 0, then X1 = 3.200

Join the points (3.200, 0) and (0, 1.600). The side of the restriction "<" is below that line.
Constraint 2: When X1 = 0, then X2 = 2.800; When X2 = 0, then X1 = 1.400
Connect the points (1.400, 0) and (0, 2.800). The side of the constraint "<" is below that line.
Constraint 3: X1 = 1.000 y X2 = 0 It is a line that starts from the x-coordinate at point 1.000.
The side of the constraint "<" is to be taken from that line, towards the side where the origin point is.

Somber, or indicate in some way, the convex set also called the feasible region.
(See Graph 1).

b) Graph the Objective Function

assigning it an arbitrary value. This value,
preferably, it should allow the objective
it shows in the solution region. For example,
the value 3,000 can be used.
cut-off points on the axes, to graph it,
the points are (1,000, 0) and (0, 600). The
Objective Function is graphed with a line of
color, in this case, to differentiate it from the
restrictions.

c) Move the Objective Function parallelly.

herself in the direction that increases her
value (upward in this case), until
I touched the last one (the last ones, if I touch them at)
same time) extreme point of the region
solution.

d) At that extreme final point, b in this case,

solve the system of equations what is
intercept. In this case, they are the equations 1
y 2. Use any of the methods to
solve pairs of linear equations with two variables.

Ing. Efraín Murillo Msc.

Operations Research I 38

e) Alternatively, to determine the optimal solution, you can calculate the coordinates of all the points.
extremes: a, b, c, d, and e, in the convex set of solutions. Then evaluate the Objective Function at each one
of them. The extreme point that provides the highest value will be the optimal extreme point.

f) In both cases, the optimal solution is obtained at the endpoint b with coordinates (800, 1,200). Thus,
the optimal solution is X1 = 800 and X2 = 1,200. Solving in the Objective Function:

Max 3X1 + 5X2, is obtained: 3(800) + 5(1,200) = 8,400

In this case 1, answer the following:

1.1 What does the coefficient of variable X2 represent in the Objective Function and in the second constraint?
1.2 What type of solution does the model present?, Why? and How is it recognized in the graph?
1.3 What decision would be recommended with the solution found?
1.4 Analyze the constraints at the optimal point and present the information obtained.
1.5 What effect would a change in the annual return of each stock have on the optimal solution found?
Type 2. Suppose you change to 9. Explain and show on the graph. What is the name of this analysis that you
What does it do?
ANSWERS:

1.1 In the Objective Function, it represents the annual return of each Type 2 stock purchased, that is, each stock
Type 2 that is purchased will provide an annual return of Bs. 5. In constraint 2, it represents the measured risk.
For each Type 2 action. That is, each Type 2 action has a risk of 0.25.

1.2 Unique Solution, because there is a unique combination of Type 1 and Type 2 actions to purchase that maximizes the
annual return on investment and is recognized in the graph because a single extreme point provides the
maximum value for the objective. In this case, point b.

1.3 Buy 800 shares Type 1 and 1,200 shares Type 2 to maximize the annual income to 8,400.
monetary units ($)

1.4 Constraint 1: 25 (800) + 50 (1200) = 80,000 It is observed that it is fulfilled exactly, that is as
an equality. This indicates that with that optimal decision, the maximum budget amount is fully utilized.
available for purchase.

Restriction 2: 0.5 (800) + 0.25 (1200) = 700 It is observed that it holds exactly, that is, as a
equality. This indicates that with that optimal decision, the maximum required amount of risk will be fully achieved.
for the purchase.

Restriction 3: 1 (800) = 800; 800 < 1,000 It is observed that it complies as a

inequality. This indicates that with that optimal decision, 800 Type 1 shares are purchased, 200 less than
maximum amount required. Remember that this is allowed because the restriction is "less than or equal to".

In the graph, it can be observed, as something logical, that the constraints that are satisfied as equalities are
crossing over the optimum point and those that are satisfied as inequalities are in the solution region
away from the optimal point.

1.5 The analysis is called Sensitivity Analysis. It is performed after obtaining the solution.
optimal. In this case, it is carried out to determine the effect it has on the optimal solution,
change in a coefficient of a variable in the Objective Function.

The new Objective Function, 3X1 + 9X2, is graphed again.

Eng. Efraín Murillo Msc.

Operations Research I 39

It is observed on the same graph 2 that

the slope of the target line (f.O2)
change. Now when moving, in
growth, the extreme point that
the limit is that of the intersection of the
lines that correspond a the
restriction of available funds and to the
ordered or non-negativity restriction,
extreme point "a". The new solution is
a unique solution, with other values for
the variables.

CASE 2. MODELS WITH ALTERNATE OR MULTIPLE OPTIMAL SOLUTIONS.

Max 6X1 + 2X2 Benefit

Subject to:
3X1 + X2 <= 4 8 working hours
3X1 + 4X2 <= 120 units of raw material
3X1 + X2 >= 36 supervision hours.
X1, X2 >= 0

The model is formulated by a company that wants to determine the quantity of units of product 1 (X1) and
Product 2 (X2) to be manufactured to meet the established goal of maximizing profit. The total amount
disponible de horas de trabajo para este período es de 48. La disponibilidad de materia prima es de 120 unidades
and the minimum amount of hours available for supervision is 36 hours.

Graphing the restrictions and obtain the

the solution space is carried out in a similar way
to the process carried out in case 1 and therefore
the instructions will not be repeated.

(See Graph 3)

The extreme points of the convex set

son: A(16,0) B(8,24) C(8/3,28) y
D(12,0).

Two extreme points provide the

maximum value of the objective, points A and B.
This allows us to state that there are solutions
optimal Alternatives for this model. They are
optimal all points on the segment
of line AB that limits the set
convex solution and correspond to the
first restriction.

Ing. Efraín Murillo Msc.

Operations Research I 40

If you use the method of graphing the Objective Function with an arbitrary value, 48 for example, you will be able to
observe that the line is completely parallel to the first and third restriction. When moving it
parallelly towards its optimization, upwards because it is being maximized, it will finally fall
completely on the first restriction, of working hours, before completely leaving the region
solution. Two extreme points would be limiting the growth of the objective, point B and point A.

Any line that has a ratio of coefficients equal to that of another line is parallel to that other line.

The advantage of models with this type of solution is that any of them can be chosen.
optimal solutions, because all present the same optimal value for the objective. For example, if one of the
solutions have fractional values for the variables and cannot be worked with fractional values, the one that
make the decision to select a solution with integer values.

Answers for some questions similar to the previous model (Case 1: Model with
Unique Optimal Solution:
1.2 It is an optimal alternative solution because there is more than one combination of products 1 and 2 to produce, that
they provide the same optimal value for profit.
It is recognized in the graph because more than one extreme point limits the optimal value of the objective or provides
its optimal value, the points: A (16,0) and point B (8,24). Therefore, all points on the line AB.
they are also optimal.
1.3 You must select one of all the solutions. Suppose you choose the endpoint B (8, 24)
In this case, the decision would be: Produce 8 units of product 1 and 24 units of product 2 to maximize.
the profit in 96 monetary units: 6 (8) + 2 (24) = 96
This value for the Objective Function would also be obtained at the other extreme point A (16, 0) and in
any point on line AB in the solution region.
1.4. Constraint 1: 3 (8) + 1 (24) <= 48 48=48

It is observed that it holds as an equality. This indicates that with that optimal decision, the resource is fully utilized.
maximum amount of working hours available for production.

Constraint 2: 3 (8) + 4 (24) = 120

It is observed that it is fulfilled exactly, that is to say as an equality. This indicates that with that optimal decision
The maximum available amount of Raw Material will be fully utilized.

Constraint 3: 3 (8) + 1(24) >= 36 48>36

It is observed that it holds as an inequality. This indicates that with that optimal decision, 12 hours are used.
about the minimum amount of available supervision hours. 48 hours are used.

1.5 Here you can practice Sensitivity Analysis on the chart, varying some element of the model.

Ing. Efraín Murillo Msc.

Operations Research I 41

CASE 3. MODELS WITHOUT POSSIBLE SOLUTION.

The elements of the model will not be defined because not

There will be a possible solution to make a decision.

(See chart 4)

Max 40 X1 + 30 X2

Subject to:
2/5X1 + 1/2X2 <= 20
1/5X2 <= 5
3/5X1 + 3/10X2 <= 21
X1 >= 30
X2 >= 15
X1, X2 >= 0

It can be observed in Graph 4 that while the first 3 constraints delimit a space in
common, the last 2 delimit another common space for them. Therefore, there is no region of points
common that satisfy both sets of constraints and the model will have no possible solution. In
in these cases it is necessary to determine which are the inconsistent constraints for the model. That is,
which are really valid for the model.

Note that if the variables X1 and X2 they take the minimum value they can take in both
last restrictions, that is X1 = 30 and X2 = 15 then the third restriction would not be fulfilled.
This is an inconsistency.

These models should not exist in the real world (14). If the modeled system works, then the
The model must represent it in such a way that allows for a possible solution.

CASE 4. MODELS THAT PRESENT SOLUTION WITH INFINITE VALUE.

Max X1 + 2X2
Subject to:
-4 X1 + 3 X2 <= 3
X1 - X2 <= 3
X1, X2 >= 0

The elements of the model will not be defined

because there will be no solution to take
any decision.

In graph 5 the convex set called

region solution, which contains all the
possible solutions, it is an open space.

It has three extreme points A, B, and C, but none delineates the growth of the objective. This function can

Ing. Efraín Murillo Msc.

Operations Research I 42

take infinite values since the variables form points with infinite values within the region
solution and none of them provide you with an optimal finite value. Therefore, given the constraints, not
It is logical to find an objective of infinite value.

In these cases, the restriction or restrictions that were omitted in the system must be sought.
model and that would limit the decision variables to feasible values.

CASE 5. MODELS WITH UNBOUNDED SOLUTION SPACE AND VALUE SOLUTION

FINISHED.

Min 0.06 X1 + 0.05 X2 (costs)

Subject to:

0.30 X1 + 0.20 X2 >= 500 Protein

0.15 X1 + 0.30 X2 >= 300 Fat
X1, X2 >= 0

The model is designed for a dog daycare that stands out for providing a balanced diet to
Pets. The food is prepared by mixing 2 well-known brands of food that we will call X1 and X2.
It is desired to determine the amount of grams of X1 and X2 to mix in the food, with the aim
established to minimize mixing costs. This,
must contain at least 500 grams of protein and
less than 300 grams of fat per day. The percentages
of fat and protein content per gram of
X1 and X2 are known and used in the model.

The solution space obtained is shown in

Graph 6. An open area is observed with the
possible solutions and extreme points A, B, C.

This indicatesthat there can be combinations of

amount of grams of food X1 and X2 with value
infinite, in this case the costs would be infinite. This
it is possible because it is not directly limiting the
cantidad de X1 and X2 in some restriction
specific and the existing restrictions are all of
Type what.

But, as long as there is at least one combination

with finite value, at some extreme point that limits the
value of the objective, to that combination is
will be considered optimal. In the cases of open region
possible solutions, it is advisable then
find the optimal value using the procedure of
Graph the Objective Function.

When graphing the Objective Function, with an arbitrary value of 120, it is observed that when displaced
parallelly towards its optimization, downwards because it is being minimized, the line falls on point B,
before completely leaving the solution region. This point will be considered an optimal extreme point.

The optimal solution is Unique with the values: X1 = 1.500, X2 = 250 F.O. = 102.5

Knowing the definition of the model, you can answer questions similar to those asked in case 1.

Ing. Efraín Murillo Msc.

Operations Research I 43

CASO 6. MODELS WITH DEGENERATE SOLUTION

Min 2500 X + 2200 Y (costs)

Subject to:
X + Y <= 10 Temporary employees
300X + 400Y >= 3,400 cards
80X + 50Y >= 680 packages
X, Y >= 0
The model is formulated by a post office that can hire up to 10 employees to manage the
email. The office knows that one employee (male) can handle 300 letters and 80 packages per day and a
An employee (woman) can handle 400 letters and 50 packages in a day. No less than 3,400 letters and 680 packages.
they are expected per day. Each male employee (X) is paid Bs. 2,500 per day and a female employee (Y)
They are paid Bs. 2,200 per day.
It is desired to determine the number of men (X) and
women (Y) who must be hired to satisfy
the restrictions and achieve the established goal of
minimize payroll costs.
Graph the constraints and obtain the space of
The solution is carried out in a manner similar to that done in the.
case 1 and therefore the instructions are not repeated.
The obtained graph is Graph 7.
A region of possible solutions is observed for a
only common point for all the restrictions and by so
so a single extreme point A.
This indicates that there is a unique combination.
possible and also optimal, in terms of number of employees
X and Y that satisfy the restrictions and optimize the
objective.
Knowing the definition of the model, you can answer.
similar questions to those asked in Case 1

1.1 What does the coefficient of the variable Y represent in the Objective Function and in the second constraint?
1.2 What type of solution does the model present?, Why? and How is it recognized in the graph?
1.3 What is the solution and the decision that would be recommended with the found solution?
1.4 Analyze the constraints at the optimal point and present the information obtained.
1.5 What effect would a change in the expected number of cards have on the optimal solution found.
Suppose it changes to 2.400. Explain and show on the graph. What is this Analysis called?
Answers:
1.1 The coefficient of variable Y in the Objective Function represents what is paid daily to each
female worker, that is, the cost of hiring a female worker per day is Bs. 2,200. In the second
The restriction represents the number of letters that each hired woman can handle per day, that is, 400 letters.
per day each hired woman can handle.
1.2 Unique and Degenerate. Normally the solution of a model contains a variable (Structural or of
slack) with a value greater than zero for each constraint of the model. In this case, more variables than normal.
they take a value of zero, in order to satisfy a greater number of constraints, at the optimal point. There is then
lower number of variables with a value greater than zero in relation to the number of constraints. That is why it is
Degenerate Solution as opposed to the Normal Solution. Furthermore, it is unique because it is a single one.
Combination of employees, men and women, provides the minimum cost.
Se debe a la presencia de restricciones redundantes en el modelo y se reconoce en el gráfico porque más de
Two constraints intersect at the optimal point. Of the total constraints that intersect the optimal point, only two
are necessary to calculate their coordinates. In this case, there is only one redundant constraint, so

Ing. Efraín Murillo Msc.

Operations Research I 44

The solution is degenerate. It is recognized as unique because there is only one extreme point that provides the
optimal value of the objective.

1.3 The solution is X = 6, Y = 4, F.O. = 23,800. The decision would be to hire 6 male employees and 4
women to minimize daily hiring costs by 23,800 monetary units: 2,500(6) + 2.200
(4).
1.4 Restriction 1
X + Y <= 10
6 + 4 = 10 With this decision, the maximum number of employees that were willing to be hired is contracted.
to hire.
Restriction 2
300X + 400Y <= 3,400
300(6) + 400(4) = 3 400 With this decision, the minimum number of letters expected will be handled.

Restriction 3
80X + 50Y >= 680
80(6) +50(4) = 680 With this decision, the minimum number of packages that is expected will be handled.

1.5 The analysis to be carried out is called Analysis of

Sensitivity.
The new constraint is graphed above the graph.
300X + 400Y = 2400
It is observed that the solution space does not change.
possible and therefore the optimal solution will continue
being the same.
In general, decrease the amount on the right side of
a Type constraint to relax the restriction and
make it easier to satisfy.
This can expand the convex set or leave it.
same. In this case, it remained the same. This is studied more
detailed in Sensitivity Analysis.

EXERCISES
1.- Which of the following relationships would not be acceptable as PL constraints?
1.7X 5Y 9
2.8XY 2X 12
3. 13X 9X 2 8
4.4X 7 / Y
5. XY 1

2.- Which of the following relationships would be acceptable as PL constraints?

Ing. Efraín Murillo Msc.

Operations Research I 45

1. X2 Y2 8
2. 3X 7Y 9
3. 12X8 9XY
7X
4.  2.2X 13Y
Y
5.6XY 4Y

3. Given the following data for products A and B

Product Resource
Resource
A B available
R1 60 20 1200
R2 40 50 2000
Contribution/unit $3 $ 1.5

a) Write the objective function and the constraints.

b) Find the optimal solution using the graphical method.

4. Given the following data regarding the products x1xy2

Product Resource
Resource available
x1 x2
R1 4 2 16
R2 410 2 8
Contribution/unit $1 $2
a) Escríbanse la función objetivo y las restricciones.
b) Find the product mix that maximizes the contribution.

5. With the following data:

Product Resource
Resource
A B available
R1 10 10 200
R2 8 4 64
R3 3 0 72
Contribution/unit $5 $5
a) Write the objective function and the constraints.
b) Find the optimal solution using the graphical method.

6. Data of products A and B such that:

Hours required per unit

Center I Center II Contribution $
A 8 20 16
B 15 10 12
Available hours 112 170
Find the optimal product mix.

7. Delisabell produces nectar yogurts for human consumption. The yogurts and nectars are made from
fruit and for the yogurt milk. The yogurts can be probiotic and commercial yogurt.
We have an availability of 40 kg of fruit and 90 liters of milk daily.

Ing. Efraín Murillo Msc.

Operations Research I 46

Table of fruit and milk requirements for each type of yogurt

Probiotic Commercial Daily availability

1kg=
Fruit 20kg 10kg 40kg
1 liter
Milk 40 liters. 35 liters 90lt

Unit utility S/.4.0 S/.3.0

It is known that the daily demand from Delisabell's customers for probiotic yogurts
it restricts its demand to 15 liters. Delisabell wants to determine the optimal mix of probiotic yogurt and the
commercial for your daily utility maximization.
For the following problems, find the optimal solution using the graphical method.

Ing. Efraín Murillo Msc.

Operations Research I 47

3 Sensitivity Analysis: Approach

Graph
. Taken from HERNANDEZ, José. Operations Research 1, 2011.

The search for the solution of a decision model is just the first step of the analysis. It is also
It is important for the manager to understand how sensitive the solution is to changes in the assumptions.
and the exogenous factors. This also applies to linear programming models, and one of the
pleasant characteristics of linear programming models is that a large part of this analysis of
sensitivity comes directly from the solution to the problem. First, we will see these concepts in
graphic form and then by interpreting the outputs of the computer programs
which are used to solve linear programming problems.

Example: A company manufactures two products, A and B. Each requires time on two machines. The
The first machine has 24 hours available and the second has 16. Each unit of product A
It requires two hours on both machines and each unit of product B needs three hours on the first.
machine and one hour in the second. The benefits are six dollars per unit of A and seven
dollars per unit of B, the company can sell all the units it produces of product A and
Due to market limitations, only 6 units of product B can be sold. Assume that the objective is
maximize profit; How many units of products A and B should be produced?

Formulation. Let:

X1Number of units of product A that will be manufactured.

X2Number of units of product B that will be manufactured.
Z = Total profit of the company.

The situation and the objective (to maximize profit) can be expressed with the following equations:

Maximize: P = 6X1+ 7X2

Subject to: 2X1+ 3X224
2X1+ X216
X2 6
X1,X2 0

The graphical solution is presented in figure 1.

Ing. Efraín Murillo Msc.

Operations Research I 48

Dual prices.
Consider the restriction equation for machine 1, which specifies a maximum of 24 hours.
available. In terms of linear programming, this capacity limit is often
denote the value of the independent term or second member of the constraint (or simply TI).
Suppose you can add an hour for the restriction to be:

2X1+ 3X225
What happens with the solution?

This case is presented graphically in figure 2.

The new optimal solution moves to point D', with X1= 5.75 y X2= 4.5. Like the previous solution
required X1= 6 y X2= 4, one additional hour available on machine 1 results in a
reduction of 0.25 units of product A and an increase of 0.5 units of product B. The change
net in the objective function Z is, then:
-(0.25)$6 + (0.50)$7 = $2

which represents an increase of two dollars in the profit. This is known as dual pricing,
marginal value or shadow price is the incremental change in benefits per unit change in
the independent term of a constraint.

Ing. Efraín Murillo Msc.

Operations Research I 49

Note that the dual price is maintained for a reduction in the value of the constant term.
For example, if there were only 23 hours available on machine 1, point D'' of figure 2 would be the
optimal solution (X1= 6.25 y X2 = 3.5), with a two-dollar reduction in profit. Then, the
dual price or marginal value represents the incremental increase in profit when a
restriction is expanded in one unit and a reduction in benefit when the restriction is tightened
in one unit.

The same analysis can be applied to the constraint of machine 2, which is shown in figure 3.
When the constraint is expanded by adding an hour, it becomes:

2X1+ X217

and the optimal point is D' with X1= 6.75 y X23.5. This represents an increase of 0.75 in the number of
units of product A and a reduction of 0.5 units of product B. The net effect on the
the benefit is:

+(0.75)$6 + (0.50)$7 = $1

A similar reduction in the constant term (that is, in the number of available hours) of the
Restriction 2 results in a solution for X.1= 5.25 y X2= 4.5, and a reduction of one dollar in
the benefit. Therefore, the dual price associated with the constraint of machine 2 is one dollar.
Consider the third constraint, the limit of the demand for product B:

X2 6

The increase of one unit in that limit, to X27 and the reduction of one unit, to X25, are shown
Graphically in figure 4. Note that none of the changes affect the solution, since the
restriction X26 is not effective. The optimal solution required only four units of product B.
therefore the limit of six units imposed by the market does not matter. Thus, the dual price is
zero. In fact, the dual price of any non-binding constraint is always zero.

Ing. Efraín Murillo Msc.

Operations Research I 50

Reduced costs: Dual prices for non-negativity constraints. Also

It is possible to determine the marginal values associated with introducing at least one unit of a
decision variable in the solution. Remember that the non-negativity constraints are X10 and X2
Including a unit in the solution implies modifying a non-negativity constraint on X. 11 or
X2The marginal values of doing this are called reduced costs.
Consider again the basic problem whose graph is presented in figure 4. The optimal solution
requires X1= 6 y X24. Both are positive values and, therefore, none of the restrictions of
no negativity is effective. So, the marginal value (that is, the reduced cost) associated with your
the modification is zero, just like for the other ineffective constraints
Suponga ahora que la función objetivo es Z = 10X1+ 3X2, as shown in figure 5. Point C
it is the optimal solution, with X1= 8 y X2= 0 and a profit of 80 dollars. Note that X2= 0, therefore the
non-negativity constraint X20 is effective. Now suppose that at least one must be produced.
unit of product B, due to a commitment to a client, which changes the restriction of
no negativity to X2With this, the optimal solution of figure 5 changes to C', with X.1= 7.5, X2= 1 y
profit of 78 dollars, a reduction of two dollars compared to the previous level. In this case the
reduced cost or marginal value associated with the non-negativity constraint of X2it is two dollars,
which represents the cost of retaining the customer.

Ing. Efraín Murillo Msc.

Operations Research I 51

Use of dual prices. Dual prices have many business applications.

Although there are restrictions and limitations in the world, most of them are not absolute.
For example, the manager who formulated the linear programming problem determined the available hours.
en cada una de las máquinas en circunstancias normales, pero podría obtener horas adicionales de
overtime work, purchasing additional equipment or rescheduling other activities. The prices
Dual indicators show whether it is worth doing and with what margin, thus helping to identify bottlenecks.
key bottle. In our example, the manager knows it is worth double (two dollars instead of one)
obtain additional hours for machine 1 than for machine 2.
SUMMARY. A dual price represents the marginal value associated with a unit change in the
independent term of a constraint. A reduced cost represents the marginal value of introducing
in the solution, one unit of a decision variable. Reduced costs can be considered
as dual prices of the non-negativity constraints. If a constraint is not binding, its price
dual is zero.

SENSITIVITY ANALYSIS: RIGHT SIDES OF THE

RESTRICTIONS
Dual prices provide the marginal value of making a small change at the limit.
from a restriction (that is, the value of the constant term), but it would be a mistake to believe that these
values would be the same if the capacity were changed arbitrarily. A point is reached where
the additional capacity is excessive and has no value, therefore, there are limits regarding the interval
of capacity in which the marginal values are maintained.
Considere de nuevo la restricción de 24 horas disponibles para la máquina 1. La figura 6 muestra lo
what happens when adding hours. Remember that each additional hour resulted in a reduction of
0.25 unidades del producto A y un aumento de 0.5 unidades del producto B. El precio dual asociado
With each hour of variation it was two dollars. With 28 hours available, the optimal solution goes from D
In J, the optimal solution is X1= 5 y X2The additional hours on machine 1 will not have
effect beyond this point, as the constraint X2= 6 is now effective. With the others
restrictions of the problem, the amount of 28 hours on machine 1 is the maximum that can be used of
profitable way. So, this increase of four hours, to reach 28 available hours,
represents the upper limit of the interval in which the dual price of two dollars is valid.

Ing. Efraín Murillo Msc.

Operations Research I 52

Similarly, by reducing the available hours of machine 1, the optimal solution shifts.
up to point C of figure 6, where 16 hours of machine time are used. When reducing the hours,
the profit decreases by increasing the number of units of product A and by reducing the number of
unidades de producto B, pero al llegar al punto C no se producen unidades de B, por lo cual ya no es
this substitution process is possible. Then, the lower limit of the dual price interval of two
Dollars for machine time 1 is a reduction of eight hours (from 24 to 16)

A similar analysis can be performed for machine 2; figure 7 shows the limits. When adding
hours to the available ones, the optimal solution shifts from D to B. At this point, eight have been added.
hours (from 16 to 24) and the additional hours will no longer add value. If the hours are reduced by four
available (from 16 to 12), the optimal solution moves from D to H. Then, the dual price of one dollar is
keeps in the interval of 12 to 24 hours available in machine 2.

The market demand constraint for product B, X26, it's a little different. Remember
that this restriction is not effective and has a dual price of zero. The optimal solution requires only four
units of product B, which allows the demand limit to be indefinitely increased without
it would have an effect; it wouldn't matter if the restriction were X210 or X2100. On the other hand, if the limit
from the demand outside of four units, so that the restriction was X2 4, it would be effective.
Any reduction below this limit of four units would reduce the profit. So, the
the limit interval of demand (that is, the independent term) of the third constraint is
four units to infinity and within this interval the dual price remains zero.
The sensitivity analysis has produced the following ranges for the values of the term
independent of the three restrictions:

SUMMARY

Ing. Efraín Murillo Msc.

Operations Research I 53

Dual prices express the marginal value of the change in the independent term of a
restriction, but these values are only maintained within certain intervals.
SENSITIVITY ANALYSIS: EVALUATION OF NEW PRODUCTS
Dual prices can help identify bottlenecks or costly restrictions that can be
modify cost-effectively. Dual prices can also be useful for evaluating new products.
Consider an extension of our [Link] research and development department of the company
It has created a product, C. It is very profitable ($10 per unit) but requires four hours on the machine.
1andthreehoursonmachine2,shouldunitsofproductCbeproduced?
Of course, the linear programming problem could be reformulated to include this.
new product, but a quick response can be obtained with the dual prices. The production of
Product C will require a reduction in the quantities of the other two products, as they all compete.
for the available time on the two machines. Remember that dual prices imply that one hour
Machine 1 costs two dollars and one hour of machine 2 costs one dollar. One unit of product C
requires four and three hours on the two machines, respectively. Then, the opportunity cost
the unit of product C is:
(Dual price of machine 1 hours) x (Hours required on machine 1)] +
(Dual price of machine 2 hours) x (Hours required on machine 2)
I mean $2 x 4 + $1 x 3 = $11
This represents the opportunity cost of producing A and B. The benefit of
Product C is only 10 dollars, so it should not be produced, since the opportunity cost
exceeds the unit benefit.
ANOTHER EXAMPLE. The same company has another item, product D, which requires one hour in
Each of the machines produces a unit profit of five dollars, should it be produced?
The opportunity cost of this product is:
$2 x 1 + $1 x 1 = $3
Since the benefit of five dollars per unit exceeds the opportunity cost of three dollars, it is necessary to
produce some units of product D.
This analysis does not indicate how many units of product D should be produced; it only indicates that it must
to be included in the product mix. The manager would reformulate the problem of
linear programming to include another decision variable for product D and then I would solve it
theproblemagain.

SUMMARY. The opportunity cost of a new product is calculated as the sum of:
(Dual price) x (Required units)
for all affected constraints. If the opportunity cost is less than the benefit per unit
of the new product, then it is profitable and should be included in the optimal solution. If the cost of
opportunity is greater than the profit per unit, then the product should not be manufactured.

SENSITIVITY ANALYSIS: FUNCTION COEFFICIENTS

OBJECTIVE

A manager may be interested in what can happen with the solution to a problem of
linear programming if one of the coefficients of the objective function changes, for example, due to a

Eng. Efraín Murillo Msc.

Operations Research I 54

increase in the price of raw materials. A graphical analysis similar to the one that was done can be carried out.
for changes in the coefficients of the constant term.
Assume that the profit per unit of product A remains fixed at six dollars, but that the
benefit per unit B, which is expected to be seven dollars may change. Figure 8 shows the
benefit functions if the benefit per unit of product B increases to eight dollars, nine
dollars and then to ten. When the coefficient of X2in the objective function is $8 (that is, profits of
eight dollars per unit of product B), the profit function is Z = 6X1+ 8X2and point D is
still the optimal solution. If the coefficient of X2it increases to $9, the slope of the profit function
is identical to the slope of the constraint of machine 1 and falls exactly on the line of
restriction. Therefore, there are several alternative optimal solutions and both D and H are vertices.
optimal. Assume that the coefficient of X2increase more, up to 10 dollars per unit, so that
The profit function is Z = 6X1+ 10X2, as shown in figure 8. Note that in the region
feasible there are points above the line (that is, with greater benefits), so point D already
it is not optimal. Now H is the only optimal solution. In other words, if the coefficient of X2exceeds
nine dollars, the optimal point moves from point D to H.

The analysis is the same if the coefficient of X is reduced.2.

Figure 9 shows the profit function for values of $5, $3, and $1 of the coefficient of X.2. With a
coefficient of 3, the profit function is P = 6X1+ 3X2, the profit function matches the
restriction of machine 2 and points C and D are optimal. If the coefficient is reduced further, to 1, the
optimal point moves to C.

Ing. Efraín Murillo Msc.

Operations Research I 55

The same analysis can be applied to the benefits of A, keeping the coefficient constant.
X2and modifying that of X1Although the graphical analysis is not displayed, the results indicate that the
coefficient of X1it can decrease from 6 to 4.67 before the optimal point changes to H. On the other hand,
the coefficient can increase from 6 to 14 before the optimal point shifts to C. These results
they can be summarized as follows:
Intervals of the coefficients of the objective function
(interval in which the basic optimal solution does not change)

Intervals of the constant term and intervals of the coefficients of the objective function.
Conceptually, the intervals of sensitivity analysis are similar for the coefficients of the
objective function that were previously developed and those corresponding to the coefficients of
independent term. However, there are important differences.
The optimal solution does not change in the interval specified for the coefficients of the function.
Objective; that is, the company must produce six units of product A and four units of
Product B. Outside the interval, the optimal solution changes abruptly to another vertex. Of course,
the total benefits vary according to changes in the coefficients, but once again, the
solution does not vary within the intervals.
On the other hand, in the intervals of the independent term, the solution changes, since the optimal point
it moves along one of the constraint lines. In our example, different occur.
quantities of products A and B as the optimal point moves. When reaching the limit of the interval, a
new vertex becomes the optimal solution.
RESUMEN:
The intervals, both for the coefficients of the constant term and for those of the function
Objective, they provide important information in the interpretation of a program's solution.
linear. The intervals of the constant term determine the limits within which it is
it maintains the dual price for each constraint. The intervals of the coefficients of the objective function
determine the limits within which the solution is the same.

Ing. Efraín Murillo Msc.

Operations Research I 56

PROPOSED PROBLEMS: SENSITIVITY ANALYSIS,

GRAPHIC APPROACH

Problem 1 Given the following P.L.

Max Z = 4X + 3Y
Sa:
2X + 3Y 18
3X + 2Y 18
Y 5
X 0,Y 0
Shade the convex polygon defined by the constraints.
b) Determine the ranges of variability of the coefficients c so that the optimal solution is the
same, and the variability ranges of the coefficients b where the dual prices remain
constants, also find the dual prices.

Problem 2 Solve the following P.L.

MinW= 5X + 4Y
To:
2X+Y 4
X+Y 3
X 0,Y 0
Shade the convex polygon defined by the constraints.
b) Determine the ranges of variability of the coefficients of the objective function so that the
the optimal solution remains the same, and the variability ranges of the coefficients b where the
dual prices remain constant, also find the dual prices.

Problem 3 Given the following P.L.:

Max Z = 4X +Y
Sa:
3X+Y 12
5X + 3Y 30
X 0,Y 0
Utiliceelmétodográficoparacontestarlaspreguntassiguientes:
a) Determine the variability intervals for each of the coefficients c for which the
theoptimalsolutionisthesame.
b) Determine the variability intervals of each coefficient b where the dual prices remain
constants.

Eng. Efraín Murillo Msc.

Operations Research I 57

Sensitivity Analysis:
COMPUTER SOLUTION

. Taken from WINSTON, Wayne L. Operations Research, Applications and Algorithms.

Mexico: Ibero-American Editorial Group, 1994.

Use of Lindo in Linear Programming Problems

LINDO (for the initials of Linear INteractive Discrete Optimizer) is a popular software package
to solve linear programming problems on PC. It is a standalone package, meaning that it does not
it is found in a spreadsheet.

Example: A company manufactures two products, A and B. Each requires time on two machines.
The first machine has 24 hours available and the second has 16. Each unit of product A
It requires two hours on both machines and each unit of product B needs three hours on the first one.
machine and one hour in the second. The profits are six dollars per unit of A and seven
dollars per unit of B, the company can sell all the units it manufactures of product A and
Due to market limitations, only 6 units of product B can be sold. Assume the objective is
maximize profit; How many units of products A and B should be produced?

In the following Table 1, the solution and comments for the problem are observed.

Table 1: Computer solution with the LINDO package

MAX 6X1 + 7X2 Formulation: X1 and X2 are decision variables
SUBJECT TO
2X1 + 3X2 <= 24 2) Machine 1 restriction
3) 2X1 + X2 <= 16 3) Machine 2 restriction
4) X2<=6 4) Market demand restriction
END

LP OPTIMUM FOUND AT STEP 2 LP optimum found after examining 2 vertices

OBJECTIVE FUNCTION VALUE

1) 64,000 The optimal utility is $64.00

VARIABLE VALUE REDUCED COST
X1 6.00 0.000 The optimal solution is 6 units of product A and 4
X2 4.00 0.000 units of product B; the reduced costs are of
zero because in both cases some amount is produced
of each product.
ROW SLACK OR SRPLUS DUAL PRICES
2) 0.00 2.00 There is no inactivity in restrictions 2 and 3.
3) 0.00 1.00 The non-use in restriction 4 implies demand not
4) 2.00 0.00 Used for product B. The dual prices are shown.
(shadow prices)
NRO INTERACTIONS = 2

RANGES IN WHICH THE BASIS IS UNCHANGED:

OBJ COEFFICIENT RANGES
VARIABLE CURRENT ALLOWABLE ALLOWABLE Ranges where the optimal solution is the same
COEF INCREASE DECREASE
X1 6.00 8.00 1.3333
X2 7.00 2.00 4.0000

Ing. Efraín Murillo Msc.

Operations Research I 58

RIGHTHAND SIDE RANGES

ROW CURRENT ALLOWABLE ALLOWABLE Ranges where the set of variables of the solution
RHS INCREASE DECREASE the basics are the same, moreover they are also ranges where
2 24.00 4.00 8.00 the dual prices are the same.
3 16.00 8.00 4.00
4 6.00 Infinite 2.00

Integral application example of LINDO

Let's consider a comprehensive example to review the concepts of linear programming that we
They were presented in this chapter. The company Lazy Life of Leisure (LLL) produces garden furniture.
The company has three products: armchairs, chairs, and tables. It requires 60 minutes of work to produce
a chair, 30 for a chair and 90 for a table. LLL has 9000 minutes of work and 600
wood units and can sell all the chairs and tables it produces. However, it cannot
Sell more than 50 armchairs. LLL makes a profit of 60 dollars for each armchair, 40 dollars for each
chair and 80 dollars per table. The formulation of LINDO and the solution for this case are shown in the
Table 2. The comments in the table describe the solution, the conclusions of the sensitivity analysis.
and the parametric analysis.

An interesting feature of this solution is that there is an alternative optimal solution. One can
see in the linear programming output that there is another solution because one of the variables of
decision, TABLES, has a value of zero in the solution and also a reduced cost of zero. This means
to say that some tables can be incorporated into the solution without affecting the benefit. From the
linear programming solution cannot determine what the alternative basic solution is, but it
knows that it exists.

Suppose that LLL is considering another product, for example, SWINGS. Each SWING requires
200 minutes of work and nine units of wood generate a profit of 100 dollars. Should it
to produce SWINGS?
The opportunity cost associated with a SWING is:
(dual price of labor) * (required labor)
(dual price of wood) * (wood required)
= 0*200 + 6.67*9 = $60

Since the benefit of 100 dollars exceeds this figure, SWINGS must be produced. It would have to
reformulate the problem and then solve it to know exactly how many SWINGS there are
what to produce.

Another feature shown in table 2 is the parametric analysis. Note that the restriction of
the amount of wood available (600 units) was effective and the dual price was 6.67 dollars
per unit. Assume that LLL can obtain additional wood. How much should they buy? The analysis
Parametric is designed to answer this question. Specifically, suppose that it is possible to
Buy 1000 units for 5000 dollars. Should they be purchased?

Note that the first 900 additional units (from 600 to 1500) have a dual price of $6.67 and
the next 100 units have a (dual price) value of 3.33. Therefore, the total value of the
1000 units is (900*$6.67) + (100*$3.33) = $6333. Since this figure is greater than the cost of 5000
dollars, we need to buy the wood.

RESUMEN
The optimal solution to a linear programming problem is a vertex of the feasible region.

Ing. Efraín Murillo Msc.

Operations Research I 59

The sensitivity analysis of the constraints first involves finding the dual price, marginal value.
from a unit change in the constant term of the restriction, and then find the interval of
values in which the dual price is maintained (variability range on the right side of the equations
of the restrictions).
The reduced cost is a marginal value analogous to the introduction of a unit of a
decision variable in the optimal solution.
The coefficient intervals of the objective function provide additional sensitivity analysis.
These intervals indicate the changes in the coefficients of the objective function that do not alter the
optimal solution.
Parametric analysis examines the changes in the solution and how the term can vary.
independent (RHS) of a constraint within a broad interval.
There are several software programs for solving linear programming problems.
including some existing within spreadsheets (EXCEL). The LINDO package is presented
as an example, but the outputs of other programs are similar.

Table 2 Computer solution Comments

1) MAX 60 ARMCHAIRS + 40 CHAIRS + 80 TABLES Problem formulation

SUBJECT TO A
60 ARMCHAIRS + 30 CHAIRS + 90 TABLES <= 9000 It is a restriction on the available working minutes.
3) 6 ARMCHAIRS + 6 CHAIRS + 12 TABLES <= 600 It is a restriction of the wood
4) ARMCHAIRS <= 50 It is the sales limit for the armchairs.
END

SOLVE Main menu command to start the solution

OPTIMAL LINEAR PROGRAMMING SOLUTION DETECTED IN STEP 2

VALUE OF THE OBJECTIVE FUNCTION

5000.00000 The benefit of the optimal solution is 5000 dollars

VARIABLE VALOR COSTO REDUCIDO The rows correspond to the restriction numbers.
ARMCHAIRS 50.00000 .00000 (row 1 is the objective function)
CHAIRS 50.00000 .00000 The solution requires 50 armchairs, 50 chairs, and no table.
TABLES .00000 .00000 Note that the reduced cost of the tables is zero, which
indicates that some tables can be produced without reducing the
benefit. This indicates that there is an alternative solution.

LOOSE FIT or EXCESS DUAL PRICES

2) 4500.00000 .00000
3) .00000 6.66667 The value of an additional unit of wood is 6.66667
4) .00000 20.00000 dollars and 20 dollars for an additional unit of demand
of armchairs.
NÚMERO DE ITERACIONES = 2

DO YOU WANT THE SENSITIVITY ANALYSIS?

YES Sensitivity analysis is requested

INTERVALS IN WHICH THE BASE DOES NOT CHANGE:

INTERVALS OF THE COEFFICIENTS OF THE OBJECTIVE FUNCTION

COEFFICIENT INCREASE REDUCTION

VARIABLE ACTUAL ALLOWED ALLOWED Intervals of the coefficients of the objective function
ARMCHAIRS 60.00000 INFINITE 20.00000 The program displays the increase or the decrease
CHAIRS 40.00000 20.00000 .00000 (but not the limit)
TABLES 80.00000 .00000 INFINITE

INTERVALS OF THE INDEPENDENT TERM (RHS)

Ing. Efraín Murillo Msc.

Operations Research I 60

RHS AUMENTO REDUCCIÓN Intervals for the restrictions. As before, the program
CURRENT LINE ALLOWED ALLOWED shows the increase or reduction, but not the limit.
2) 9000.00000 INFINITE 4500.00000
3) 600.00000 900.00000 300.00000
4) 50.00000 50.00000 50.00000

REPORTS/PARAMETRICS Request the parametric analysis.

ROW 3 The program requests the row for analysis. The selected one is
NEW VALUE OF THE INDEPENDENT TERM (RHS) line 3, the restriction of wood, as well as a limit
2000 superior to 2000 units.

VARIABLE VARIABLE FILA VALUE VARIABLE VALUE

OUTPUT OF INPUT RHS PIVOT DUAL OBJECTIVE There are three steps
CLEARANCE 2 CLEARANCE 4 2 1500.00 6.66667 11000.0 from 600 to 1500 units: dual price = 6.66667
ARMCHAIRS CLEARANCE 3 4 1800.00 3.33334 12000.0 from 1500 to 1800 units: dual price = 3.33333
ART ART 0 1800.00 .00000 12000.0 from 1800 to more units: dual price = 0.00000

Signs convention in LINDO

An important aspect of the LINDO computing program that tends to confuse users is
the convention used for the signs (positive or negative) of dual prices (values
marginal values or shadow prices) and the reduced costs. As a general rule, it is considered that making
the stricter a restriction is harmful (that is, it reduces benefits or increases costs), while
that the expansion of a restriction helps (increases benefits or reduces costs).
In the case of equality constraints, it is sometimes not clear if the constraint becomes stricter.
or looser. The following table should serve for its interpretation.

SIGN OF MAXIMIZATION OR Effect of an increase of one unit

DUAL PRICE MINIMIZACIÓN in the value of the constant term (RHS)
+ Maximization Increase in profits
- Maximization Reduction in benefits
+ Minimization Reduction in costs
- Minimization Increase in costs

Reduced costs have a value of zero or are positive. In minimization, the effect of including
one unit in the solution is the increase in costs; in maximization it is the reduction of
benefit.

DEGENERATION
In a linear programming problem with m constraints, the basic solution will have exactly m
basic or solution variables. In LINDO outputs, the basic variables can be determined with just
identify the decision variables or the slack or surplus variables that are different from zero.
Sometimes (in fact, frequently in large-scale linear programming problems)
It happens that there are fewer than m variables different from zero in the optimal solution. This means that by
At least one of the basic variables is zero, a situation known as degeneration.
Generally, degeneration is not a problem, but care must be taken in the interpretation of
some of the dual prices. Specifically, when examining the intervals of the constant term,
a dual price can have a zero value as an allowed increase, a zero allowed decrease, or
both. In other words, the dual price will remain only within a limited range or perhaps will not exist.
such interval.

Ing. Efraín Murillo Msc.

Operations Research I 61

ALTERNATIVE SOLUTIONS
Sometimes a linear programming problem can have more than one optimal solution. It is
It is easy to detect these cases in LINDO results if the solution does not present degeneration.
If there is a decision variable that has a value of zero and also a reduced cost of zero, this
indicates that there is an alternative solution. Furthermore, a constraint with slack (or surplus)
equaltozeroandwithadualpriceofzeroalsoindicatesanalternativesolution.

CASE STUDY: Sensitivity Analysis

A company sells its product through sales agents by means of sales visits to three types of
clients: Commercial, Industrial, and Professional. For each sales visit to a commercial client, you obtain
revenue from sales of $2,000, for each visit to an industrial client you earn $5,000 and for each visit to a
Professional client earns $10,000 from sales.
En el mes actual se dispone de 3.200 horas de los agentes vendedores para efectuar las visitas y de $ 10.000 para
travel expenses.
The management does not allow more than 20% of the time for sales visits to be allocated to visiting clients.
commercials, nor does it accept that more than 30% of the travel budget be used for client visits
professionals.
It takes 5 hours to visit a Commercial Client, 8 hours for an Industrial Client, and 11 hours for a Client.
Professional.
The travel expenses for each visit to a Commercial client are $10; $14 for each visit to an Industrial client and
$35 per client visit. The goal is to maximize sales revenue.
Formulation and construction of the model:
Decision Variables: X1, X2, and X3, representing visits to be made to Commercial, Industrial, and
Professionals respectively.
Objective Function: Maximize sales revenue.
Restrictions:
Restriction 1: Limited availability of hours to conduct visits
Restriction 2: Limited availability of $ for travel expenses.
Restriction 3: Requirement of no more than 20% of the time for visits to commercial clients.
Restriction 4: Requirement of no more than 30% of the budget for travel expenses on visits to professional clients.
In the solution obtained from the model using the LINDO program on the computer, the construction of
model.
MAX 2000 X1 + 5000 X2 + 10000 X3
SUBJECT TO
2) 5 X1 + 8 X2 + 11 X3 <= 3200
3) 10 X1 + 14 X2 + 35 X3 <= 10000
5 X1 <= 640
35 X3 <= 3000
END
LP OPTIMUM FOUND AT STEP 2
OBJECTIVE FUNCTION VALUE
2267857.00
VARIABLE VALUE REDUCED COST
X1 0.000000 1125.000000
X2 282.142900 .000000
X3 85.714290 .000000

Ing. Efraín Murillo Msc.

Operations Research I 62

ROW SLACK OR SURPLUS DUAL PRICES

2) .000000 625.000000
3) 3050.000000 .000000
4) 640.000000 .000000
5) 0.000000 89.285710
NO. ITERATIONS= 2
RANGES IN WHICH THE BASIS IS UNCHANGED:
OBJ COEFFICIENT RANGES
VARIABLE CURRENT ALLOWABLE ALLOWABLE
COEF INCREASE DECREASE
X1 2000.000000 1125.000000 INFINITY
X2 5000.000000 2272.727000 1800.000000
X3 10000.000000 INFINITY 3125.000000
RIGHTHAND SIDE RANGES
ROW CURRENT ALLOWABLE ALLOWABLE
RHS INCREASE DECREASE
2 3200.000000 1742.857000 2257.143000
3 10000.000000 INFINITY 3050.0000
4 640.000000 INFINITY 640.000000
5 3000.000000 6777.778000 3000.000000
Note: If you want to copy the model to display it along with the results, you must use the commands
necessary points outlined in the program manual.
RESULTS REPORT:
Optimal Solution:
X1 = 0 X2 = 282.142900 X3 = 85.714290
Interpreting those values, in light of the meanings of the variables in the model, we have the following:
Do not make visits to commercial clients, make 282.1429 visits to professional clients and 85.714290 visits to
industrial clients. The use of Integer Linear Programming would be advisable.
Objective Function.
The maximum income obtained from sales is 2,267,857.00 monetary units. (Remember that any
the amount placed after the decimal point in the results is a decimal number, since the programs use
English notation.
Clearances:
Restriction 1: No hours are available for visits; all available hours are used.
Restriction 2: From the total maximum budget for travel expenses, $3050 is available.
Restriction 3: The maximum required of 640 hours for visits to Commercials remains unused. Or also:
The maximum 640 hours required for visits to commercial clients are still available since
Visits to these clients are not made.
Restriction 4: Use the total maximum required amount of $ for travel expenses for visits to professional clients. Or
Also: There is no availability of the maximum allowances for these visits.

Sensitivity Analysis:
a) The ranges of variation for the coefficients of the variables in the Objective Function:
-  Cx1 3125 3200 Cx2 7272.727 6875 Cx3  
The income generated by each visit to a commercial client can vary between - infinity and 3125 and the solution
The basic will remain the same. Within that range, the value of total revenue, Objective Function, does not change.

Ing. Efraín Murillo Msc.

Operations Research I 63

Since the value of variable X1 is zero, therefore multiplying it by any value within that range does not
will have no effect on the target value.
The income generated by each visit to a professional client can vary between 3200 and 7272.727 and the solution
basic will remain the same. Within that range, the total income value, Objective Function, can
vary depending on the value that the income takes within that range. If it changes to 3,200 or 7272.727, it would exist
also an optimal alternate solution.
The income generated by each visit to an industrial client can vary between 6875 and and the basic solution
it will remain the same. Within that range, the value of total income, Objective Function, may vary.
depending on the value that income takes within that range. If it changes to 6875, there would also be a solution.
optimal alternative.

b) The ranges of variation within which the Basis does not change (RANGES IN WHICH THE BASIS IS
) indicates the following ranges for the right-hand sides of the constraints (RHS):
942.857 b1 4942.857 6.950 b2    0 b3  0 b 4 9777,7780
The maximum number of hours available for visits can vary between 942.857 and 4942.857 and the solution
the basics will remain the same.
The maximum amount of dollars available for per diem can vary between 6,950 and and the basic solution
it will remain the same.
The amount of 20% of hours required for visits to commercial clients can vary between 0 and and the solution
the basic will remain the same.
The amount of 30% of the travel budget required for visits to professional clients may vary between
0 and and the basic solution will remain the same.

Dual Variables:
For the first constraint, the value of 625 of the dual variable indicates that the optimal total income will increase.
for each hour that the maximum number of hours currently available for visits increases,
as long as that increase is used. Therefore, in the absence of further information, it would be
it is convenient to increase this maximum amount available. The maximum amount that can be increased is
the number of hours is 4942.857 as indicated by the sensitivity analysis of the solution. Beyond that
amount, the basic solution would change, by changing the amount of the dual variable and forcing an analysis
different.
For the second constraint, the value 0 of the dual variable indicates that the optimal total income will not increase.
for every monetary unit by which the maximum budget currently available is increased. Therefore, in
In the absence of more information, it would be irrelevant to increase this maximum amount available. The quantity
the maximum to which this amount can be theoretically increased is an infinite amount as indicated by the
sensitivity analysis of the solution. In practice, it refers to any finite amount.
For the third constraint, the value 0 of the dual variable indicates that the optimal total revenue will not vary for
every hour in which the maximum of 20% currently required to make customer visits increases
commercials. Therefore, in the absence of further information, it would be indifferent to increase this maximum amount.
For the fourth constraint, the value 89.28571 of the dual variable indicates that the optimal total income.
they will increase by that amount for each $ that the maximum of 30% currently required increases for
travel expenses for visits to be made to professional clients. Therefore, in the absence of further information, it would be
It is convenient to increase this maximum available amount. The maximum amount to which this can be increased
the amount is 9777.778 as indicated by the sensitivity analysis of the solution. Beyond that amount, the
basic solution would change, changing the amount of the dual variable and forcing a different analysis.
NOTE: The use of Integer Linear Programming is more convenient in this model.

Ing. Efraín Murillo Msc.

Operations Research I 64

Exercises: Sensitivity Analysis:

Computer Solution
CASO 1.-LaCompañía DAKOTAfabrica escritorios, mesas y sillas. Hacer cada uno de estos productos
requires wood and hours of touch-up and carpentry. The amount of resources needed to produce each one of
these products are shown in the attached table.

Resource Desk Table Chair

Wooden Tables 8 6 1
Hours Retouch 4 2 1.5
Hrs Carpentry 2 1.5 0.5

Currently, there are 48 wooden tables, 20 hours for retouching, and 8 for carpentry. A desk is
sells for US$60, a table for US$30, and a chair for US$20. DAKOTA thinks that the demand for desks and
Chairs are unlimited, but at most 5 tables will be sold. Since the available resources have already been purchased.
DAKOTA wants to maximize profits.

1. How much can the price of chairs vary for the production plan to remain optimal?
What happens when a desk can be sold for US $55?
What happens if only 30 wooden tables are available?
4. Dakota is considering making computer tables. A computer table is sold for US
$36 and it uses 6 wooden boards, 2 hours of touch-up work, and 5 hours of carpentry. Is it worth it for Dakota?
to make these tables?
5. Find and Interpret the dual prices.
6. Formulate the Dual Model and solve it.
7. Analyze in the Dual the effect of changing the selling price of the tables
8. Suppose Dakota wants to manufacture a new type of chairs that sell for $15 and use 1 board.
of wood, 1 hour of touch-up and 1 hour of carpentry. What happens to the optimal solution?

CASE 2.- An industrial company produces 4 models of Keyboards, each of which is treated in the
assembly and finishing departments. The number of man-hours of labor needed, by keyboard in
each department is:
Modelo1 Modelo2 Modelo3 Modelo4
Assembled 4 10 7 10
Finished 1 1 3 4
There are 6000 man-hours available in the assembly department and 4000 in the
finished in the next 6 months. The profits in dollars for each keyboard model are:
2.2, 5, 3 and 4 respectively for keyboards 1, 2, 3 and 4.
There are restrictions to produce at least 100 units of model 2, at most 200 units of model 3, and to
more than 250 units of model 4. The production of model 1 is unrestricted.
Let X1, X2, X3, and X4 be the variables of the problem that represent the number of keyboards of models 1, 2, 3, and 4.
respectively to produce in the next 6 months.
a) Formulate the Linear Programming model to determine how much the manufacturer should produce of each
model, so that the profits are maximized.
Using Lindo Software or WinQsb for your solution, answer the following:
b) The maximum utility of the company. What quantity is produced of each model?
c) The slack or excess of assembly and finishing hours.

Ing. Efraín Murillo Msc.

Operations Research I 65

d) How many hours of assembly are used in the production of model 1? How many finishing hours are used?
used in the production of model 2?
e) What happens to current utility if the minimum demand for model 2 increases or decreases by 1 unit?
What range of variability is this value valid in?
f) How much should they pay at most to increase an extra hour of assembly? How much for one hour?
extra finish?, within what ranges of variation are these values valid?
g) How much should be invested at most to hire 1000 overtime assembly hours?
h) How much must the unit profit of model 3 improve to justify its production?, How much the
of model 4?
i) What ranges can the unit utility of model 1 and 2 vary within, without affecting the plan of
current production?
j) A fifth model is under consideration. It requires 4 hours of assembly and 5 hours of finishing.
The profit per unit is 2 dollars. Should it be produced? Why?

CASE 3.- A factory produces 4 products: A, B, C, and D. Each unit of product A requires two hours of
machining, one hour of assembly and costs $10 in work in process inventory. Each unit of product B requires
one hour of machining, three hours of assembly, and it is worth $5 in work in progress inventory. Each unit of C needs
2 ½ hours of machining, 2 ½ hours of assembly, and it costs $2 in work in progress inventory. Finally, each unit
Product D requires five hours of machining, does not need assembly time, and costs $12 in inventory.
in process.
The factory has 1200 hours of machining and 1600 hours of assembly. Furthermore, it cannot allocate more than
$10,000 in work in progress inventory. Each unit of product A has a profit of $40; each unit of B, of
$24; each unit of C, $36, and each unit of D, $23. No more than 200 units of A can be sold; not
more than 160 of C, and any amount of products B and D can be sold. However, to comply with
a contract, at least 100 units of product D must be produced and sold. The factory's goal is
maximize the resulting utility from the sale of the four products.
Answer the following questions using the output from LINDO or WinQsb.
a) How many units of each product should be produced? What is the optimal profit?
b) What is the slack or surplus of each of the resources?
c) How much should be paid at most for an additional hour of machining? How much for one hour?
additional mounting? How much does an additional $1 generate in working capital for inventory?
process?
d) Suppose that the company can invest money in advertising in order to increase the demand for the product
How much could the company invest to increase the demand for product A in one unit?
e) If the contract for product D were for 130 units (instead of 100), what effect would it have on the
utility?
f) What would happen to the production plan and optimal profit if the unit profit of product A
Does it decrease by $5? And what would happen if it decreases by $10?
g) For question f. What is the range of variability of the coefficients of the objective function, within
of which the production plan does not vary?
h) How much should the unit profit of product C improve to justify its production.
A new product, product E, is under consideration. It requires 2 hours of machining, 5 hours of
assembly and $20 in working capital. The profit per unit is $50. Should any unit be produced?
of product E?

CASE 4.- A television manufacturer produces four models: A, B, C, and D, A and B are black and white models, and C and
It gives colored models. Each television requires time for assembly and testing. Table 1 indicates the requirements.
of assembly and testing for each model, as well as the amount of time available for each operation. In addition, by
Due to a strike, there is a shortage of kinescopes and the supplier indicates that they will not be able to provide more than 180.
Cinescopesnextmonthand,ofthese,atmost100canbecolored.

Table1 Model Total

A B C D available
Tiempo de montaje (horas) 8 10 12 15 2000
Testing time (hours) 2 2 4 5 500

Eng. Efraín Murillo Msc.

Operations Research I 66

Beneficio marginal (dólares) 40 60 80 100

The problem is formulated and solved in Table 2. The decision variables Xa, Xb, Xc, and Xd represent the
number of units produced for the model.
Responda las siguientes preguntas con base en los resultados de LINDO de la tabla 2.

a) Cuál es el programa de producción óptimo para el fabricante de televisores? ¿Cuál es el beneficio máximo?
b) What is the marginal value of an additional hour of assembly time? In what interval is this valid?
marginal value?
c) Suppose that 80 additional hours of testing time can be obtained at a cost of four dollars each.
Time? Should it be done? What would be the increase in benefits?
d) Suppose that a price change is instituted that modifies the marginal benefit of model D, from 100 to
130 dollars. Would this change the optimal production plan? Suppose the price of model D changes.
from 100 to 160 dollars. In this case, would the production plan vary?
e) How much would the price of model A have to change in order to justify its production?
f) Management wants to introduce a new color model (model E). Model E would only require 10 hours of
assembly and 3 hours of testing. The marginal benefit of the model would be 70 dollars. Should it be produced the
new model? If so, what would be the marginal value of the production of one unit of model E?

Table 2
Max 40Xa + 60Xb + 80Xc + 100Xd
s.a.
2) 8Xa + 10Xb + 12Xc + 15Xd <= 2000
3) 2Xa + 2Xb + 4Xc + 5Xd <= 500
4) Xa + Xb + Xc + Xd <= 180
5) Xc + Xd <= 100
End.
OBJECTIVE FUNCTION VALUE
1) 12500.00
VARIABLE VALUE REDUCED COST
Xa 0.000 10,000
Xb 125,000 0.000
Xc 0.000 0.000
Xd 50,000 0.000

ROW SLACK OR SURPLUS DUAL PRICES

2) 0.0000 5.0000
3) 0.0000 5.0000
4) 5.0000 0.0000
5) 50.0000 0.0000

NO. ITERATIONS= 5

RANGES IN WHICH THE BASIS IS UNCHANGED:

OBJ COEFFICIENT RANGES
VARIABLE CURRENT ALLOWABLE ALLOWABLE
COEF INCREASE DECREASE
Xa 40.00 10,000 INFINITY
Xb 60.00 6.667 20,000
Xc 80.00 0.000 INFINITY
Xd 100.00 50,000 0.0 00

RIGHTHAND SIDE RANGES

ROW CURRENT ALLOWABLE ALLOWABLE
RHS INCREASE DECREASE
2 2000.00 33.333 500.00
3 500.00 100,000 20.00
4 180.00 INFINITY 5.00
5 100.00 INFINITY 50.00

CASE 5.-HAL produces two types of computers: PC and VAX. The computers are manufactured in two locations:
New York and Los Angeles. New York can produce up to 800 computers, and Los Angeles up to 1000.
HAL can sell up to 900 PCs and 900 VAXes. The profits associated with each production and sales site are
los siguientes: Nueva York, PC, 1000 dólares; VAX, 800 dólares; Los Ángeles, PC, 1000 dólares; VAX, 1300

Eng. Efraín Murillo Msc.

Operations Research I 67

dollars. The time of specialized labor required to build a computer at each site is given at
continuation: New York, PC, 2 h; VAX, 2 h; Los Angeles, PC, 3 h; VAX, 4 h. A total of 4000 is available.
h of work. Let it be:
XNP = PCs produced in New York
XLP = PCs produced in Los Angeles
XNV= VAX produced in New York
XLV= VAX produced in Los Angeles
So we get the output from LINDO from the following figure. Use this output to answer the questions.
following:
a) If there were 3000 hours of specialized labor time, what would HAL's profit be?
b) Suppose an external contractor offers to raise the capacity of New York to 850 computers, at a
cost of 5000 dollars. Would HAL have to accept the contractor?
c) How much would the profit have to increase for a VAX produced in New York, for HAL?
Will it consider producing VAX computers in that city?
d) What is the maximum amount that HAL would have to be willing to pay for an extra hour of
work?

LP OPTIMUM FOUND AT STEP 5

OBJECTIVE FUNCTION VALUE

1582500.

VARIABLE VALUE REDUCED COST

XNP 800.000000 0.000000
XLP 100.000000 0.000000
XNV 0.000000 175.000000
45 525.000000 0.000000

ROW SLACK OR SURPLUS DUAL PRICES

2) 0.000000 325.000000
3) 0.000000 325.000000
4) 375.000000 0.000000
5) 0.000000 25.000000
6) 375.000000 0.000000

NO. ITERATIONS= 5
RANGES IN WHICH THE BASIS IS UNCHANGED:

OBJ COEFFICIENT RANGES

VARIABLE CURRENT ALLOWABLE ALLOWABLE
COEF INCREASE DECREASE
XNP 1,000,000.00 INFINITY 175.000000
XLP 1,000,000.00 175.000000 25.000000
XNV 800.000000 175.000000 INFINITY
XLV 1300.000000 33.333332 233.333328

RIGHTHAND SIDE RANGES

ROW CURRENT ALLOWABLE ALLOWABLE
RHS INCREASE DECREASE
2 4000.000000 1,500,000 2100.000000
3 800.000000 100.000000 500.000000
4 1000.000000 INFINITY 375.000000
5 900.000000 700.000000 100.000000
6 900.000000 INFINITY 375.000000

CASE 6.- "JC" S.A. wants to spend $1500 on local advertising. The overall objective is to achieve the maximum
possible audience while reaching at least 9000 children and 1000 elderly people.
There are 3 media available; their costs and the audience they have are given in the following table:

Ing. Efraín Murillo Msc.

Operations Research I 68

NEWSPAPER RADIO T.V.

COST PER AD ($) 200 100 500
TOTAL AUDIENCE 31000 15000 50000
CHILDREN 1000 1000 3000
GRANDPARENTS 200 100 500
Use the output from Lindo (next table) to address the following concerns:
a) What is the optimal advertising mix? And how much audience would be reached?
b) How much would the audience increase with an additional dollar in the budget? What is the range for this
Is marginal value valid?
"JC" is worried since the total audience of the Radio is only approximate. About what range of
Will the total audience remain on the radio in the solution?
d) How much would the total audience of television have to increase for it to become profitable?
consideration in the optimal mix of advertising?
e) If the total audience of the newspaper decreased by 1000, what would be the new optimal solution for 'JC'?
f) A senior year is under consideration. It requires a budget of 350 dollars, reaches 2000 children and 400.
Grandparents. The total audience is 40,000. Should it be considered? Why?
g) Show the dual PPL and define only the variable associated with the capital constraint.

Output of the Lindo for case 6:

Where P, R, and T are: the number of advertisements in Newspaper, Radio, and TV, respectively.

Max 31000P+15000R+50000T Total audience for 'JC'

St
200P + 100R + 500T <= 1500 Capital budget constraint
1000P + 1000R + 3000T >= 9000 Minimum audience restriction for children
200P + 100R + 500T >= 1000 Minimum audience restriction for grandparents

LP OPTIMUM FOUND AT STEP 0

OBJECTIVE FUNCTION VALUE

231000.0

VARIABLE VALUE REDUCED COST

P 6.000000 0.000000
R 3.000000 0.000000
T 0.000000 27000.000000

ROW SLACK OR SURPLUS DUAL PRICES

2) 0.000000 160.000000
3) 0.000000 -1.000000
4) 500.000000 0.000000

NO. ITERATIONS= 0

RANGES IN WHICH THE BASIS IS UNCHANGED:

OBJ COEFFICIENT RANGES

VARIABLE CURRENT ALLOWABLE ALLOWABLE
COEF INCREASE DECREASE
P 31000.000000 INFINITY 1000.00000
R 15000.000000 500.000000 27000.00000
T 50000.000000 27000.000000 INFINITY

RIGHTHAND SIDE RANGES

ROW CURRENT ALLOWABLE ALLOWABLE
RHS INCREASE DECREASE
2 1500.000000 300.000000 500.000000
3 9000.000000 6000.000000 1500.000000
4 1,000,000.00 500.000000 INFINITY

Ing. Efraín Murillo Msc.

Operations Research I 69

CASE 7.- A jewelry store produces three models of jewelry (rings, bracelets, and necklaces). To do this, it uses two types of
materials (gold and silver) of which there are 4000 and 6000 grams available respectively. The requirements
The material costs per unit for the three models are given below:

Material ring bracelet necklace Availability

Gold 2gr. 3gr. 5gr. 4000 g.
Silver 4 2 7 6000 g.
Requerimiento(hrs) 10 5 3 7000 hours

A market study indicates that the minimum demand for the three models is 200, 200 and
150 units, respectively. The unit profits are: 30, 20, and 50 soles for a ring.
bracelet and necklace respectively.
Use the Lindo or WinQsb software and use its output to answer the following questions:
a) ¿Cuál es el plan de producción óptimo?.y Cuál es la utilidad máxima?.
b) What is the maximum amount that the jewelry store should be willing to pay for an extra hour?
What range of values is this value valid for?
c) If the profit per collar were S/. 20 (instead of S/. 50), how would it affect the production plan?
and current optimal profit?
What would happen to the company's optimal production plan if the profit from a ring triples?
What happens with the optimal utility?
e) What amount of gold and what amount of silver does the jewelry use?
f) How many man-hours are spent on the production of rings?
g) If the minimum demand contract for the collars were doubled, what effect would it have on the profit?
the company?
h) If a new product (Lapels) is considered that requires 3 grams of gold, 2 grams of silver, and 10
working hours and the profit per unit is S/. 45, should it be produced?
i) Obtain the DUAL LP model, define its variables, and show its optimal solution.

Eng. Efraín Murillo Msc.

Operations Research I 70

4 Dual Analysis

. Taken from WINSTON, Wayne L. Operations Research, Applications and Algorithms.

Mexico: Ibero-American Publishing Group, 1994.

Obtain the dual of a PL

Associated with a PL, there is another PL, called dual. Understanding the relationship between a PL and its dual is very important.
for the understanding of advanced topics in linear and nonlinear programming. This relationship is important
because it provides interesting economic interpretations. The knowledge of duality also
provides other perspectives on sensitivity analysis.
When we take the dual of a given LP, we refer to a given LP as the primal. If the primal is a problem
max, the dual will be a min. problem, and vice versa. This means that if one of the LPs is a min. problem, the
another will be a max problem. For our convenience, we define the variables for the max problem
z , x1,…and
x, 2 xfor
n
the min problem. w , y1y, 2 ,…. y n
We start by explaining how to obtain the dual of a max problem, in which it is required that all the variables.
no negative ones and all the constraints are less than or equal to restrictions (normal max problem). It can be written as a
maximum normal problem as:

maxz c1 x1 c2x2  cn xn

s.a a11x1 a12 x2  a1nxn b 1

a21x1 a22 x2  a2nxn b2 (1)

am1 x1 am2x2  amn xn bm

x j 0 ( j 1,2,...,n)

The dual of a maximum normal problem, like (1), is defined as

mniw b1y1 b2y 2  bny n

s.a a11y1 a12y 2  a1ny n c1
a 21y1 a 22y 2  a 2ny n c 2 (2)

a m1y1 a m2y 2  a mny n cm

yi 0 (i 1,2,...,m)

A min problem, like (2), that has all the constraints ≥ and all the variables non-negative, is called
min problem normal. If the primal is a normal min problem, like (2), we define the dual of (2) as (1).

OBTAIN THE DUAL OF A NORMAL MAX OR MIN PROBLEM

A table facilitates the obtaining of the dual of a LP. If the primal is a standard max program, it will be possible to
find it using Table 1 horizontally; the dual will be obtained using Table 1 vertically. In
In a similar way, if the primal is a standard minimization problem, it can be obtained using Table 1 vertically; the
dual will be found using the same table horizontally. It illustrates the use of this table, obtaining the dual
the Dakota problem and the food problem. The Dakota problem is:

Ing. Efraín Murillo Msc.

Operations Research I 71

maxz 60x1 30x2 20x3

s.a8x1 6x 2 x3 48 Wood restriction
4x 1 2x 2 1.5x3 20 (Finish Restriction)
2x 1 1.5x2 0.5x3 8 (Carpentry Restriction)
x1 ,x2,x3 0

Where:
x1 produced desks
x2 produced tables
x 3 chairs produced

TABLES

Max z
Min. w (x 1 0) (x 2 0) … (x n 0)
x1 x2 xn
a1n b1
( y 1 0)
y1 a11 a12 …
a 21 a 22 a 2n b2
( y 2 0)
y2 …
… bm
ym a m1 am2 a mn
( y m 0)
c1 c2 cn

Table 1 Find the dual problem of max or min.

We obtain the Dakota problem using Table 2 horizontally, using the format of Table 1.
Then, using Table 2 vertically we will obtain Dakota's dual as:

mniw 48y 1 20 years 2 8y 3

s.a8y1 4 y 2 2 y 3 60
6 y 1 2 y 2 1.5y3 30
y1 1.5y2 0.5y3 20
y1 ,y2,y3 0

Max z
Min. w (x 1 0) (x 2 0) (x 3 0)
x1 x2 x3
( y 1 0) 48
y1 8 6 1
20
( y 2 0) y2 4 2 1.5
8
( y 3 0) y3 2 1.5 0.5
60 30 20

Table 2 Obtain the dual of the Dakota problem

The tabular method for finding the dual clearly shows that the i-th constraint of the dual corresponds to the
i-th primal variable For example,
xi the first constraint of the dual corresponds to (desks), because
x1
each number in the first dual constraint comes from the commune of x1 (desk) of the primal. Similarly the
the second dual constraint corresponds to x2 (months), and the third restriction of the dual corresponds to x 3 (chairs).
In a similar way, the dual variable y I it relates to the i-th constraint of the primal. For example y1 if

Ing. Efraín Murillo Msc.

Operations Research I 72

it relates to the first constraint of the primal (the wood constraint), because each coefficient of y1 in the
dual sale of the restriction of wood or the availability of wood. The importance of these
correspondences between the primal and the dual will be clarified later.
Now we find the dual of the feeding problem. As the feeding problem is a
My problem, we continue with the agreement to use w to indicate the objective function y,y 1,y2,y3y 4 y for
the variables. Then the problem of feeding can be written as:

mniw 50y 1 20 years 2 30 years 3 80y 4

s.a400y1 200y 2 150y 3 500y 4 500 Calorie restriction
3y 1 2 y 2 6 (Chocolate restriction)
2 y 1 2 y 2 4 y e a r s 3 4 y 4 10 (Sugar restriction)
2 y 1 4 y 2 y 3 5y 4 8 (Fat restriction)
y1,y2,y3,y4 0
Where
y1 biscuits consumed daily
y 2 chocolate ice cream scoops consumed daily
y 3 bottles of cola soda consumed daily
y 4 slices of pineapple cheesecake consumed daily

As it is primal, it is a normal min problem, we can find it vertically in Table 3, and its dual
horizontally in the same table. We then obtain the dual of the feeding problem, as

maxz 500 times1 6x 2 10x 3 8x 4

A400X1 3x 2 2x 3 2x 4 50
200 times1 2x 2 2x 3 4x 4 20
150x1 4x 3 1x4 30
500x1 4x 3 5x 4 80
x1,x2,x3,x4 0

Max z
Min. w (x 1 0) (x 2 0) (x 3 0) (x 4 0)
x1 x2 x3 x4
( y 1 0) y1 400 3 2 2 50
( y 2 0) 200 2 2 4 20
y2
( y 3 0) 150 0 4 1 30
y3
500 0 4 5 80
( y 4 0) y4
500 6 10 8

Table 3 Obtain the dual of the feeding problem

Like in the Dakota problem, we observe that the i-th dual constraint corresponds to the i-th
variable of the primal. For example, one can consider the third constraint of the dual as the constraint of
refreshment. Also, the i-th variable of the dual corresponds to the i-th constraint of the primal. For example, it
you can consider x 3 (the third variable of the dual) like the sugar variable.

Ing. Efraín Murillo Msc.

Operations Research I 73

OBTAIN THE DUAL OF A NON-NORMAL PL

Unfortunately, many LPs are not standard max problems or standard min problems. For example,
Applying the rules we have:
Max z 3x 1 4x 2
Max z 3x 1 4x 2
s.a. 2x1 5x 2 40
s.a. 2x1 5x 2 40
3x 1 2x 2 50
3x 1 2x 2 50
3x 1 2x 2 50
8 times 1 3x 2 25
 8x 1 3x 2 25
x1 ,x2 0
x1 ,x2 0
in this case, it is necessary to present the
mathematical model in any of
the previous forms, for that they Changing to the address of the
recommend using the following rules restriction 3, there would be a normal LP:
of transformation:
Max from 3x 1 4x 2
a) Max z Min z s.a. 2x1 5x 2 40
b) a ij x j b i
 a x ij b j i 3x 1 2x 2 50
c) a ij x j b i
a x by
ij j i  3x 1 2x 2 50
a ij x j b I
 8x 1 3x 2 25
x1 ,x2 0

ECONOMIC INTERPRETATION OF THE DUAL PROBLEM

INTERPRETATION OF THE DUAL OF A MAX PROBLEM

The dual of the Dakota problem was

mniw 48 years 1 20 years 2 8y 3

s.a8y1 4 y 2 2 y 3 60 (Desk restriction)
6 y 1 2 y 2 1.5y3 30 (Table restriction) (3)
y1 1.5y2 0.5y3 20 (Seat restriction)
y1 ,y2,y3 0

The first dual constraint relates to the desks, the second dual constraint relates to the tables; and the
third dual restriction with the chairs. Furthermore, it isy1related to wood; with the finishing
y 2 hours; and
y 3 with the hours of carpentry. The information relevant to Dakota's problem is shown in Table 4.
Now we are ready to interpret the dual of Dakota (3). Suppose a businessman wants to buy all
the resources of Dakota. So the businessman has to determine the price he would be willing to pay for
one unit of each of the resources of Dakota. With this in mind, we define

y1 price paid for 1 p.t. of wood

y 2 = price paid for 1 hour of finishing
y 3 price paid for 1 hour of carpentry

Ing. Efraín Murillo Msc.

Operations Research I 74

We now demonstrate that the prices of the resources would need to be determined y1y, 2 y y 3 , solving the
Dakota dual (3). The total price that would have to be paid for these resources is 48y1 20y2 8y 3 Since
it is necessary to minimize the cost of purchasing resources,

mniw 48y1 20y2 8y 3

This is the objective function for the dual of Dakota.

When setting the prices of resources,

RESOURCE/PRODUCT AVAILABLE QUANTITY

OF THE RESOURCE
Resource Desktop Table Chair
Wood 8 p.m. 6 p.m. 1 p.t. 48 p.t.
Finished 4h 2h 1.5 hours 8 PM
Carpentry 2h 1.5h 0.5 hours 8h
Selling Price 60 dollars 30 dollars 20 dollars

Table 4 Information about the Dakota problem

What are the restrictions faced by the entrepreneur? Prices of resources need to be quoted.
high enough to persuade Dakota to sell her resources. For example, the businessman has to offer
at least 60 dollars for a combination of resources that includes 8 feet of lumber, 4 hours
of finishing, and 2 hours of carpentry, because Dakota could, if they wanted to, use these resources to produce a
desk, which can be sold for 60 dollars. Therefore, the eventual buyer has to pay at least 60
dollars for the combination of resources required to manufacture a desk, otherwise Dakota would have none
reason to sell these resources. As the entrepreneur offers 8y 1 4 y 2 2 y 3 for the resources used to
to produce a desk, you have to choose y1y, 2y 3 y to satisfy
8y 1 4 y 2 2 y 3 60

But this is exactly the first constraint (of the desks) of Dakota's dual. A similar reasoning
It shows that you have to pay at least 30 dollars for the resources used to make a table (6 feet)
wood table, 2 hours of finishing, and 1.5 hours of carpentry). This means that y1y, 2 y y 3 has to
satisfy

6 y 1 2 y 2 1.5y3 30

This is the second constraint (of the tables) of Dakota's dual.

Similarly, the third constraint (of the chairs) of the dual,

y1 1.5y2 0.5y3 20

It indicates that you have to pay at least 20 dollars (the price of a chair) for the required resources to
to produce a chair (1 foot of wood, 1.5 hours of finishing and 0.5 hours of carpentry). It is also necessary to
satisfy the sign constraints y1 0,y2 0y3 y 0 Putting everything together, we observe that the solution of the dual

from the Dakota problem provides prices for wood, finishing hours, and carpentry hours. The
the previous analysis also shows that the i-th variable of the dual corresponds effectively, in a natural way,
to the i-th constraint of the primal.
In summary, when the primal is a standard max problem, the dual variables relate to the value of the
resources available for those who make decisions. For this reason, it often refers to the variables of
dual like the shadow prices of resources. In the next section, the resources are studied more deeply.
shadow prices.

INTERPRETATION OF THE DUAL OF A MINIMIZATION PROBLEM

To interpret the dual of a minimization problem, we consider the dual of the feeding problem. In the section
previously, we found that the dual of the feeding problem was:

Eng. Efraín Murillo Msc.

Operations Research I 75

maxz 500x1 6x 2 10x 3 8x 4

s.a400x1 3x 2  2x 3 2x 4 50 (Restriction of the cake)
200x1 2x 2 2x 3 4x 4 20 (Ice cream restriction) (4)
150 times1 4x 3 1x4 30 (Restriction on cola refreshment)
500x1 4x 3 5x 4 80 (Restriction of the cheesecake)
x1,x2,x3,x4 0

Table 5 contains the data for the food problem. To interpret (4), assume that Candice is
a seller of 'nutritional substances' who offers calories, chocolate, sugar, and fat. Candice wants
make sure that a dietitian meets your daily requirements by purchasing calories, sugar, fat with her
chocolate. Then Candice has to determine:

x1 Price per calorie that must be charged to the dietitian

x2 Price per ounce of chocolate
x 3 Price per ounce of Sugar
x 4 Price per ounce of Fat

Candice wants to minimize her income by selling the daily ration of nutrients to the dietitian.
required. Since Candice will receive 500x1 6x 2 10x3 8x 4 cents as profit with the dietitian, the
Candice's objective is:
maxz 500x1 6x 2 10x3 8x 4
This is the objective function of the dual of the nutrition problem. But by fixing the prices of the substances
nutritious, Candice needs to set the prices low enough so that the dietitian has an interest
economical in buying all the nutrients with it. For example, when buying a cake for 50
cents, the dietitian can obtain 400 calories, 3 oz. of chocolate, 2 oz. of sugar, and 2 oz. of fat. From this
In any case, Candice cannot charge more than 50 cents for this combination of nutrients. This will arrive
to the following restriction (cake restriction):

4i0t0e1s
m 3x 2 2x 3 2x 4 50

CALORIAS CHOCOLATE AZÚCAR GRASA PRICE

Ounces Ounces (Onzas) (Centavos)
Sponge cake 400 3 2 2 50
Ice cream 200 2 2 4 20
Soft drink 150 0 4 1 30
Cheesecake 500 0 4 5 80
Requirements 500 6 10 8

Table 5 Relevant information on the issue of food

What is the first constraint in the dual of the food problem. A similar reasoning provides the
second restriction (ice cream restriction), the third restriction (cola soda restriction), and the fourth
restriction (cheesecake restriction). Once again, the sign constraints must be satisfied. x1 0,
x2 0x,3 0 , andx4 0 .
The analysis shows that the seventh values can be interpreted as x I as a price for one unit of the
nutritional substance associated with the i-th dual constraint. In this way, x1 it would be the price of 1 calorie,x2
it would be the price of 1 ounce of chocolate and so on. Furthermore, we see that it is reasonable to relate the i-
the i-th variable ofxthe
I
dual with the i-th of the primal.
In summary, we have shown that when the primal is a standard max problem, or a standard min problem, the
The dual problem has an intuitive economic interpretation.

Ing. Efraín Murillo Msc.

Operations Research I 76

5 The SIMPLEX method for solving

Linear Programming Problems

Taken from [Link]ó[Link]

It is an iterative procedure that allows for improving the solution at each step. The process concludes
when it is no longer possible to continue improving that solution.

Starting from the value of the objective function at any vertex, the method consists of searching
successively another vertex that improves upon the previous one. The search is always madeand through the sides
of the polygonor of the edges of the polyhedron, if the number of variables is greater). As the number of
vertices (and edges) is finite, a solution can always be found.
The simplex method is based on the following property: if the objective function, f, does not take its value
maximum at vertex A, then there is an edge that goes from A, along which it increases.

The simplex method was created in 1947 by the mathematician George Dantzig.
The simplex method is primarily used to solve linear programming problems in which
three or more variables intervene.
Matrix algebra and the Gauss-Jordan elimination process for solving a system of equations
Linear equations form the basis of the simplex method.

With a view to understanding the methodology applied in the SIMPLEX Method, we will solve the
next problem:
Maximize Z = f(x,y) = 3x + 2y
subject to: 2x + y 18
2x + 3y 42
3x + y 24
x 0, y 0
The following phases are considered:
1. Convert inequalities into equalities
A slack variable is introduced for each of the constraints, to convert them into
equalities, resulting in the system of linear equations:

2x + y + s1 = 18

2x + 3y + s2 = 42

3x + y + s3 = 24

2. Set the objective function to zero

- 3x - 2y + Z = 0

Ing. Efraín Murillo Msc.

Operations Research I 77

3. Write the initial simplex table

In the columns, all the variables of the problem will appear, and in the rows, the coefficients of the
equalities obtained, one row for each constraint and the last row with the coefficients of the function
objective:

Table I. Iteration No. 1

Base Variable de decisión Variable de holgura Valores solución
x y s1 s2 s3
h 2 1 1 0 0 18
s 2 3 0 1 0 42
d 3 1 0 0 1 24
Z -3 -2 0 0 0 0

4. Find the decision variable that enters the basis and the slack variable that exits.
base
A. To choose the decision variable that enters the basis, we look at the last row, that of
the coefficients of the objective function and we choose the variable with the negative coefficient
greater (in absolute value).
In our case, the coefficient variable is -3.
If there were two or more equal coefficients that meet the previous condition, then
choose any one of them.
If there are no negative coefficients in the last row, it means that the
optimal solution. Therefore, what will determine the end of the application process of
simplex method, is that in the last row there are no negative elements.
The column of the variable that enters the base is called the pivot column (In a bluish color).

B. To find the slack variable that has to leave the base, each term is divided by
the last column (solution values) by the corresponding term of the pivot column,
as long as the latter are greater than zero. In our case:
18/2 [=9] , 42/2 [=21] y 24/3 [=8]
If there were any element less than or equal to zero, that quotient is not made. In the case of
that all elements were less than or equal to zero, then we would have a solution no
bounded and cannot continue.
The term of the pivot column that in the previous division gives rise to the smallest quotient.
Positive, the 3, already 8 is the smallest, indicates the row of the slack variable that leaves the base, d.
This row is called the pivot row (In bluish color).
If when calculating the quotients, two or more are equal, it indicates that any of the variables
corresponding ones can leave the base.
C. At the intersection of the pivot row and pivot column, we have the operational pivot element,
3.

5. Find the coefficients of the new table.

Ing. Efraín Murillo Msc.

Operations Research I 78

The new coefficients are obtained by dividing all the coefficients of the row by the pivot.
operational, 3, which is the one that needs to be converted into 1.

Next, by using Gaussian elimination, we make the remaining terms of its column zero.
with which we obtain the new coefficients of the other rows including those of the objective function
Z.
It can also be done using the following scheme:
Pivot row:
New pivot row = (Old pivot row) / (Pivot)
Rest of the rows:
New row = (Old row) - (Coefficient of the old row in the entering variable column) X
(New pivot row)
Let's see it with an example once the pivot row (x row in Table II) has been calculated:

Old row of s 2 30 1 0 42
- - - - -
Coefficient 2 22 2 2 2
x xx x X x
New pivot row 1 1/3 0 0 1/3 8
= == = = =
New row of s 0 7/3 0 1 -2/3 26

Table II. Iteration No. 2

Base Variable de decisión Variable de holgura Valores solución
x y h s d
h 0 1/3 1 0 -2/3 2
s 0 7/3 0 1 -2/3 26
x 1 1/3 0 0 1/3 8
Z 0 -1 0 0 1 24

Since there is one negative element in the last row, -1, it means that we have not reached yet
the optimal solution. The process needs to be repeated:
The variable that enters the basis is y, as it is the variable that corresponds to the coefficient -1.
B. To calculate the variable that comes out, we divide the terms of the last column by the
corresponding terms of the new pivot column:
["2:1/3 [=6]","26:7/3 [=78/7]","8:1/3 [=8]"]
And since the smallest positive quotient is 6, we have that the slack variable that exits is h.
C. The pivot element, which now needs to be made 1, is 1/3.

Eng. Efraín Murillo Msc.

Operations Research I 79

Operating in a manner similar to the previous one, we obtain the table:

Table III. Iteration No. 3

Base Variable de decisión Variable de holgura Valores solución

x y h s d
y 0 1 3 0 -2 6
s 0 0 -7 0 4 12
x 1 0 -1 0 1 6
Z 0 0 3 0 -1 30

Since there is one negative element in the last row, -1, it means that we have not yet reached
the optimal solution. The process needs to be repeated:
A. The variable that enters the base esd, because it is the variable corresponding to the coefficient -1
B. To calculate the outgoing variable, we divide the terms of the last column by the
corresponding terms of the new pivot column:
["6/(-2) = -3","12/4 = 3","y 6:1 = 6"]
And since the smallest positive quotient is 3, we have that the slack variable that goes out is ess.
C. The pivot element, which now needs to be made 1, is 4.
We obtain the table:

Table IV. End of the process

Base Variable de decisión Variable de holgura Valores solución
x y h s d
y 0 1 -1/2 0 0 12
d 0 0 -7/4 0 1 3
x 1 0 -3/4 0 0 3

Z 0 0 5/4 0 0 33

Since all the coefficients in the objective function row are positive, we have reached the solution.
optimal.
The optimal solution is given by the value of Z in the solution values column, in our
case:33. In the same column, you can observe the vertex where it is reached, observing the rows.
corresponding to the decision variables that have entered the base: D(3,12)
If in the maximization problem there were inequalities of the form: ax + by as constraints c;
multiplying them by -1 transforms them into inequalities of the form -ax - by and we are in the
previous case
If instead of maximizing it is a minimizing problem, the same process is followed, but
changing the sense of the criterion, that is, to enter the base the variable whose value is chosen,

Ing. Efraín Murillo Msc.

Operations Research I 80

row of the objective function, let it be the largest of the positives and the iterations are completed when all the
the coefficients of the objective function row are negative

Geometric interpretation of the simplex method

The successive tables we have built provide the value of the objective function in the
different vertices, adjusting at the same time the coefficients of the initial and slack variables.

In the first iteration(Table I)all the coefficients have remained the same, the has been calculated the
value of the objective function at vertex A(0,0), which is 0.

Next, it moves along edge AB, calculating the value of f, until it reaches B.
This step contributes to theTable II.

In this second iteration, the value corresponding to vertex B(8,0) has been calculated: Z=f(8,0) = 24
Continue along the edge BC until you reach C, where you stop and deploy the data from theTable III.
In this third iteration, the value corresponding to vertex C(6,6) has been calculated: Z=f(6,6)=30.
Continue calculating through edge CD until reaching vertex D. The data that is reflected
they are from theTable IV.

Concludes with this table, noting that it has finished (before verifying that the solution does not
improvement when moving along the edge
The maximum value of the objective function is 33, and it corresponds to x = 3 and y = 12 (vertex D).
If we calculate the value of the objective function at the vertex E(0,14), its value does not exceed 33.

Ing. Efraín Murillo Msc.

Operations Research I 81

CASE STUDY: LINEAR PROGRAMMING

You, an Industrial Engineering student, have a father named Lucas, a businessman, who is dedicated to
business producing 2 types of wooden furniture for children (desks and drawing boards)
for which it has 2 types of workers: carpentry and finishing. Currently, it has 5
woodworking operators and 14 finishing operators, each works 6 effective hours per day.
A cost analysis estimates that the profit per desk is 30 dollars and for each table 25.
dollars.
A desk requires 2 hours of carpentry and 2 hours of finishing, while a tabletop
drawing requires 1 hour of carpentry and 4 hours of finishing.
A market study indicates that there is a maximum demand of 10 desks, but the demand for
drawing boards have no limit.
Requirement
Finishing Carpentry Utility/unit
Desk 2 hrs. 2 hrs. $ 30
Drawing board 1 hour. 4 hrs. $ 25
Daily availability 30 hrs. 84 hrs.
Currently, the company produces 10 desks and 10 drawing boards.
You have just received the Operations Research course, therefore you dare to
suggest some optimization alternatives in the company. What alternatives would you recommend?
You, regarding your current Production Plan assuming that you are in the company
only with your pocket calculator?
b) What is the new utility of the company?
c) Lucas observed that his finishing workers had too much free time, so he
Request an analysis of the labor force, what is your opinion on this?
d) Is there unmet demand? What is it?
Lucas, faced with the surprising results, decides to implement the new production plan. After a while
Some weeks he observes situations where decision making is complicated for him, so
I turn to you to hire your services in order to analyze the following concerns:
e) The operators of both carpentry and finishing frequently request permits, for which
Lucas believes that production will decrease and consequently his profit as well.
How much would I stop producing and earning daily for each type of permit hour?
granted?
f) On the other hand, there are workers both in carpentry and finishing who offer to work.
overtime. Lucas wants to know what effect hiring an hour has on his current solution.
extra carpentry and an extra hour of finishing. How much would you be willing to pay as
máximo por una hora adicional de carpintería?, ¿Cuánto por una hora adicional de acabado?
g) Knowing that the extra wage of a carpenter worker is 2.5 dollars per hour and the salary
extra for a finishing worker 3 dollars, what would be the net profit per additional hour
for each type?, What would you recommend?, hiring extra hours for carpentry or finishing.
h) Lucas realizes now that by increasing overtime, his profit increases and if he grants
permissions, decreases its utility by a constant amount, so he would like to know until
how many hours of carpentry and finishing can and should be increased and up to how many hours of
Carpentry and finishing must grant permits. For this, you researched that the value in the
what increases or decreases the utility when increasing or decreasing one unit on the right side, it
remains constant as long as the optimal solution (optimal point) remains in the
intersection of the lines that generated it.

Ing. Efraín Murillo Msc.

Operations Research I 82

i) Concerned about the selling prices, which have recently fluctuated and affect
directly the utility per article produced, Lucas wants to find out what happens with his
optimal solution to the variation in the unit profit of each product? Within what
variation ranges for each of these coefficients, can Lucas be at ease or not
change the current production plan?
There is a new desktop model for children, which has a high demand. This model uses
1.5 hours of carpentry and 3 hours of finishing generate a profit of $35.
j) Lucas asks him to assess the feasibility of producing the new product or not. If the answer is
Yes, what should be the new production plan and the new utility for the company?
k) Lucas congratulates his son on the advice, but he would like to know why the production stopped.
old desks? And what should happen for it to be included back in the plan?
production?
In a market study, it was determined that the new desktop model has a demand
maximum daily of 10 units.
Lucas wants to know how it affects his current production plan and what his new profit would be.
maximum.
Given your constant concern for the business, since you intend to inherit it, he invented a
device to be used by carpentry workers, which allows reducing the time in the
production of each item by 10%.
m) What would be the new optimal solution to the problem?
n) Lucas realizes that the values are now in decimals and finds it difficult to carry it to the
practice, How could you help Lucas with his new concern?
o) Assuming that the company works 5 days a week and that its device has a cost of
$700, would it be worth implementing it?

Solution:
You have just received the Operations Research course, therefore you dare to suggest
some optimization alternatives in the company. What alternatives would you recommend?
regarding your current Production Plan assuming that you are only in the company
with your pocket calculator?

Maximize:
Z = 30X1 + 25X2
S.A.
2X1 + X2 <= 30 hours of carpentry
2X1+4X2<=84 finishing hours
X1 <= 10
X1, X2>=0

Applying the graphical method:

Ing. Efraín Murillo Msc.

Operations Research I 83

Regarding the production plan, I would recommend that they produce 6 desks and 18 boards to maximize.
its usefulness with the restrictions it has

What is the new utility of the company?

The new profit of the company would be $630.00

Lucas noticed that his finishing workers had too much free time, so he asks them
an analysis of the workforce, what is your opinion on it?

By currently producing 10 desks and 10 boards, the finishing hour requirement would be:
2X1 + 4X2 <= 84 finishing hours
2(10)+4(10)=60 hours, leaving 24 hours free.
But with the new production plan, it would be:
2(6)+4(18)=84 hours, eliminating idle time.

Is there unsatisfied demand? What is it?

With the new production plan, if there is unmet demand, 4 out of the 10 desks are not manufactured.
what the market demands.

Lucas, faced with the surprising results, decides to implement the new production plan. After a few.
Some weeks he observes situations in which decision-making is complicated for him, so he resorts to
You. To hire your services in order to analyze the following concerns:

Both the carpentry and finishing workers frequently request permits, so

Lucas believes that production will decrease and consequently his profit will too. How much
Would you stop producing and earning daily for each type of permit granted?

WinQsb Software Output:

Ing. Efraín Murillo Msc.

Investigación de Operaciones I 84

The utility decreases by $11.67 for a missed hour of carpentry and $3.33 for an hour of
unfinished finishes

On the other hand, there are workers both in carpentry and finishing who offer to work.
overtime. Lucas wants to know what effect it has on his current solution to hire an extra hour
for carpentry and an extra hour of finishing. How much would you be willing to pay at most?
For an extra hour of carpentry?, How much for an extra hour of finishing?

You must pay a maximum of $11.67 per hour for carpentry and $3.33 per hour for finishing.

Knowing that the salary of a carpenter worker is 2.5 dollars per hour and the salary of a
finish worker 3 dollars, what would be the net gain per hour for each type?
What would you recommend, hiring extra hours for carpentry or for finishing?

The net profit per hour of carpentry would be: (11.67-2.5) =$9.17
The net profit per finishing hour would be: (3.33-3.00)=$0.33
Recomendaríacontratar horas extras de carpinteríaya que nos da mayor utilidad.

Lucas realizes now that by increasing overtime, his profit increases and if he grants
permissions, decreases its usefulness by a constant amount, so you would like to know until
how many hours of carpentry and finishing can and should increase and up to how many hours of
Carpentry and finishing must grant permissions. For this, you investigated the value in which
increases or decreases the utility by increasing or decreasing one unit on the right side, it
remains constant as long as the optimal solution (optimal point) remains at the intersection
of the lines that generated it.
In Carpentry, it can increase up to 36 (this is 6 hours) and decrease up to 21 (this is in 9 hours).
In Finishing, it can increase up to 120 (this is within 36 hours) and decrease down to 60 (this is within 24 hours).

Worried about the selling prices, which have recently varied and affect
directly the utility per produced item, Lucas wants to find out what happens with his
optimal solution in the face of the variation of the unit utility of each product? Within what
variation ranges for each of these coefficients can Lucas be calm and not
change the current production plan?

It must be within the following ranges:

For the desks: from 12.5 to 50
For the boards: from 15 to 60

Ing. Efraín Murillo Msc.

Operations Research I 85

There is a new desktop model for children, which has a high demand. This model uses 1.5
hours of carpentry and 3 hours of finishing and generates a profit of $35.

Lucas asks you to evaluate the feasibility of producing the new product or not. If it is the
Yes answer, what should be the new production plan and the new utility for the
company?

A new variable must be added to the model and after solving, the following result is obtained (according to
the software WinQsb):

It is observed that total profit increases, therefore the new model should indeed be produced.
desk, and stop producing the old model. The new production plan is to produce 12 desks
of the new model and 12 boards.

Lucas congratulates his son on the consultancy, but he would like to know why the production stopped.
old desks? And what should happen for it to be included again in the plan?
production?

It should improve its unit utility by at least $15.00 (Reduced Cost).

In a market study, it was determined that the new desktop model has a daily demand.
maximum of 10 units.

Lucas wants to know how it affects his current production plan and what his new utility would be.
maximum.

We need to add a new constraint to the model. The new solution is:

Ing. Efraín Murillo Msc.

Operations Research I 86

El nuevo plan de producción sería: 1 escritorio, 13 tableros y 10 escritorios para niños. La nueva
utilidad sería: $705.00

Given your constant concern for the business, since you plan to inherit it, you invented a
device to be used by woodworking operators, which allows for reducing time in the
production of each item by 10%.

What would be the new optimal solution to the problem?

The new mathematical model would be:

Desks
x2: Drawing boards
x3: Desks for children
Max 30x1 + 25x2 + 35x3
ST
1.8x1 + 0.9x2 + 1.35x3 <= 30 hours of carpentry
2x1 + 4x2 + 3x3<=84
x1 <=10
x3 <= 10
WinQsb output:

3.22 desks, 11.89 boards, and 10 children's desks must be made, yielding a profit of
$743.89.

Lucas realizes that the values are now with decimals and finds it difficult to carry it to the
practice, how could you help Lucas with his new concern?

In WinQsb, we indicate that the type of variables must be non-negative integers, obtaining the
next solution:

Ing. Efraín Murillo Msc.

Operations Research I 87

This solution is whole, harming the total utility.

Assuming that the company works 5 days a week and that its device costs $
700, would it be worth implementing it?

The profit without the device is $705 per day, and with the device, it is $740 per day, then the
total increase per week will be:
5*(740-705) = $135
Per month it will be:

4*135 = $540
In two months it will be:
2*540 = $1080 > $700, therefore in my opinion it would be worth implementing it.

Eng. Efraín Murillo Msc.

Operations Research I 88

BANK OF PROBLEMS:
1. FORMULATION OF PROBLEMS
2. SOLUTION BY THE GRAPHIC METHOD
3. SOLUTION BY THE SIMPLEX METHOD
4. SENSITIVITY ANALYSIS
5. DUAL ANALYSIS

1. Given the following mathematical model of linear programming:

Min y1 + 2y2
St
y1 + 4y2 <= 21
2y1 + y2 >= 7
3y1 + 1.5y2 <= 21
-2y1 + 6y2 >= 0
y1, y2 >= 0
a) Obtain the optimal solution by the graphical method (value of the decision variables and value of
the objective function). Show the respective graph.
b) Show the respective Dual mathematical model.

[Link] the following mathematical model of linear programming:

Min Z = 2X1 + 3X2
St
X1 + X2 <= 8
X1 - 3X2 <= 2
2X1 + X2 >= 11
X1, X2 >= 0
It is requested: (From your answers with only 2 decimals)

a) The optimal solution by the graphic method. Show the respective graph.
b) The dual price of the exhausted RHS (right-hand sides).
c) The sensitivity analysis of RHS1 and RHS2.
d) The sensitivity analysis of the coefficients of the objective function.
e) How would an increase of 3 units in RHS2 affect the objective function?
f) The solution using the generalized simplex method. Show only the initial table and the
first iteration.

3.-The company XXX produces 2 items: car covers and awnings. For the next week, it has available
the following supplies:
400 man-hours
1200 meters2of linen
2700 meters2of cloth
600 machine hours
The utility of car covers is 150 soles and of awnings is 500 soles.

Ing. Efraín Murillo Msc.

Operations Research I 89

To produce a car cover requires 1 man-hour, 6 meters 2of linen and 1 hour-machine, and
To produce a tent, it requires 2 man-hours, 18 meters. 2of canvas and 1.5 hours-machine. There is
a minimum demand of 140 awnings.
To obtain your answers, work with all data rounded to 3 decimal places.
1) Define the decision variables of the problem.
Xi: Number of units to be produced weekly of product i
i=1(cubiertas de carro), 2(tolderas)
2) Formulate the objective function of the problem.
Max 150x1 + 500x2
3) Formulate the structural constraints of the system.
x1 + 2x2 <= 400 man-hours
6x1 <= 1200 meters2of linen
18x2 <= 2700 meters2of canvas
x1 + 1.5x2 <= 600 machine hours
X2 >= 140
Xi >= 0,i=1,2
4) Obtain the solution by the graphical method.

5) Determine the production plan that maximizes the company's profit. What is the
maximum utility?
Production plan:
X1 = 100 car covers
X2 = 150 towels
Maximum income: S/. 90,000
To answer the following questions, you can use the output from Figure 1 of the software.
Nice the output of figure 2 from the WinQsb software shown below.
6) Determine the total amount used in meters.2of linen.
6*100=600 meters2of linen
Are there idle machine hours? How many?
Yes, 275 hrs. machine
8) How many man-hours were used in the production of the awnings?

Eng. Efraín Murillo Msc.

Operations Research I 90

2*150=300 man hours

How many machine hours were used in the production of the car covers?
1*100=100 machine hours
10) How much would you be willing to pay at most for 1 meter?2of additional canvas?
$11,111
11) How much would you pay at most for an extra machine hour?
$0
12) What range of variability is the dual price of a man-hour valid within?
[300,500]
13) Determine the range of variability of the unit utility of a awning, within which the
Current production plan does not change.
[300, infinity]
14) Obtain the solution of the previous mathematical model using the SIMPLEX algorithm and
compare the answers to questions 5), 10) and 11).
FINAL TABLE OF THE SIMPLEX

The answers are the same.

15) Build the corresponding dual mathematical model.

Min 400y1 + 1200y2 + 2700y3 + 600y4 - 140y5
St
y1 + 6y2 + Y4 >= 150
2y1 + 18y3 + 1.5y4 - Y5 >= 500
Yi>=0,i=1,2,3,4,5
16) Define the variables of the dual problem.
Yi: Price per unit of the RHSi resource
i =1:hrs. hombre, 2: mts2 de lino, 3: mts2 de lona, 4: hrs. máquina, 5: demanda min. de x2
17) Show the respective optimal solution of the dual model (value of the objective function and of the
decision variables)
Z= S/. 90,000
y1=S/. 150

Ing. Efraín Murillo Msc.

Operations Research I 91

y3=S/. 11.111

Entry of the model to Lindo:

Figure 1: Output of the LINDO Software

Entry of the model into WinQsb:

Figure 2: Output of the WinQsb Software

Ing. Efraín Murillo Msc.

Operations Research I 92

4.-With the following data:

Product Resource
Resource
A B available
R1 10 10 200
R2 8 4 64
R3 3 0 72
Contribution/unit $5 $5

Find the optimal solution using the simplex method. Show the iterations.

5.-Company ABC S.A. has a budget of up to $1500 to spend on local advertising.

next week. The global objective of the company is to reach the maximum audience of potential
clients. Likewise, they want to reach at least 9000 children and 1000 elderly people.
There are 3 advertising media available; their costs and the audience they have are given in the table.
next
Newspaper Radio TV
Cost per ad ($) 200 100 500
Total Audience
(persons/ad) 31000 15000 50000
Children/announcement 1000 1000 3000
Grandparents/announcement 200 100 500
Sea: X1, X2 y X3: número de anuncios en Periódico, Radio y TV por semana, respectivamente.
a) Formulate the mathematical model of linear programming. (state the restrictions in the
next order: budget spending, restriction on children and restriction on grandparents
Using the output of the WinQSB Software for your solution, please address the following concerns:
b) What is the optimal mix of advertising? And how much audience would be reached?
c) How many grandparents would be reached with the newspaper.

d) How much would the audience increase with an additional dollar in the budget?
e) How much would the total audience of television need to increase for it to become profitable?
your consideration in the optimal advertising mix?
f) Show the respective Dual mathematical model.
g) Show the optimal solution of the dual model (value of the decision variables and the value of the
objective function

Ing. Efraín Murillo Msc.

Operations Research I 93

h) Interpret the variables y1 of the dual problem.

SOLUTION:
a) Mathematical Model:
Max 31000X1 + 15000X2 + 50000X3
St
200X1 + 100X2 + 500X3 <= 1500 budget
1000X1 + 1000X2 + 3000X3 >= 9000 children
200X1 + 100X2 + 500X3 >= 1000 grandparents
X1, x2, x3 >= 0

WinQsb Software Output:

b) What is the optimal mix of advertising? And what audience would be reached?
6 ads in the newspaper and 3 on the radio.
Total audience = 231,000 listeners.
c) How many grandparents would be reached with the newspaper.

200(6) = 1200 grandparents

d) How much would the audience increase with an additional dollar in the budget?
In 160 audiences.
e) How much would the total audience of television have to increase for it to become profitable?
Your consideration in the optimal mix of advertising?
In 27,000 audiences.
f) Show the corresponding Dual mathematical model.
Min 1500Y1 + 9000Y2 + 1000Y3
St
200Y1 - 1000Y2 - 200Y3 >= 31000
100Y1 - 1000Y2 - 100Y3 >= 15000
500Y1 - 300Y2 - 500Y3 >= 50000
Y1, Y2, Y3 >= 0
g) Show the optimal solution of the dual model (value of the decision variables and the value of the
objective function
Y1= 160
Y2= 1
Y3= 0
W=231000
h) Interpret the variables y1 of the dual problem.

Eng. Efraín Murillo Msc.

Operations Research I 94

Y1 = value in number of listeners per dollar of budget.

6.-For the following mathematical model, which represents a cost minimization problem:
MIN 4X1 + 5X2
St
5X1 + 2X2 >= 120
-X1 + 3X2 <= 230
X1<=30
X2>=30
X1, X2 >= 0
Sepide:
a) The optimal solution using the graphic method.
b) The state of the RHS
c) The Dual Price of exhausted RHS.
SOLUTION
a) Respective graph.

Point X1 X2 Z
Po 12 30 198 Min
P1 0 60 300
P2 0 76.7 383.5
P3 30 86.7 553.5
P4 30 30 270

Optimal solution:
X1=12
X2=30
Z=198

b) Status of the RHS

RHS Available Used Slack or Surplus

1 120 120 0 exhausted

Ing. Efraín Murillo Msc.

Operations Research I 95

2 230 78 152 slack

3 30 12 18 slack
4 30 30 0 sold out

c) Dual price of the exhausted RHS:

PD of RHS1:
5X1 + 2X2 = 121
X2=30
Resolving:
X1= 12.2
X2=30
Z’= 198.8
PD(RHS1)=198–198.8 = -0.8

PD of RHS4:
5X1+2X2=120
X2=31
Resolving:
X1= 11.6
X2=31
Z’= 201.4
PD(RHS4)=98 -201.4 = -3.4

Exit of WinQsb:

d) Sensitivity Analysis
Sensitivity Analysis of RHS1:
Max P4(30,30): 5(30)+2(30)=210
Min P7(0,30): 5(0)+2(30)=60
60 <= RHS1 = 120 <= 210

Sensitivity Analysis of RHS2:

Max M
Min P0(12,30): -(12)+3(30)=78
78 <= RHS2 = 230 <= M

Ing. Efraín Murillo Msc.

Operations Research I 96

Sensitivity Analysis of RHS3:

Max M
Min P0(12,30): 12
12 <= RHS3 = 30 <= M

Sensitivity Analysis of RHS4:

Max P1(0,60): 60
Min P5(24,0): 0
0<=RHS4=30<=60

Sensitivity Analysis of C1, then C2=5

P(z)=4/5=0.8
P(R1)=5/2=2.5
P(R4)=0/1=0
P(R4) <= P(z) <= P(R1)
0<=C1/5<=2.5
0<=C1<=12.5

Sensitivity Analysis of C2: then C1=4

0 <= 4/C2 <= 2.5
1.6 <= C1 <= M

7.-A company produces three types of products: A, B, and C. The required man-hours are H1 and
H2 in departments 1 and 2 respectively for product A. One unit of B requires H3.
and H1 hours in departments 2 and 3 respectively, and a product C requires H3, H2, and H1
hours in departments 1, 2, and 3 respectively. The hourly wage is "J" in any
department. It is desired to manufacture up to a maximum of FT units for the three products; a
maximum of F1 units between products A and B and a minimum of F2 units of C. If the products
A, B y C se vende a V1, V2 y V3 cada unidad respectivamente y la disponibilidad de horas son HA,
HB and HC for departments 1, 2, and 3 respectively. Formulate a Programming Model.
Linear for this problem.
SOLUTION:

Hour requirement
A B C Availability
DEPARTMENT 1 H1 H3 HA
DEPARTMENT 2 H2 H3 H2 HB
DEPT 3 H1 H1 HC

Variables:
Xi: Number of units to be produced of product i
i=A, B, C

Ing. Efraín Murillo Msc.

Operations Research I 97
Max V1XA + V2XB + V3XC - J(H1XA + H3XC + H2XA + H3XB + H2XC + H1XB + H1XC)
St
XA + XB + XC <= FT maximum production of the three products
XA + XB <= F1 maximum production between products A and B
XC>=F2 minimum production of product C
H1XA+H3XC<=HA Available hours in the DEPT 1
H2XA + H3XB + H2XC <= HB Available hours in the DEPT 2
H1XB + H1XC <= HC Available hours in the DEPT 3
Xi >= 0

8.-Amanufacturing company produces three models weekly (Green, Blue, and Pink) of a certain
product. For this, it uses two types of materials (Onyx and Plaqué) of which 4000 and 6000 are available.
units, respectively. The material requirements per unit for the 3 models are given to
continuation:
Requirement per unit of the model
Materials Green Blue Pink
Onyx 2 3 5
Plated 4 2 7
The labor time required for each unit of the Green model is one man-hour, which is the
twice what each unit of the Blue model requires and 10/3 times what the Pink model requires.
The factory workforce has 700 man-hours available. The market study indicates that the
the minimum demand for the three models is 200, 200, and 150 units, respectively.
The unit profits for the Green, Blue, and Pink models are 30, 20, and 50 soles.
Output of the WinQsb software:

Using the output from the WinQsb software, address the following concerns:
a) Define the decision variables of the problem.
b) Show the respective linear programming mathematical model. (consider the following
order for the restrictions: Onix, Plaqué, Man-Hours, green demand, demand for
blue and demand for pink.
c) Determine the maximum profit of the company. How many units of each model are produced?
d) How much Onix is used in the production of the Blue model?
e) What amount of Plaqué is used in the production of the Rosado model?
f) How much should he pay at most to increase the Onix stock by 1 unit?

Eng. Efraín Murillo MSc.

Operations Research I 98

g) How much should be invested at most to hire 200 hours of overtime work?
h) Show the corresponding Dual mathematical model.
i) Show the optimal solution of the dual model (value of the decision variables and the value of the
objective function
j) Interpret the variables Y1, Y2, and Y3 of the dual problem.

SOLUTION:
a) Define the decision variables of the problem.
Xi: Number of units to be produced weekly of product i
i=1,2,3
b) Show the corresponding mathematical model of Linear Programming. (consider the following
order for restrictions: Onix, Plaqué, Man-Hours, green demand, demand of
blue and demand for pink).
Max30X1+20X2+50X3

ST

2X1 + 3X2 + 5X3 ≤ 4000

4X1 + 2X2 + 7X3 ≤ 6000
X1 + 0.5X2 + 0.3X3 ≤ 700
X1≥200
X2≥200
X3≥150
Xi≥0

c) Determine the maximum profit of the company. What quantity is produced of each model?
Maximum Profit: S/. 42500
Production:
X1= 450 units of the Green model
X2=200 units of the Blue model
X3=500 units of the Pink model
d) How much Onix is used in the production of the Blue model?
3*200= 600 units
e) What amount of Plaqué is used in the production of the Rosado model?
7*500=3500 units
f) How much should he pay at most to increase Onix's inventory by 1 unit?
9.3182 soles
g) How much should be invested at most to contract 200 extra work hours?
11.3636*200=2272.72 soles
h) Show the respective Dual mathematical model.
Min 4000Y1 + 6000Y2 + 700Y3 - 200Y4 - 200Y5 - 150Y3
ST
2Y1 + 4Y2 + Y3 - Y4 ≥ 30
3Y1+2Y2+0.5Y3-Y5≥20
5Y1+7Y2+0.3Y3-Y6≥50

Ing. Efraín Murillo Msc.

Operations Research I 99

Yi≥0
i=1,2,3,4,5,6
i) Show the optimal solution of the dual model (value of the decision variables and the value of the
objective function
Y1=9.3182
Y2=0
Y3=11.3636
Y4=0
Y5=13.6364
Y6=0
W=42500
j) Interpret the variables Y1, Y2, and Y3 of the dual problem.
Y1: precio por unidad de Onix
Y1: price per unit of Plaqué
price per man-hour

A furniture maker has two different types of wood; he has 1500 board feet of type A and
1000 of type B, also has 800 man-hours available to carry out the work. The demand
What has been estimated is the following: at least 40 tables, 130 chairs, 30 desks and no more.
of 10 bookshelves. The quantities of wood A and B, and the man-hours required for the production
For each unit of article, they are indicated in the following table:

Wood Hours Demand Utilities

Article A B Man Dear per unit

Table 5 2 3 no less than 40 $ 12
Chair 1 3 2 no less than 130 5
Escritorio 9 4 5 no less than 30 15
Bookshelf 12 1 10 no more than 10 10
Availability
weekly 1500 1000 800
Formulate a Linear Programming Mathematical Model to determine how much should be produced.
furniture maker of each item, so that the profits are maximized?

Solution
Decision variables:
X1: Número de mesas a producir semanalmente
X2: Number of chairs to be produced weekly
X3: Number of Desks to be produced weekly
X4: Número de Libreros a producir semanalmente

Mathematical model:
Max 12X1 + 5X2 + 15X3 + 10X4
ST
5X1 + X2 + 9X3 + 12X4 <= 1500 feet wooden table A
2X1 + 3X2 + 4X3 + X4 <= 1000 feet wooden table B
3X1 + 2X2 + 5X3 + 10X4 <= 800 man-hours
X1>=40
X2>=130
X3>=30

Ing. Efraín Murillo Msc.

Operations Research I 100

X4<=10
X1, X2, X3, X4 >= 0
10.-Dado el siguiente modelo matemático de programación lineal:
MIN Z = 2X1 + 3X2
ST
X1 - 3X2 <= 2
2X1 + X2 >= 11
X1 + X2 <= 8
X1, X2 >= 0
a) Show the respective graph and the evaluation table of the intersection points of the region
feasible.
Respectivegraph

Table:

Point x1 x2 Z
P1 3 5 21
P2 6,5 1,5 17.5
P3 5 1 13 Min

b) Indicate the value of the decision variables and the value of the objective function.
X1=5

X2=1

Z=13

c) Determine the Dual Price of RHS2

Optimal point P3 (R1∩R2)

R1:X1-3x2=2

R2:2x1+x2=11+1

2*R1-R2: -7x2=-8 X2=1.14

In R1: X1-3(1.14)=2 X1= 5.42

Ing. Efraín Murillo Msc.

Operations Research I 101

Z' = 2(5.42) + 3(1.14) = 14.26

PD(RHS2)= 13-14.26=-1.26

11.-An industrial company produces 5 models of Keyboards, each of which is handled in the
assembly and finishing departments. The number of man-hours of labor required,
by keyboard in each department is:
["Model1","Model2","Model3","Model4","Model5"]
Assembly 2 5 7 9 7
Finished 3 4 1 6 5
There are 10,000 man-hours available in the assembly department and 5,000 in the other one.
finished next week. The profits in dollars for each keyboard model are: 6, 3,
5, 4 and 5 respectively for keyboards 1, 2, 3, 4 and 5.
There are restrictions to produce at least 100 units of model 3, at most 150 units of
model 4 and at most 250 units of model 5. The production of models 1 and 2 is unrestricted.
a) Define the decision variables of the problem.
Xi: Number of units to be produced weekly of model i
i=1,2,3,4,5
b) Show the respective mathematical model.
Max 6X1 + 3X2 + 5X3 + 4X4 + 5X5
ST
2X1 + 5X2 + 7X3 + 9X4 + 7X5 <= 10000
3X1 + 4X2 + 1X3 + 6X4 + 5X5 <= 5000
X3 >= 100
X4<=150
X5 <= 250
Xi >= 0
c) Show the Dual model.
Primal Model:
Max 6X1 + 3X2 + 5X3 + 4X4 + 5X5
ST
2X1 + 5X2 + 7X3 + 9X4 + 7X5 <= 10000
3X1 + 4X2 + 1X3 + 6X4 + 5X5 <= 5000
-X3 <=-100
X4 <=150
X5<=250
Xi>=0
Dual Model:
Min 10000Y1 + 5000Y2 - 100Y3 + 150Y4 + 250Y5
ST
2Y1 +3Y2 greater than or equal to 6
5Y1 +4Y2 greater than or equal to 3
7Y1 +1Y2 -Y3 greater than or equal to 5
9Y1 +6Y2 +Y4 >=4
7Y1 +5Y2 +Y5 >= 5
Yi>=0

d) Define the variables of the Dual model.

Hourly assembly price
Y2: Price per hour of Finishing

Ing. Efraín Murillo Msc.

Operations Research I 102

Y3: Price per unit of demand for model 3

Y4: Price per unit of demand of model 4
Y5: Price per unit of demand for model 5
12.-For the following mathematical model, which represents a production problem:
MAX Z = 3X1 + 2X2 utility in soles
ST
X1 + X2 <= 80 man-hours
2X1 + X2 <= 100 machine hours
X1 >=40 demand for product 1
X1, X2 >= 0
It is requested:

a) The optimal solution using the graphical method. Show the respective graph, the value of the
decision variables and objective function value.
b) The optimal solution using the simplex method. Show the iteration and indicate it.
value of the decision variables and the value of the objective function.
SOLUTION:
a) Graphical Solution:

a) The optimal solution is:

X1=40 units
X2=20 units
Z= S/. 160

13.-A company is considering seven major capital investments. Each investment can be
do it only once. These investments differ in the estimated long-term gain (present value
net) that will be generated, as well as the amount of capital required, as shown in the following table
(in millions of dollars):
Investment opportunity
1 2 3 4 5 6 7
Estimated profit 17 10 15 19 7 13 9
Required capital 43 28 34 48 17 32 23
There is $100 million available as total capital for these investments. The opportunities
Investment 1 and 2 are mutually exclusive, as are 3 and 4. Furthermore, opportunity 3 does not
It can only be taken advantage of if one of the first two opportunities is invested in. The goal is
choose the combination of capital investments that maximizes the estimated long-term profit
net present value
Formulate a binary programming model for this problem.
Solution
Xi=1 if project i is accepted, =0 otherwise

Ing. Efraín Murillo Msc.

Operations Research I 103

i=1,2,3,4,5,6,7
Max 17x1 + 10x2 + 15x3 + 19x4 + 7x5 + 13x6 + 9x7
St
43x1+28x2+34x3+48x4+17x5+32x6+23x7<=100
x1 + x2 <= 1
x3 + x4 <= 1
x3 - x1 <= 0
x3 - x2 <= 0
end
Xi {0,1}
14.-For the following mathematical model, which represents a cost minimization problem:
MIN Z = 4X1 + 3X2
ST
X1 + 3X2 >= 15
2X1 + X2 >= 10
X1 + X2 >= 8
X1, X2 >= 0
It is requested:

a) The respective graph and the optimal solution (value of the decision variables and value of the
objective function

Z=26
X1=2
X2=6
b) The dual price of the RHS.
PD1=0
PD2=-1
PD3=-2

15.-A factory produces 4 products: A, B, C, and D. Each unit of product A requires two hours.
of machining, one hour of assembly and is worth $10 in work-in-progress inventory. Each unit of the product
B requires one hour of machining, three hours of assembly, and has a value of $5 in work in progress.
Each unit of C requires 2 ½ hours of machining, 2 ½ hours of assembly and is worth $2.
inventory in process. Finally, each unit of product D requires five hours of
machining, does not require assembly time and is worth $12 in work in progress inventory.
The factory has 1200 hours of machining and 1600 hours of assembly. Additionally, it cannot
to have more than $10,000 in work-in-progress inventory. Each unit of product A has a
utility of $40; each unit of B, $10; each unit of C, $36, and each unit of D, $23. No
More than 200 units of A can be sold; no more than 160 of C, and any can be sold.

Ing. Efraín Murillo Msc.

Operations Research I 104

quantity of products B and D. However, to fulfill a contract, they must be produced and
to sell at least 100 units of product D. The factory's goal is to maximize profit.
resulting from the sale of the four products.
Answer the following questions using the output from the LINDO Software (See fig. 1)
Fig. 1

j) Show the respective mathematical model and define its respective decision variables.
MAX 40XA + 10XB + 36XC + 23XD
ST
2XA + XB + 2.5XC + 5XD <= 1200
XA + 3XB + 2.5XC <= 1600
10XA + 5XB + 2XC + 12XD <= 10000
XA<=200
XC<=160
XD>=100
Xi >= 0
Decision variable:
Xi: Number of units to be produced of product i
i=A,B,C,D
k) What is the optimal production plan?
Production Plan:
200 units of product A
120 units of product C
100 units of product D
l) What is the utility obtained from the production of product type A?
200*40=$ 8000
m) How many hours of assembly does the production of product type D consume?

Ing. Efraín Murillo Msc.

Operations Research I 105

None
n) How much would you be willing to pay for an extra hour of machining?
Up to 14.4 dollars
o) How many extra hours would you increase?
Up to 1100 hours.

p) How much should the unit profit of product B improve to justify its production?
In an amount greater than 4.4 dollars
What would happen to the optimal production plan if the unit profit of product C
decreases by $15?
The production plan changes
r) Indicate what the idle capital of the company is.
The idle capital is $6560
A new product, product E, is under consideration. It requires 5 hours of machining.
and 3 hours of assembly. The utility per unit is $30. Should any unit be produced of the
product E?
Opportunity cost = 14.4*5 + 0*3 = $70 > $30, therefore it should not be produced.
t) Obtain the dual model, define its decision variables, and show its optimal solution.
Mathematical model:
MIN 1200Y1 + 1600Y2 + 10000Y3 + 200Y4 + 160Y5 - 100Y6
ST
2Y1 + Y2 + 10Y4 + Y5 >= 40
Y1 + 3Y2 + 5Y3 >= 10
2.5Y1 + 2.5Y2 + 2Y3 + Y5 >= 36
5Y1 + 12Y3 - Y6 >= 23
Yi>=0
i=1,2,3,4,5,6
Decision variables:
Y1: Precio por hora de maquinado
Y2: Price per hour of assembly
Y3: Price per unit of work in progress inventory
Y4: Price per unit of demand for product A
Y5: Price per unit of demand for product C
Y6: Price per unit of demand for product D
Optimal solution:
Y1=14.4
Y2=0
Y3=0
Y4=11.2
Y5=0
Y6=49
W=$ 14620
16.-For the following mathematical model, which represents a production problem:

MAX Z= 3X1+2X2+5X3 ganancia en soles

ST
X1 + X2 + X3 <= 430 man-hours

Ing. Efraín Murillo Msc.

Operations Research I 106

3X1 + 2X3 <= 460 machine hours

X1 + 4X2 <= 420 kg of raw material
X1, X2, X3 >= 0
Applying the Simplex method, obtain your solution and answer the following concerns (your
answers to reach them with 2 decimals:
a) The optimal production plan. The maximum profit.

Optimal production plan:

0 units of X1
105 units of X2
230 units of X3
Maximum profit = 1360 soles
b) The dual prices of each RHS
PD(RHS1)=0
PD(RHS2)=2.5
PD(RHS3)=0.5
c) The slack or surplus of each RHS
RHS1 slack = 95 man-hours
Holgura de RHS2=0 horas máquina
RHS3 clearance = 0 Kg of raw material
d) How much would you be willing to pay at most for an extra man-hour?, For a
extra hour-machine?, for an additional kg of raw material?
For an extra man hour = 0 soles
For an extra machine hour = 2.5 soles
For an additional kg of raw material = 0.5 soles
17.-A company produces 4 products (1,2,3,4) in two departments (A,B). Product 1 goes through
A and B at a rate of 1/3 and 1/2 unit/hour respectively, the 2 goes through A, B, A at a rate of 1, 1, 1/2
unit/hour respectively, the 3 passes through B at a rate of 1/2 unit/hour and the 4 passes through A at a rate of
1/3 unit/hour. If the profits of the products are $30, $20, $45, and $15 and the availabilities
The hours in the departments are 300 and 200 hours respectively, determine:
a) The optimal solution
b) Marginal values and reduced costs
c) If you had to produce 5 units of a product that you do not produce, which one would you choose and what would it be?
its new usefulness?
d) The optimal solution of the dual

18.-Atelevision manufacturer produces four models:A, B, C, and [Link] B are black and white models and
C and D are color models. Each television requires a cathode ray tube, and it also requires time to
assembly and testing. Table 1 shows the assembly and testing requirements for each model, as
as the amount of time available for each operation. Additionally, due to a strike, there is a

Ing. Efraín Murillo Msc.

Operations Research I 107

scarcity of kinescopes and the supplier indicates that it wil not be able to provide more than 180 kinescopes
next month and, of these, at most 100 may be in color. The maximum demand for model B is
120 units and at least 10 units of modelAare required.
Table 1 Model Total
A
B C D available
Assembly time (hours) 8 10 12 15 2000
Testing time (hours) 2 2 4 5 1000
Beneficio marginal (dólares) 40 60 70 100
Using the output of the POMQM software shown above, answer the following
concerns in their answer sheet:
g) What is the optimal production program for the television manufacturer? What is the benefit?
maximum?
h) How much would you pay at most for an additional hour of setup time? What range?
Is this value valid?
i) Suppose that 100 additional hours of assembly time can be obtained at a cost of
Four dollars per hour? Should it be done? What would the net increase in benefits be?
j) Suppose a price change is instituted that modifies the marginal benefit of model C,
from 70 to 80 dollars. Would this change the optimal production plan?
k) How much would the optimal marginal benefit increase if 5 hours of testing are added?
l) Suppose that the price of model C changes from 70 to 90 dollars. In this case, would the plan change?
of production?
m) How many hours of testing are used in the production of model B?
What would happen to the optimal marginal benefit if it is required to produce at least 11 units of the
model A?
o) How many hours of assembly are used in the production of model A?
p) If I had to choose between increasing the assembly hours or color cinescopes, what
Would you decide? Why?
q) How much would the objective function increase if 20 hours of assembly are added?
r) How much would the marginal benefit of model C have to change to justify its
production?
In the previous question, what would be the new optimal benefit of the problem?
t) Obtain the dual model (write it on your answer sheet).
u) ¿Cuál sería la solución óptima del modelo dual? (indique sólo el valor de las variables y el valor
from Z).
v) What does each variable of the dual model represent?
In the primal model, suppose it is required to produce at least 5 units of model C, what
Would it be the new optimal marginal benefit?
x) Without taking the previous question into account, management wants to introduce a new color model.
(model E). Model E would only require 10 hours of assembly and 3 hours of testing. The benefit
The margin of the model would be 70 dollars. Should the new model be produced? Why?
What would be the new production plan if the E model is considered for production?
z) Taking into account the initial primal model, what would happen to the optimal marginal benefit if the
Model A only requires 4 hours of assembly?
19.-“JC” S.A. wants to spend $1500 on local advertising. The overall objective is to achieve the maximum
potential audience while reaching at least 9000 children and 1000 grandparents.
There are 3 media; their costs and the audience they have are given in the following table:

Ing. Efraín Murillo Msc.

Operations Research I 108

NEWSPAPER RADIO T.V.

COSTO POR ANUNCIO ($) 200 100 500
TOTAL AUDIENCE 31000 15000 50000
CHILDREN 1000 1000 3000
GRANDPARENTS 200 100 500

Max 31000P + 15000R + 50000T

St
200P+100R+500T<=1500
1000P + 1000R + 3000T >= 9000
200P+100R+500T>=1000
PomQm output:

Using the output of the POMQM software shown above, answer the following.
concerns on their answer sheet:
h) What is the optimal mix of advertising? And how much audience would be reached?
i) How much would the total audience increase with an additional dollar in the budget? Why? Which
What is the range for this marginal value to be valid?
j) "JC" is concerned, as the total audience of the Radio is only approximate. About what range
Will the Radio remain in the solution for total audience?
k) How much would the total audience increase if 500 dollars were added for advertising?
l) If the total audience of the Newspaper decreased by 1000, what would be the new optimal solution for 'JC'?
Why?
m) Suppose that the audience reached by the radio increases by 700. Would this change the solution?
optimal?
n) What total audience is reached with the radio?
o) What would happen to the total audience if it is required to reach at least 9001 children? Why.
p) How many grandparents are reached in total?
q) How many grandparents can you reach just with the newspaper?
r) How much would the total audience increase if 2000 dollars of capital are added for
advertising?
s) By how much would the total audience of television have to increase for it to become profitable?
consideration in the optimal mix of advertising? Why?
In the previous question, what would be the new total audience reached?
u) Obtain the dual model (write it on your answer sheet).
v) ¿Cuál sería la solución óptima del modelo dual? (indique sólo el valor de las variables y el valor
of Z).

Eng. Efraín Murillo Msc.

Operations Research I 109

w) What does each variable of the dual model represent?

In the primal model, suppose it is required to produce at least 5 units of model C, what
Would it be the new optimal marginal benefit?
y) Not taking into account the previous question, a quarter medium is under consideration. It requires 350.
budget dollars, reaches 2000 children and 400 grandparents. The total audience is 40000.
Should it be considered? Why?
What would the new value of z be if the fourth medium is included?
aa) Taking into account the initial primal model, what would happen to the optimal total audience if the cost
The TV ad is for $300 instead of $500?
20.- A manufacturing company produces three models weekly (Green, Blue, and Pink) of a
certain product. For this, it uses two types of materials (Onyx and Plaqué) which are available
4000 and 6000 units, respectively. The material requirements per unit of the 3
models are given below:
Requirement per unit of the model
Materials Green Blue Pink
Onix 2 3 5
Plaqué 4 2 7
The labor time required by each unit of the Green model is one man-hour which is the
double what each unit of the Blue model requires and 10/3 times what the Pink model requires. The
The factory's workforce has 700 man hours available. The market study indicates that the
the minimum demand for the three models is 200, 200, and 150 units, respectively.
Las utilidades por unidad de los modelos Verde, Azul y Rosado son 30, 20 y 50 dólares.
a) Show the corresponding linear programming mathematical model.
b) Determine the maximum profit of the company. What quantity is produced of each model?
c) How much Onix is used in the production of the Blue model? In what range is this valid?
value?.
The company wants to increase its profits to $44,000 by acquiring extra time from its workforce.
work how much more extra time is needed?
e) What quantity of Plaqué is used in the production of the Rosado model?
f) What would be the new utility of the company if 500 units of Onix are added? Explain.
g) If you could obtain additional amounts of only one of the three resources
Which one would you recommend? Explain.
h) What would happen to the production plan if the profit from the green model decreases by 5%?
i) How much should be paid at most to increase the Onix stock by 1 unit?
Explain.
j) How much should you invest at most to hire 200 extra work hours? Explain.
k) If the minimum demand for the blue model increases by 2%, what would happen to the profit of the
company?
l) What gain coefficient or coefficients could be doubled while keeping fixed
all the other coefficients, without affecting the optimal production plan? Explain.
m) If a new product is considered that requires 3 units of Onix, 2 units of Plaqué and
2 man-hours and the utility per unit is $50, should it be produced?
n) Indicate and interpret the clearances and excesses of the right sides.
o) Obtain the DUAL LP model of the primal (initial) problem.
p) Define the decision variables of the dual model.
q) Show your optimal solution of the dual model.
r) Show the dual values of the dual model.

Ing. Efraín Murillo Msc.

Operations Research I 110

21.-Cauchos Peru manufactures three rubber products: Airtex (foamy material), Extendex (material
elastic) and Resistex (rigid material). The three products require the same three chemical polymers.
and a base. The amount of each ingredient used per pound of the final product is shown in the table:

Ingredient (oz/lb of product)

PRODUCTO Polímero A Polymer B Polymer C Base
Airtex 4 2 4 5
Extendex 2 2 3 8
Resistex 5 3 5 2

Cauchos Perú is committed to producing at least 1200 pounds of Airtex, 500 pounds of
Extendex and 300 pounds of Resistex for next week, but the company management knows that
you can sell more of each of the three products. The current inventories of the ingredients
500 pounds of polymer A, 425 pounds of polymer B, 650 pounds of polymer C, and 1100 pounds of the
Base. Each pound of Airtex produces a profit of $7 for the company, each pound of Extendex a
profit of $6.5 and a profit of $6 for each pound of Resistex.

Output of the POMQM Software:

a) Show the respective mathematical model.

Xi: Number of pounds to be produced of product i
i=1,2,3
Max 7x1 + 6.5x2 + 6x3
St

Ing. Efraín Murillo Msc.

Operations Research I 111

4x1 + 2x2 + 5x3 <= 8000

2x1 + 2x2 + 3x3 <= 6800
4x1+3x2+5x3<=10400
5x1 + 8x2 + 2x3 <= 17600
X1>=1200
X2>=500
X3>=300
Xi >= 0
Based on the output of the POMQM software shown above, answer the following
concerns:

b) How much should Cauchos Perú produce of each product? How much profit would it report?
the sale of these products.
Airtex=1200 libras.
Extension = 850 pounds.
Resistex=300 pounds.

Ganancia=$ 15725

c) How much would you pay at most for an additional ounce of polymer A? What range is it in?
Is this value valid?
An amount <$3.25, range [7300 to 9033.33]

d) Indicate which resources were depleted. Are there any surpluses? How much?
Resource exhausted: Polymer A
Excedente: Extendex, 350 libras.

e) If you could obtain additional amounts of only one of the three polymers
Which one would you recommend? Explain.
Polymer A, since it is the only one with Dual Price > 0

f) The company wishes to increase its profits to $18000 by acquiring more quantity of the
polymer A How much more of polymer A is needed?
18000-15725= 2275
2275/3.25=700 ounces

g) If the demand for Airtex increases by 2%, what would happen to the company's profit?
0.02*1200=24, is within the permitted range
Then the utility decreases by $144 (24*6)

What would happen to the production plan if Extendex's profit decreases by 20%?
0.2*6.5=1.3, it is within the allowed range, THUS the Production Plan DOES NOT CHANGE

What would happen to the company's optimal profit if the commitment to produce 300 pounds?
is it falling by 10%?
0.1*300=30, it is within the allowed range, THEN the PROFIT INCREASES by $307.5
(30*10.25=307.5)

j) What coefficient or coefficients of gain could be doubled while keeping fixed

all other coefficients, without affecting the optimal production plan? Explain. (2
points)
Airte C1: 7*2=14, out of range, cannot be duplicated
Extendex C2: 6.5*2=13, is within the range, YES It can be doubled.
Resistex C3: 6*2=12, is within range, YES It can be doubled

Ing. Efraín Murillo Msc.

Operations Research I 112

k) What quantity of polymer B is used in the company?

6800-1800=5000 ounces
l) If a new product is considered that requires 2.2 ounces of polymer A, 3 ounces of the
polymer B, 1 ounce of polymer C and 2 ounces of Base and the utility per unit is $7.5, must
to occur?. (2 points)
2.2*3.25+3*0+1*0+2*0=7.15<7.5, then IT MUST OCCUR.
If the new product is produced, what would the new production plan be?
X1=1200
X2=500
X3=300
X4=318.8
Z=15836.36

n) Obtenga el modelo de PL DUAL del problema primal.

Min 8000Y1 + 6800Y2 + 10400Y3 + 17600Y4 - 1200Y5 - 500Y6 - 300Y7

St
4Y1 + 2Y2 + 4Y3 + 5Y4 - Y5 >= 7
2Y1 + 2Y2 + 3Y3 + 8Y4 - Y6 >= 6.5
5Y1 + 3Y2 + 5Y3 + 2Y4 - Y7 >= 6
Yi>=0
o) Define the decision variables of the dual model. (2 points)
Y1: precio/onza de polímero A
Y2: price/ounce of polymer B
Y3: price/ounce of polymer C
Y4: price/ounce of Base
Y5: price/pound of Airtex demand
Y6: price/pound of demand for Extendex
Y7: price/pound of demand for Resistex
p) Show your optimal solution of the dual model.
Y1=3.25
Y2=0
Y3=0
Y4=0
Y5=6
Y6=0
Y7=10.25
Z=15727
Show the dual values of the dual model.
Dual value 1=-1200
Dual value 2=-850
Dual value 3=-300

Ing. Efraín Murillo Msc.