CHAPTER 10

Integer Programming,
Goal Programming, and
Nonlinear Programming

LEARNING OBJECTIVES
After completing this chapter, students will be able to:

1. Understand the difference between LP and integer 3. Formulate and solve goal programming problems
programming. using Excel and QM for Windows.
2. Understand and solve the three types of integer pro- 4. Formulate nonlinear programming problems and
gramming problems. solve using Excel.

CHAPTER OUTLINE
10.1 Introduction 10.4 Goal Programming
10.2 Integer Programming 10.5 Nonlinear Programming
10.3 Modeling with 0–1 (Binary) Variables

Summary • Glossary • Solved Problems • Self-Test • Discussion Questions and Problems • Internet Homework
Problems • Case Study: Schank Marketing Research • Case Study: Oakton River Bridge • Bibliography

395
396 CHAPTER 10 • INTEGER PROGRAMMING, GOAL PROGRAMMING, AND NONLINEAR PROGRAMMING

10.1 Introduction
This chapter presents a series of other important mathematical programming models that arise
when some of the basic assumptions of LP are made more or less restrictive. For example,
one assumption of LP is that decision variables can take on fractional values such as
X1 = 0.33, X2 = 1.57, or X3 = 109.4. Yet a large number of business problems can be solved
only if variables have integer values. When an airline decides how many Boeing 757s or Boeing
777s to purchase, it can’t place an order for 5.38 aircraft; it must order 4, 5, 6, 7, or some other
Integer programming is the integer amount. In this chapter we present the general topic of integer programming, and we
extension of LP that solves specifically consider the use of special variables that must be either 0 or 1.
problems requiring integer A major limitation of LP is that it forces the decision maker to state one objective only. But
solutions. what if a business has several objectives? Management may indeed want to maximize profit, but
it might also want to maximize market share, maintain full employment, and minimize costs.
Many of these goals can be conflicting and difficult to quantify. South States Power and Light,
for example, wants to build a nuclear power plant in Taft, Louisiana. Its objectives are to maxi-
mize power generated, reliability, and safety, and to minimize cost of operating the system and
Goal programming is the the environmental effects on the community. Goal programming is an extension to LP that can
extension of LP that permits permit multiple objectives such as these.
more than one objective to be Linear programming can, of course, be applied only to cases in which the constraints and
stated. objective function are linear. Yet in many situations this is not the case. The price of various
products, for example, may be a function of the number of units produced. As more are made,
the price per unit decreases. Hence an objective function may read as follows:
Nonlinear programming is the Maximize profit = 25X1 - 0.4X1 2 + 30X2 - 0.5X2 2
case in which objectives or
constraints are nonlinear. Because of the squared terms, this is a nonlinear programming problem.
Let’s examine each of these extensions of LP—integer, goal, and nonlinear programming—
one at a time.

10.2 Integer Programming

Solution values must be whole An integer programming model is a model that has constraints and an objective function identi-
numbers in integer programming. cal to that formulated by LP. The only difference is that one or more of the decision variables
has to take on an integer value in the final solution. There are three types of integer program-
ming problems:
There are three types 1. Pure integer programming problems are cases in which all variables are required to have
of integer programs: pure integer values.
integer programming; 2. Mixed-integer programming problems are cases in which some, but not all, of the decision
mixed-integer programming;
variables are required to have integer values.
and 0–1 integer programming.
3. Zero–one integer programming problems are special cases in which all the decision
variables must have integer solution values of 0 or 1.
Solving an integer programming problem is much more difficult than solving an LP prob-
lem. The solution time required to solve some of these may be excessive even on the fastest
computer.

Harrison Electric Company Example of Integer Programming

The Harrison Electric Company, located in Chicago’s Old Town area, produces two products
popular with home renovators: old-fashioned chandeliers and ceiling fans. Both the chandeliers
and fans require a two-step production process involving wiring and assembly. It takes about
2 hours to wire each chandelier and 3 hours to wire a ceiling fan. Final assembly of the chande-
liers and fans requires 6 and 5 hours, respectively. The production capability is such that only
12 hours of wiring time and 30 hours of assembly time are available. If each chandelier
 10.2 INTEGER PROGRAMMING 397

produced nets the firm $7 and each fan $6, Harrison’s production mix decision can be formu-
lated using LP as follows:
Maximize profit = $7X1 + $6X2
subject to 2X1 + 3X2 … 12 1wiring hours2
Although enumeration is feasible 6X1 + 5X2 … 30 1assembly hours2
for some small integer program- X1, X2 Ú 0
ming problems, it can be difficult
or impossible for large ones. where
X1 = number of chandeliers produced
X2 = number of ceiling fans produced
With only two variables and two constraints, Harrison’s production planner, Wes Wallace,
employed the graphical LP approach (see Figure 10.1) to generate the optimal solution of
X1 = 3.75 chandeliers and X2 = 1.5 ceiling fans during the production cycle. Recognizing that
the company could not produce and sell a fraction of a product, Wes decided that he was dealing
with an integer programming problem.
It seemed to Wes that the simplest approach was to round off the optimal fractional solu-
tions for X1 and X2 to integer values of X1 = 4 chandeliers and X2 = 2 ceiling fans. Unfortu-
Rounding off is one way to reach nately, rounding can produce two problems. First, the new integer solution may not be in the
integer solution values, but it feasible region and thus is not a practical answer. This is the case if we round to X1 = 4, X2 = 2.
often does not yield the best Second, even if we round off to a feasible solution, such as X1 = 4, X2 = 1, it may not be the
solution. optimal feasible integer solution.
Listing all feasible solutions and selecting the one with the best objective function value is
called the enumeration method. Obviously this can be quite tedious for even small problems,
and it is virtually impossible for large problems as the number of feasible integer solutions is
extremely large.

FIGURE 10.1
X2
Harrison Electric Problem
6

6X1 + 5X2 ≤ 30

+ = Possible Integer Solution

Optimal LP Solution
2 (X1 = 3.75, X2 = 1.5, Profit = $35.25)

2X1 + 3X2 ≤ 12
1

0 1 2 3 4 5 6 X1
398 CHAPTER 10 • INTEGER PROGRAMMING, GOAL PROGRAMMING, AND NONLINEAR PROGRAMMING

TABLE 10.1 CHANDELIERS CEILING FANS PROFIT

Integer Solutions to 1X12 1X22 1$7X1 # $6X22
the Harrison Electric 0 0 $0
Company Problem
1 0 7
2 0 14
3 0 21
4 0 28
5 0 35 Optimal solution
to integer programming
0 1 6
problem
1 1 13
2 1 20
3 1 27
4 1 34 Solution if rounding
is used
0 2 12
1 2 19
2 2 26
3 2 33
0 3 18
1 3 25
0 4 24

An important concept to Table 10.1 lists the entire set of integer-valued solutions to the Harrison Electric problem.
understand is that an integer By inspecting the right-hand column, we see that the optimal integer solution is
programming solution can never
be better than the solution to the X1 = 5 chandeliers, X2 = 0 ceiling fans, with a profit = $35
same LP problem. The integer
Note that this integer restriction results in a lower profit level than the original optimal LP solu-
problem is usually worse in terms
of higher cost or lower profit.
tion. As a matter of fact, an integer programming solution can never produce a greater profit
than the LP solution to the same problem; usually, it means a lesser value.

Using Software to Solve the Harrison Integer Programming Problem

QM for Windows and Excel spreadsheets are capable of handling integer programming prob-
lems such as the Harrison Electric case. Program 10.1A illustrates the input data to QM for
Windows, and Program 10.1B provides the results.
To use QM for Windows, select the Integer & Mixed Integer Programming module. Specify
the number of constraints and the number of variables. Program 10.1A provides the input
screen, with data entered for the Harrison Electric example. The last row of the table allows you
to classify each variable according to type of variable (Integer, Real, or 0–1). Once all variables
have been correctly specified, click Solve, and you will see the output in Program 10.1B.
Solver in Excel 2010 can also be used to solve this problem, as shown in Program 10.2. The
Solver parameters and selections are shown, and the key formulas are displayed. To specify that
the variables must be integers, a special constraint is entered in Solver. After opening the Solver
Parameters window, select Add just as you would do to enter other constraints. When the
Add Constraint window opens, enter the range containing the solution values, as shown in
Program 10.2. Then click the tab to open the drop-down menu and then change the type of con-
straint to int, for integer. Click OK to return to the Solver Parameters window. Then enter the
other constraints, specify the parameters and selections, and click Solve. The solution is shown
to be 5 chandeliers and 0 fans, for a profit of $35.
 10.2 INTEGER PROGRAMMING 399

PROGRAM 10.1A
QM for Windows Input
Screen for Harrison
Electric Problem

PROGRAM 10.1B
QM for Windows
Solution Screen for
Harrison Electric Problem

PROGRAM 10.2
Excel 2010 Solver
Solution for Harrison
Electric Problem

Solver Parameter Inputs and Selections Key Formulas

Set Objective: D5
By Changing cells: B4:C4
To: Max
Subject to the Constraints:
Copy D5 to D8:D9
D8:D9 <= F8:F9
B4:C4 = integer
Solving Method: Simplex LP
! Make Variables Non-Negative
400 CHAPTER 10 • INTEGER PROGRAMMING, GOAL PROGRAMMING, AND NONLINEAR PROGRAMMING

Mixed-Integer Programming Problem Example

Although the Harrison Electric example was a pure integer problem, there are many situations
in which some of the variables are restricted to be integers and others are not. The following is
an example of such a mixed-integer programming problem.
Bagwell Chemical Company, in Jackson, Mississippi, produces two industrial chemicals.
The first product, xyline, must be produced in 50-pound bags; the second, hexall, is sold by the
pound in dry bulk and hence can be produced in any quantity. Both xyline and hexall are com-
posed of three ingredients—A, B, and C—as follows:

AMOUNT PER 50-POUND BAG AMOUNT PER POUND AMOUNT OF INGREDIENTS

OF XYLINE (LB) OF HEXALL (LB) AVAILABLE
30 0.5 2,000 lb—ingredient A
18 0.4 800 lb—ingredient B
2 0.1 200 lb—ingredient C

Bagwell sells 50-pound bags of xyline for $85 and hexall in any weight for $1.50 per pound.

Integer Programming
MODELING IN THE REAL WORLD: at the USPS

Defining
Defining the Problem
the Problem The U. S. Postal Service (USPS) operates one of the largest transporation networks in the world, delivering
one-fifth of a trillion items every year. The inherent transportation-related problems are, obviously, very
large. Nevertheless, USPS’s problem is how to deliver mail in the most cost-efficient manner possible.

Developing Developing a Model

a Model A large-scale integer program known as the Highway Corridor Analytical Program (HCAP) was developed
to help solve the problem. More specifically, the HCAP solves a Vehicle Routing Problem (VRP) with known
pickup and delivery locations. That is, the model takes into account all the different modes of transporta-
tion, the capacities inherent to the system, all the pickup locations, and all the delivery locations and then
assigns trucks to routes as the (binary) decision variable of interest.

Acquiring
Acquiring Input Data
Input Data Geographic information system (GIS) data of all the pickup and delivery locations is integrated into the
model. Realistic time and distance constraints were placed on the model to prevent drivers from being
assigned a pickup in one area and a delivery in an area too far away.

Testing the
Testing the Solution
Solution The model was loaded into a large-scale mathematical programming solver. Several versions and models
were tested.

Analyzing
Analyizing the Results
the Results Decision makers found improvements in several areas. For example, one of the model outputs resulted in
a 20% reduction in redundant trips.

Implementing
Implementing the Results
the Results USPS has already realized over $5 million in transportation savings due to its implementation of the HCAP
integer programming optimization model. Efforts are under way to seek out additional efficiencies through
the use of HCAP.

Source: Based on A. Pajunas, E. J. Matto, M. Trick, and L.F. Zuluaga. “Optimizing Highway Transporation at the United States
Postal Service,” Interfaces 37, 6(2007): 515-525.
 10.2 INTEGER PROGRAMMING 401

If we let X = number of 50-pound bags of xyline produced and Y = number of pounds of hexall
(in dry bulk) mixed, Bagwell’s problem can be described with mixed-integer programming:
Maximize profit = $85X + $1.50Y
subject to 30X + 0.5Y … 2,000
18X + 0.4Y … 800
2X + 0.1Y … 200
X, Y Ú 0 and X integer.
Note that Y represents bulk weight of hexall and is not required to be integer valued.
USING QM FOR WINDOWS AND EXCEL TO SOLVE BAGWELL’S INTEGER PROGRAMMING
MODEL The solution to Bagwell’s problem is to produce 44 bags of xyline and 20 pounds of
hexall, yielding a profit of $3,770. (The optimal linear solution, by the way, is to produce 44.444
bags of xyline and 0 pounds of hexall, yielding a profit of $3,777.78.) This is first illustrated in
Program 10.3, which uses the Mixed Integer Programming module in QM for Windows. Note
that variable X is identified as Integer, while Y is Real in Program 10.3.
In Program 10.4, we use Excel to provide an alternative solution method.

PROGRAM 10.3
QM for Windows
Solution for Bagwell Limits are used, and the best solution
Chemical Problem available after a certain time is presented.

Notice that only X must be integer,

while Y may be any real number.

Mixed-Integer Programming in the IBM

IN ACTION Supply Chain

T he manufacture of semiconductors is a very expensive opera-

tion, often requiring investments in the billions of dollars. The IBM
While LP is commonly used in supply-chain modeling, it is neces-
sary to use models in which some of the variables are required to
be integers. The resulting MIP models are so large (millions of vari-
Systems and Technology Group has used mixed-integer program- ables) that they cannot be solved with even the fastest computers.
ming (MIP) along with other operations research techniques to Therefore, the IBM Systems and Technology Group developed
plan and optimize its semiconductor supply chain. The business heuristic methods to solve the optimization models as part of the
of supply-chain optimization (SCO) has been deemed business company’s advanced planning systems.
critical and must incorporate a variety of planning criteria and The benefits of the CPE are many. The on-time deliveries
constraints. improved by 15%. A 25% to 30% reduction in inventory was
IBM uses a central planning engine (CPE) to balance the supply observed as a result of the model. The company also observed an
chain’s resources against the semiconductor demand. The MIP is an improvement in asset allocation of between 2% and 4% of costs.
important part of this CPE. The MIP model is a cost-minimization This model allowed “what-if” questions to be quickly answered—
problem with constraints related to material flows and other aspect a process that was not possible in the past. Strategic planning
of the supply chain. was also facilitated by the CPE. IBM benefited greatly from the
Some of the models involved in SCO include sourcing among use of MIP in managing the semiconductor supply chain.
multiple plants, the logistics of interplant shipping, and the devel-
opment of production plans for all the plants within the system. Source: Based on Brian. T. Denton, John Forrest, and R. John Milne. “IBM
They may involve a planning horizon of a few days, a few months, Solves a Mixed-Integer Program to Optimize Its Semiconductor Supply
or even a few years, depending on the specific type of application. Chain,” Interfaces 36, 5 (September–October 2006): 386–399.
402 CHAPTER 10 • INTEGER PROGRAMMING, GOAL PROGRAMMING, AND NONLINEAR PROGRAMMING

PROGRAM 10.4
Excel 2010 Solver
Solution for Bagwell
Chemical Problem

Solver Parameter Inputs and Selections Key Formulas

Set Objective: D5
By Changing cells: B4:C4
To: Max
Subject to the Constraints:
D8:D10 <= F8:F10 Copy D5 to D8:D10
B4 = integer
Solving Method: Simplex LP
! Make Variables Non-Negative

10.3 Modeling with 0–1 (Binary) Variables

In this section we demonstrate how 0–1 variables can be used to model several diverse situa-
tions. Typically a 0–1 variable is assigned a value of 0 if a certain condition is not met and 1 if
the condition is met. Another name for a 0–1 variable is a binary variable. A common problem
of this type, the assignment problem, involves deciding which individuals to assign to a set of
jobs. (This is discussed in Chapter 9.) In this assignment problem, a value of 1 indicates a per-
son is assigned to a specific job, and a value of 0 indicates the assignment was not made. We
present other types of 0–1 problems to show the wide applicability of this modeling technique.

Capital Budgeting Example

A common capital budgeting decision involves selecting from a set of possible projects when
budget limitations make it impossible to select all of these. A separate 0–1 variable can be
defined for each project. We will see this in the following example.
Quemo Chemical Company is considering three possible improvement projects for its plant:
a new catalytic converter, a new software program for controlling operations, and expanding the
warehouse used for storage. Capital requirements and budget limitations in the next two years
prevent the firm from undertaking all of these at this time. The net present value (the future value
of the project discounted back to the present time) of each of the projects, the capital require-
ments, and the available funds for the next two years are given in Table 10.2.
To formulate this as an integer programming problem, we identify the objective function
and the constraints as follows:
Maximize net present value of projects undertaken
subject to Total funds used in year 1 … $20,000
Total funds used in year 2 … $16,000
 10.3 MODELING WITH 0–1 (BINARY) VARIABLES 403

TABLE 10.2
PROJECT NET PRESENT VALUE YEAR 1 YEAR 2
Quemo Chemical
Company Information Catalytic Converter $25,000 $8,000 $7,000
Software $18,000 $6,000 $4,000
Warehouse Expansion $32,000 $12,000 $8,000
Available Funds $20,000 $16,000

We define the decision variables as

1 if catalytic converter project is funded
X1 = b
0 otherwise
1 if software project is funded
X2 = b
0 otherwise
1 if warehouse expansion project is funded
X3 = b
0 otherwise
The mathematical statement of the integer programming problem becomes
Maximize NPV = 25,000X1 + 18,000X2 + 32,000X3
subject to 8,000X1 + 6,000X2 + 12,000X3 … 20,000
7,000X1 + 4,000X2 + 8,000X3 … 16,000
X1, X2, X3 = 0 or 1
Program 10.5 provides the Solver solution in Excel 2010. You specify the variables to be binary
(0–1) by selecting bin from the Add Constraint window. The optimal solution is X1 = 1,

PROGRAM 10.5
Excel 2010 Solver
Solution for Quemo
Chemical Problem

Solver Parameter Inputs and Selections Key Formulas

Set Objective: E5
By Changing cells: B4:D4
To: Max
Subject to the Constraints: Copy E5 to E8:E9
E8:E9 <= G8:G9
B4:D4 = binary
Solving Method: Simplex LP
! Make Variables Non-Negative
404 CHAPTER 10 • INTEGER PROGRAMMING, GOAL PROGRAMMING, AND NONLINEAR PROGRAMMING

X2 = 0, X3 = 1with an objective function value of 57,000. This means that Quemo should fund
the catalytic converter project and the warehouse expansion project but not the new software
project. The net present value of these investments will be $57,000.

Limiting the Number of Alternatives Selected

One common use of 0–1 variables involves limiting the number of projects or items that are se-
lected from a group. Suppose that in the Quemo Chemical Company example, the company is
required to select no more than two of the three projects regardless of the funds available. This
could be modeled by adding the following constraint to the problem:
X1 + X2 + X3 … 2
If we wished to force the selection of exactly two of the three projects for funding, the following
constraint should be used:
X1 + X2 + X3 = 2
This forces exactly two of the variables to have values of 1, whereas the other variable must have
a value of 0.

Dependent Selections
At times the selection of one project depends in some way upon the selection of another project.
This situation can be modeled with the use of 0–1 variables. Now suppose in the Quemo Chem-
ical problem that the new catalytic converter could be purchased only if the software was pur-
chased also. The following constraint would force this to occur:
X1 … X2
or, equivalently,
X1 - X2 … 0
Thus, if the software is not purchased, the value of X2 is 0, and the value of X1 must be 0 also
because of this constraint. However, if the software is purchased 1X2 = 12, then it is possible
that the catalytic converter could be purchased 1X1 = 12 also, although this is not required.
If we wished for the catalytic converter and the software projects to either both be selected
or both not be selected, we should use the following constraint:
X1 = X2
or, equivalently,
X1 - X2 = 0
Thus, if either of these variables is equal to 0, the other must be 0 also. If either of these is equal
to 1, the other must be 1 also.

Fixed-Charge Problem Example

Often businesses are faced with decisions involving a fixed charge that will affect the cost of fu-
ture operations. Building a new factory or entering into a long-term lease on an existing facility
would involve a fixed cost that might vary depending upon the size of the facility and the loca-
tion. Once a factory is built, the variable production costs will be affected by the labor cost in
the particular city where it is located. An example follows.
Sitka Manufacturing is planning to build at least one new plant, and three cities are being
considered: Baytown, Texas; Lake Charles, Louisiana; and Mobile, Alabama. Once the plant or
plants have been constructed, the company wishes to have sufficient capacity to produce at least
38,000 units each year. The costs associated with the possible locations are given in Table 10.3.
In modeling this as an integer program, the objective function is to minimize the total of the
fixed cost and the variable cost. The constraints are: (1) total production capacity is at least 38,000;
(2) number of units produced at the Baytown plant is 0 if the plant is not built, and it is no more
than 21,000 if the plant is built; (3) number of units produced at the Lake Charles plant is 0 if the
plant is not built, and it is no more than 20,000 if the plant is built; (4) number of units produced at
the Mobile plant is 0 if the plant is not built, and it is no more than 19,000 if the plant is built.
 10.3 MODELING WITH 0–1 (BINARY) VARIABLES 405

TABLE 10.3
ANNUAL VARIABLE COST ANNUAL
Fixed and Variable SITE FIXED COST PER UNIT CAPACITY
Costs for Sitka
Baytown, TX $340,000 $32 21,000
Manufacturing
Lake Charles, LA $270,000 $33 20,000
Mobile, AL $290,000 $30 19,000

Then we define the decision variables as

1 if factory is built in Baytown
X1 = b
0 otherwise
1 if factory is built in Lake Charles
X2 = b
0 otherwise
1 if factory is built in Mobile
The number of units produced X2 = b
0 otherwise
must be 0 if the plant is not built.
X4 = number of units produced at Baytown plant
X5 = number of units produced at Lake Charles plant
X6 = number of units produced at Mobile plant
The integer programming problem formulation becomes
Minimize cost = 340,000X1 + 270,000X2 + 290,000X3 + 32X4 + 33X5 + 30X6
subject to X4 + X5 + X6 Ú 38,000
X4 … 21,000X1
X5 … 20,000X2
X6 … 19,000X3
X1, X2, X3 = 0 or 1; X4, X5, X6 Ú 0 and integer
Notice that if X1 = 0 (meaning Baytown plant is not built), then X4 (number of units produced
at Baytown plant) must equal zero also due to the second constraint. If X1 = 1, then X4 may be
any integer value less than or equal to the limit of 21,000. The third and fourth constraints are
similarly used to guarantee that no units are produced at the other locations if the plants are not
built. The optimal solution, shown in Program 10.6, is
X1 = 0, X2 = 1, X3 = 1, X4 = 0, X5 = 19,000, X6 = 19,000
Objective function value = 1,757,000
This means that factories will be built at Lake Charles and Mobile. Each of these will produce
19,000 units each year, and the total annual cost will be $1,757,000.

Financial Investment Example

Numerous financial applications exist with 0–1 variables. A very common type of problem in-
Here is an example of stock
volves selecting from a group of investment opportunities. The following example illustrates
portfolio analysis wiht 0–1 this application.
programming The Houston-based investment firm of Simkin, Simkin, and Steinberg specializes in rec-
ommending oil stock portfolios for wealthy clients. One such client has made the following
specifications: (1) at least two Texas oil firms must be in the portfolio, (2) no more than one in-
vestment can be made in foreign oil companies, (3) one of the two California oil stocks must
be purchased. The client has up to $3 million available for investments and insists on purchas-
ing large blocks of shares of each company that he invests in. Table 10.4 describes various
stocks that Simkin considers. The objective is to maximize annual return on investment subject
to the constraints.
406 CHAPTER 10 • INTEGER PROGRAMMING, GOAL PROGRAMMING, AND NONLINEAR PROGRAMMING

PROGRAM 10.6
Excel 2010 Solver
Solution for Sitka
Manufacturing Problem

Solver Parameter Inputs and Selections Key Formulas

Set Objective: H5
By Changing cells: B4:G4
To: Min
Subject to the Constraints: Copy H5 to H8:H11
H8 >= J8
H9:H11 <= J9:J11
B4:D4 = binary
Solving Method: Simplex LP
! Make Variables Non-Negative

TABLE 10.4
EXPECTED ANNUAL COST FOR BLOCK OF
Oil Investment STOCK COMPANY NAME RETURN ($1,000s) SHARES ($1,000s)
Opportunities
1 Trans-Texas Oil 50 480
2 British Petroleum 80 540
3 Dutch Shell 90 680
4 Houston Drilling 120 1,000
5 Texas Petroleum 110 700
6 San Diego Oil 40 510
7 California Petro 75 900

To formulate this as a 0–1 integer programming problem, Simkin lets Xi be a 0–1 integer
variable, where Xi = 1 if stock i is purchased and Xi = 0 if stock i is not purchased:
Maximize return = 50X1 + 80X2 + 90X3 + 120X4 + 110X5 + 40X6 + 75X7
subject to X1 + X4 + X5 Ú 2 1Texas constraint2
X2 + X3 … 1 1foreign oil constraint2
X6 + X7 = 1 1California constraint2
480X1 + 540X2 + 680X3 + 1,000X4 + 700X5 + 510X6 + 900X7 … 3,000
1$3 million limit2
Xi = 0 or 1 for all i
The solution from using Solver in Excel 2010 is shown is Program 10.7

10.4 Goal Programming

In today’s business environment, profit maximization or cost minimization are not always the
Firms usually have more than only objectives that a firm sets forth. Often, maximizing total profit is just one of several goals,
one goal. including such contradictory objectives as maximizing market share, maintaining full employ-
ment, providing quality ecological management, minimizing noise level in the neighborhood,
and meeting numerous other noneconomic goals.
 10.4 GOAL PROGRAMMING 407

PROGRAM 10.7
Excel 2010 Solver
Solution for Financial
Investment Problem

Solver Parameter Inputs and Selections Key Formulas

Set Objective: I5
By Changing cells: B4:H4
To: Max
Subject to the Constraints: Copy I5 to I7:I10
I7 >= K7
I8 <= K8
I9 = K9
I10 <= K10
B4:H4 = binary
Solving Method: Simplex LP
! Make Variables Non-Negative

Continental Airlines Saves $40 Million

IN ACTION Using CrewSolver

A irlines use state-of-the-art processes and automated tools to

develop schedules that maximize profit. These schedules will
In late 2000 and throughout 2001, Continental and other air-
lines experienced four major disruptions. The first two were due to
severe snowstorms in early January and in March 2001. The Hous-
assign aircraft to specific routes, and then schedule pilots and ton floods caused by Tropical Storm Allison closed a major hub in
flight attendant crews to each of these aircraft. When disruptions June 2001 and left aircraft in locations where they were not sched-
occur, planes and personnel are often left in positions where they uled to be. The terrorist attacks of September 11, 2001 left aircraft
are unable to adhere to the next day’s assignments. Airlines face and crews scattered about and totally disrupted the flight sched-
schedule disruptions from a variety of unexpected reasons such ules. The CrewSolver system provided a faster and more efficient
as bad weather, mechanical problems, and crew unavailability. recovery than had been possible in the past. It is estimated that the
In 1993, Continental Airlines began an effort to develop a CrewSolver system saved approximately $40 million for these ma-
system of dealing with disruptions in real time. Working with jor disruptions in 2001. This system also saved additional money
CALEB Technologies, Continental developed the CrewSolver and and made recovery much easier when there were minor disruptions
OptSolver systems (based on 0–1 integer programming models) due to local weather problems at other times throughout the year.
to produce comprehensive recovery solutions for both aircraft and
crews. These solutions retain revenue and promote customer sat- Source: Based on Gang Yu, Michael Arguello, Gao Song, Sandra M.
isfaction by reducing flight cancellations and minimizing delays. McCowan, and Anna White. “A New Era for Crew Recovery at Continental
These crew recovery solutions are low cost while maintaining a Airlines,” Interfaces 33, (January–February 2003): 5–22.
high quality of life for pilots and flight attendants.

The shortcoming of mathematical programming techniques such as linear and integer

Goal programming permits programming is that their objective function is measured in one dimension only. It’s not possible
multiple goals. for LP to have multiple goals unless they are all measured in the same units (such as dollars), a
highly unusual situation. An important technique that has been developed to supplement LP is
called goal programming.
408 CHAPTER 10 • INTEGER PROGRAMMING, GOAL PROGRAMMING, AND NONLINEAR PROGRAMMING

Goal programming is capable of handling decision problems involving multiple goals.

A four-decade old concept, it began with the work of Charnes and Cooper in 1961 and was
refined and extended by Lee and Ignizio in the 1970s and 1980s (see the Bibliography).
In typical decision-making situations, the goals set by management can be achieved only at
the expense of other goals. It is necessary to establish a hierarchy of importance among these
Goal programming “satisfices,” goals so that lower-priority goals are tackled only after higher-priority goals are satisfied. Since
as opposed to LP, which tries to it is not always possible to achieve every goal to the extent the decision maker desires, goal pro-
“optimize.” This means coming gramming attempts to reach a satisfactory level of multiple objectives. This, of course, differs
as close as possible to reaching from LP, which tries to find the best possible outcome for a single objective. Nobel laureate
goals. Herbert A. Simon, of Carnegie-Mellon University, states that modern managers may not be able
to optimize, but may instead have to “satisfice” or “come as close as possible” to reaching goals.
This is the case with models such as goal programming.
The objective function is the How, specifically, does goal programming differ from LP? The objective function is the
main difference between goal main difference. Instead of trying to maximize or minimize the objective function directly, with
programming and LP. goal programming we try to minimize deviations between set goals and what we can actually
achieve within the given constraints. In the LP simplex approach, such deviations are called
slack and surplus variables. Because the coefficient for each of these in the objective function is
In goal programming we want to zero, slack and surplus variables do not have an impact on the optimal solution. In goal program-
minimize deviational variables, ming, the deviational variables are typically the only variables in the objective function, and the
which are the only terms in the objective is to minimize the total of these deviational variables.
objective function. When the goal programming model is formulated, the computational algorithm is almost
the same as a minimization problem solved by the simplex method.

Example of Goal Programming: Harrison Electric Company Revisited

To illustrate the formulation of a goal programming problem, let’s look back at the Harrison
Electric Company case presented earlier in this chapter as an integer programming problem.
That problem’s LP formulation, you recall, is
Maximize profit = $7X1 + $6X2
subject to 2 X1 + 3X2 … 12 1wiring hours2
6X1 + 5X2 … 30 1assembly hours2
X1, X2 Ú 0
where
X1 = number of chandeliers produced
X2 = number of ceiling fans produced
We saw that if Harrison’s management had a single goal, say profit, LP could be used to
find the optimal solution. But let’s assume that the firm is moving to a new location during a
particular production period and feels that maximizing profit is not a realistic goal. Management
sets a profit level, which would be satisfactory during the adjustment period, of $30. We now
have a goal programming problem in which we want to find the production mix that achieves
this goal as closely as possible, given the production time constraints. This simple case will pro-
vide a good starting point for tackling more complicated goal programs.
We first define two deviational variables:
d1 - = underachievement of the profit target
d1 + = overachievement of the profit target
Now we can state the Harrison Electric problem as a single-goal programming model:
Minimize under or overachievement of profit target = d1 - + d1 +
subject to $7X1 + $6X2 + d1 - - d1 + = $30 1profit goal constraint2
2X1 + 3X2 … 12 1wiring hours constraint2
6X1 + 5X2 … 30 1assembly hours constraint2
X1, X2, d1 -, d1 + Ú 0
 10.4 GOAL PROGRAMMING 409

Note that the first constraint states that the profit made, $7X1 + $6X2, plus any underachieve-
ment of profit minus any overachievement of profit has to equal the target of $30. For example,
if X1 = 3 chandeliers and X2 = 2 ceiling fans, then $33 profit has been made. This exceeds $30
by $3, so d1+ must be equal to 3. Since the profit goal constraint was overachieved, Harrison did
not underachieve and d1- will clearly be equal to zero. This problem is now ready for solution
by a goal programming algorithm.
If the target profit of $30 is exactly achieved, we see that both d1+ and d1- are equal to zero.
Deviational variables are zero if The objective function will also be minimized at zero. If Harrison’s management was only con-
a goal is completely obtained. cerned with underachievement of the target goal, how would the objective function change? It
would be as follows: minimize underachievement = d1- . This is also a reasonable goal since
the firm would probably not be upset with an overachievement of its target.
In general, once all goals and constraints are identified in a problem, management should
analyze each goal to see if underachievement or overachievement of that goal is an acceptable
situation. If overachievement is acceptable, the appropriate d + variable can be eliminated from
the objective function. If underachievement is okay, the d - variable should be dropped. If man-
agement seeks to attain a goal exactly, both d - and d + must appear in the objective function.

Extension to Equally Important Multiple Goals

We need a clear definition of Let’s now look at the situation in which Harrison’s management wants to achieve several goals,
deviational variables, such as each equal in priority.
these.
Goal 1: to produce profit of $30 if possible during the production period
Goal 2: to fully utilize the available wiring department hours
Goal 3: to avoid overtime in the assembly department
Goal 4: to meet a contract requirement to produce at least seven ceiling fans
The deviational variables can be defined as follows:
d1- = underachievement of the profit target
d1+ = overachievement of the profit target
d2- = idle time in the wiring department 1underutilization2
d2+ = overtime in the wiring department 1overutilization2
d3- = idle time in the assembly department 1underutilization2
d3+ = overtime in the assembly department 1overutilization2
d4- = underachievement of the ceiling fan goal
d4+ = overachievement of the ceiling fan goal
Management is unconcerned about whether there is overachievement of the profit goal, over-
time in the wiring department, idle time in the assembly department, or more than seven ceiling
fans are produced: hence, d1+, d2+, d3- , and d4+ may be omitted from the objective function. The
new objective function and constraints are
Minimize total deviation = d1- + d2- + d3+ + d4-
subject to 7X1 + 6X2 + d1- - d1+ = 30 1profit constraint2
2X1 + 3X2 + d2- - d2+ = 12 1wiring hours constraint2
6X1 + 5X2 + d3- - d3+ = 30 1assembly constraint2
X2 + d4- - d4+ = 7 1ceiling fan constraint2
All Xi, di variables Ú 0.

Ranking Goals with Priority Levels

A key idea in goal programming In most goal programming problems, one goal will be more important than another, which in
is that one goal is more turn will be more important than a third. The idea is that goals can be ranked with respect to their
important than another. importance in management’s eyes. Lower-order goals are considered only after higher-order
Priorities are assigned to each goals are met. Priorities (P’i s) are assigned to each deviational variable—with the ranking that
deviational variable. P1 is the most important goal, P2 the next most important, then P3, and so on.
410 CHAPTER 10 • INTEGER PROGRAMMING, GOAL PROGRAMMING, AND NONLINEAR PROGRAMMING

Let’s say Harrison Electric sets the priorities shown in the following table:

GOAL PRIORITY
Reach a profit as much above $30 as possible P1
Fully use wiring department hours available P2
Avoid assembly department overtime P3
Produce at least seven ceiling fans P4

Priority 1 is infinitely more This means, in effect, that the priority of meeting the profit goal 1P12 is infinitely more
important than Priority 2, which important than the wiring goal 1P22, which is, in turn, infinitely more important than the assem-
is infinitely more important than bly goal 1P32, which is infinitely more important than producing at least seven ceiling fans 1P42.
the next goal, and so on. With ranking of goals considered, the new objective function becomes
Minimize total deviation = P1d1- + P2d2- + P3d3+ + P4d4-
The constraints remain identical to the previous ones.

Goal Programming with Weighted Goals

When priority levels are used in goal programming, any goal in the top priority level is infinitely
more important than the goals in lower priority levels. However, there may be times when one
goal is more important than another goal, but it may be only two or three times as important.
Instead of placing these goals in different priority levels, they would be placed in the same pri-
ority level but with different weights. When using weighted goal programming, the coefficients
in the objective function for the deviational variables include both the priority level and the
weight. If all goals are in the same priority level, then simply using the weights as objective
function coefficients is sufficient.
Consider the Harrison Electric example, in which the least important goal is goal 4 (pro-
duce at least seven ceiling fans). Suppose Harrison decides to add another goal of producing at
least two chandeliers. The goal of seven ceiling fans is considered twice as important as this
goal, so both of these should be in the same priority level. The goal of 2 chandeliers is assigned

The Use of Goal Programming for Tuberculosis

IN ACTION Drug Allocation in Manila

A llocation of resources is critical when applied to the health in-

dustry. It is a matter of life and death when neither the right sup-
minimizing the underachievement in the allocation of anti-TB
drugs to the 45 centers). Four goal constraints considered the
interrelationships among variables in the distribution system.
ply nor the correct quantity is available to meet patient demand. Goal 1 was to satisfy the medication requirement (a 6-month
This was the case faced by the Manila (Philippines) Health Center, regimen) for each patient. Goal 2 was to supply each health
whose drug supply to patients afflicted with Category I tubercu- center with the proper allocation. Goal 3 was to satisfy the
losis (TB) was not being efficiently allocated to its 45 regional cure rate of 85%. Goal 4 was to satisfy the drug requirements
health centers. When the TB drug supply does not reach patients of each health center.
on time, the disease becomes worse and can result in death. Only The goal programming model successfully dealt with all of
74% of TB patients were being cured in Manila, 11% short of the these goals and raised the TB cure rate to 88%, a 13% improve-
85% target cure rate set by the government. Unlike other dis- ment in drug allocation over the previous distribution approach.
eases, TB can only be treated with four medicines and cannot be This means that 335 lives per year were saved through this
cured by alternative drugs. thoughtful use of goal programming.
Researchers at the Mapka Institute of Technology set out
to create a model, using goal programming, to optimize the Source: Based on G. J. C. Esmeria. “An Application of Goal Programming in
allocation of resources for TB treatment while considering sup- the Allocation of Anti-TB Drugs in Rural Health Centers in the Philippines,”
ply constraints. The objective function of the model was to Proceedings of the 12th Annual Conference of the Production and Operations
Management Society (March 2001), Orlando, FL: 50.
meet the target cure rate of 85% (which is the equivalent of
 10.5 NONLINEAR PROGRAMMING 411

PROGRAM 10.8A
Harrison Electric’s Goal
Programming Analysis
Using QM for Windows:
Inputs

PROGRAM 10.8B
Summary Solution
Screen for Harrison
Electric’s Goal
Programming Problem
Using QM for Windows

a weight of 1, while the 7 ceiling fan goal will be given a weight of 2. Both of these will be in
priority level 4. A new constraint (goal) would be added:
X1 + d5- - d5+ = 2 1chandeliers2
The new objective function value would be
Minimize total deviation = P1d1- + P2d2- + P3d3+ + P412d4-2 + P4d5-
Note that the ceiling fan goal has a weight of 2. The weight for the chandelier goal is 1. Techni-
cally all of the goals in the other priority levels are assigned weights of 1 also.
USING QM FOR WINDOWS TO SOLVE HARRISON’S PROBLEM QM for Windows goal program-
ming module is illustrated in Programs 10.8A and 10.8B. The input screen is shown first, in
Program 10.8A. Note in this first screen that there are two priority level columns for each con-
straint. For this example, the priority for either the positive or the negative deviation will be zero
since the objective function does not contain both types of deviational variables for any of these
goals. If a problem had a goal with both deviational variables in the objective function, both pri-
ority level columns for this goal (constraint) would contain values other than zero. Also, the
weight for each deviational variable contained in the objective function is listed as 1. (It is 0 if
the variable is not appearing in the objective function.) If different weights are used, they would
be placed in the appropriate weight column within one priority level.
The solution with an analysis of deviations and goal achievement is shown in Program
10.8B. We see that the first two constraints have negative deviational variables equal to 0, indi-
cating full achievement of those goals. In fact, the positive deviational variables both have val-
ues of 6, indicating overachievement of these goals by 6 units each. Goal (constraint) 3 has both
deviational variables equal to 0, indicating complete achievement of that goal, whereas goal 4
has a negative deviational variable equal to 1, indicating underachievement by 1 unit.

10.5 Nonlinear Programming

Linear, integer, and goal programming all assume that a problem’s objective function and con-
straints are linear. That means that they contain no nonlinear terms such as X13, 1>X2, log X3, or
5X1X2. Yet in many mathematical programming problems, the objective function and/or one or
more of the constraints are nonlinear.
Unlike with linear programming methods, computational procedures for solving many
nonlinear programming (NLP) problems do not always yield the optimal solution. In many
412 CHAPTER 10 • INTEGER PROGRAMMING, GOAL PROGRAMMING, AND NONLINEAR PROGRAMMING

NLP problems, a particular solution may be better than any other point nearby, but it may
not be the overall best point. This is called a local optimum, and the overall best solution is
called the global optimum. Thus, for a particular problem, a solution technique may indicate
that an optimum solution has been found, but it is only a local optimum, so there may be a
global optimum that is better. The mathematics involved in solving these problems is beyond
the scope of this text. We will rely on Solver in Excel to solve the nonlinear problems pre-
sented in this section.
In this section, we examine three categories of NLP problems and illustrate how Excel
can be used to search for the solution to these problems. In Solved Problem 10–3, we will see
how NLP in Excel can help find the best parameter to use in an exponential smoothing fore-
casting model.

Nonlinear Objective Function and Linear Constraints

Here is an example of a The Great Western Appliance Company sells two models of toaster ovens, the Microtoaster 1X12
nonlinear objective function and the Self-Clean Toaster Oven 1X22. The firm earns a profit of $28 for each Microtoaster re-
gardless of the number sold. Profits for the Self-Clean model, however, increase as more units
are sold because of fixed overhead. Profit on this model may be expressed as 21X2 + 0.25X22.
Hence the firm’s objective function is nonlinear:
Maximize profit = 28X1 + 21X2 + 0.25X22
Great Western’s profit is subject to two linear constraints on production capacity and sales
time available:
X1 + X2 … 1,000 1units of production capacity2
0.5X1 + 0.4X2 … 500 1hours of sales time available2
X1, X2 Ú 0
When an objective function contains squared terms (such as 0.25X22) and the problem’s con-
Quadratic programming contains straints are linear, it is called a quadratic programming problem. A number of useful problems
squared terms in the objective in the field of portfolio selection fall into this category. Quadratic programs can be solved by a
function. modified method of the simplex method. Such work is outside the scope of this book but can be
found in sources listed in the Bibliography.
The solution to the Great Western Appliance nonlinear programming problem is shown in
Program 10.9. This was found using Solver in Excel 2010, and there are two important features
of this spreadsheet that are different from the previous linear and integer programming examples.

Goal Programming Model for Prison

IN ACTION Expenditures in Virginia

P risons across the United States are overcrowded, and there is

need for immediate expansion of their capacity and replacement
each expenditure item; the impact of facility construction and
renovation on imprisonment, sentence length, and early releases
and parole; the mix of different facility types required by the sys-
or renovation of obsolete facilities. This study demonstrates how tem; and the personnel requirements resulting from the various
goal programming was used on the capital allocation problem capital expenditures for correctional facilities.
faced by the Department of Corrections of Virginia. The solution results for Virginia were one new maximum
The expenditure items considered by the Virginia corrections security facility for drug, alcohol, and psychiatric treatment
department included new and renovated maximum, medium, activities; one new minimum security facility for youthful
and minimum security facilities; community diversion programs; offenders; two new regular minimum security facilities; two
and personnel increases. The goal programming technique forced new community diversion programs in urban areas, renovation
all prison projects to be completely accepted or rejected. of one existing medium security and one minimum security
Model variables defined the construction, renovation, or es- facility; 250 new correctional officers; four new administrators;
tablishment of a particular type of correctional facility for a spe- 46 new treatment specialist/counselors; and six new medical
cific location or purpose and indicated the people required by personnel.
the facilities. The goal constraints fell into five categories: addi-
tional inmate capacity created by new and renovated correc- Source: Based on R. Russell, B. Taylor, and A. Keown. Computer Environ-
tional facilities; operating and personnel costs associated with mental Urban Systems 11, 4 (1986): 135–146.
 10.5 NONLINEAR PROGRAMMING 413

PROGRAM 10.9
Excel 2010 Solver
Solution for Great
Western Appliance NLP
Problem

Solver Parameter Inputs and Selections Key Formulas

Set Objective: E8
By Changing cells: B4:C4
To: Max
Subject to the Constraints:
E11:E12 <= G11:G12
Solving Method: GRG Nonlinear
! Make Variables Non-Negative

First, the solving method used for this in Solver is GRG Nonlinear instead of Simplex LP. The
second change involves the objective function and the changing cells. For the sake of consistency,
values for both X2 (cell C4) and X22 (cell D7) are shown in Program 10.9. However, cell D7 is
simply cell C4 squared. Thus, when cell C4 changes, D7 will automatically change, and the
Changing cells specified in Solver are B4:C4, while D7 is not included.

Both Nonlinear Objective Function and Nonlinear Constraints

An example in which the The annual profit at a medium-sized (200–400 beds) Hospicare Corporation–owned hospital
objective and constraints are depends on the number of medical patients admitted 1X12 and the number of surgical patients
both nonlinear. admitted 1X22. The nonlinear objective function for Hospicare is
$13X1 + $6X1X2 + $5X2 + $1>X2
An example of nonlinear The corporation identifies three constraints, two of which are also nonlinear, that affect opera-
constraints. tions. They are
2X12 + 4X2 … 90 1nursing capcity, in thousands of labor–days2
X1 + X23 … 75 1x–ray capacity, in thousands2
8X1 - 2X2 … 61 1marketing budget required, in thousands of $2
414 CHAPTER 10 • INTEGER PROGRAMMING, GOAL PROGRAMMING, AND NONLINEAR PROGRAMMING

PROGRAM 10.10
Excel 2010 Solution for
Hospicare NLP Problem

Solver Parameter Inputs and Selections Key Formulas

Set Objective: H8
By Changing cells: B4:C4
To: Max
Subject to the Constraints: Copy H8 to H11:H13
H11:H13 <= J11:J13
Solving Method: GRG Nonlinear
! Make Variables Non-Negative

Excel’s Solver is capable of formulating such a problem. The optimal solution is provided in
Program 10.10.

Linear Objective Function with Nonlinear Constraints

Thermlock Corp. produces massive rubber washers and gaskets like the type used to seal joints
on the NASA Space Shuttles. To do so, it combines two ingredients: rubber 1X12 and oil 1X22.
The cost of the industrial-quality rubber used is $5 per pound and the cost of the high-viscosity
oil is $7 per pound. Two of the three constraints Thermlock faces are nonlinear. The firm’s
objective function and constraints are
Minimize costs = $5X1 + $7X2
subject to 3X1 + 0.25X12 + 4X2 + 0.3X22 Ú 125 1hardness contraint2
13X1 + X13 Ú 80 1tensile strength2
0.7X1 + X2 Ú 17 1elasticity2
To solve this nonlinear programming, we turn again to Excel. The output is provided in
Program 10.11.
 GLOSSARY 415

PROGRAM 10.11
Excel 2010 Solution for
Thermlock NLP Problem

Solver Parameter Inputs and Selections Key Formulas

Set Objective: D5
By Changing cells: B4:C4
To: Min
Subject to the Constraints:
G10:G12 >= I10:I12
Solving Method: GRG Nonlinear
! Make Variables Non-Negative
Copy G10 to G11:G12

Summary
This chapter addresses three special types of LP problems. The The latter part of the chapter deals with goal program-
first, integer programming, examines LP problems that cannot ming. This extension of LP allows problems to have multiple
have fractional answers. We also note that there are three types goals. Again, software such as QM for Windows is a powerful
of integer programming problems: (1) pure or all-integer pro- tool in solving this offshoot of LP. Finally, the advanced topic
grams, (2) mixed problems, in which some solution variables of NLP is introduced as a special mathematical programming
need not be integer, and (3) 0–1 problems, in which all solu- problem. Excel is seen to be a useful tool in solving simple
tions are either 0 or 1. We also demonstrate how 0–1 variables NLP models.
can be used to model special situations such as fixed charge However, it is important to remember that the solution
problems. QM for Windows and Excel are used to illustrate found for an NLP problem might be a local optimum and not a
computer approaches to these problems. global optimum.

Glossary
Deviational Variables Terms that are minimized in a goal Local Optimum A solution to a nonliner programming
programming problem. Like slack variables in LP, they problem which is better than any nearby point, but which
are real. They are the only terms in the objective function. may not be the global optimum.
Global Optimum The overall best solution to a nonlinear Nonlinear Programming A category of mathematical pro-
programming problem. gramming techniques that allows the objective function
Goal Programming A mathematical programming and/or constraints to be nonlinear.
technique that permits decision makers to set and prioritize Satisficing The process of coming as close as possible to
multiple objectives. reaching your set of objectives.
Integer Programming A mathematical programming tech- 0–1 Integer Programming Problems in which all decision
nique that produces integer solutions to linear programming variables must have integer values of 0 or 1. This is also
problems. called a binary variable
416 CHAPTER 10 • INTEGER PROGRAMMING, GOAL PROGRAMMING, AND NONLINEAR PROGRAMMING

Solved Problems

Solved Problem 10-1

Consider the 0–1 integer programming problem that follows:
Maximize 50X1 + 45X2 + 48X3
subject to 19X1 + 27X2 + 34X3 … 80
22X1 + 13X2 + 12X3 … 40
X1, X2, X3 must be either 0 or 1
Now reformulate this problem with additional constraints so that no more than two of the three vari-
ables can take on a value equal to 1 in the solution. Further, make sure that if X1 = 1, then X2 = 1 also.
Then solve the new problem using Excel.
Solution
Excel can handle all-integer, mixed-integer, and 0–1 integer problems. Program 10.12 below shows two
new constraints to handle the reformulated problem. These constraints are
X1 + X2 + X3 … 2
and
X1 - X2 … 0
The optimal solution is X1 = 1, X2 = 1, X3 = 0, with an objective function value of 95.

PROGRAM 10.12
Excel 2010 Solution for
Solved Problem 10-1

Solver Parameter Inputs and Selections Key Formulas

Set Objectives: E5
By Changing cells: B4:D4
To: Max
Subject to the Constraints: Copy E5 to E8:E11
E8:E11 <= G8:G11
B4:D4 = binary
Solving Method: Simplex LP
! Make Variables Non-Negative
 SOLVED PROBLEMS 417

Solved Problem 10-2

Recall the Harrison Electric Company goal programming problem seen in Section 10.4. Its LP formula-
tion was
Maximize profit = $7X1 + $6X2
subject to 2X1 + 3X2 … 12 1wiring hours2
6X1 + 5X2 … 30 1assembly hours2
X1, X2 Ú 0
where
X1 number of chandeliers produced
X2 number of ceiling fans produced
Reformulate Harrison Electrical as a goal programming model with the following goals:
Priority 1: Produce at least 4 chandeliers and 3 ceiling fans.
Priority 2: Limit overtime in the assembly department to 10 hours and in the wiring department to
6 hours.
Priority 3: Maximize profit.
Solution
Minimize P11d1- + d2-2 + P21d3+ + d4+2 + P3d5-
subject to X1 + d1- - d1+ = 4
f Priority 1
X2 + d2- - d2+ = 3
2X1 + 3X2 + d3- - d3+ = 18
f Priority 2
6X1 + 5X2 + d4- - d4+ = 40
7X1 + 6X2 + d5- - d5+ = 99,999} Priority 3
In the priority 3 goal constraint, the 99,999 represents an unrealistically high profit. It is just a mathe-
matical trick to use as a target so that we can get as close as possible to the maximum profit.

Solved Problem 10-3

In Chapter 5, the exponential smoothing method for forecasting time series was presented. Program
10.13A provides an example of this, with the smoothing constant (α) selected to be 0.1. The fore-
cast in time period 1 is assumed to be perfect, so the error and the absolute value of the error are
both 0. The mean absolute deviation (MAD) is the measure of accuracy, and the first time period er-
ror is not used in computing this since it was simply assumed to be perfect. The MAD is very de-
pendent on the value of α. Use Excel to find the value of α that will minimize the MAD. Hint:
Instead of writing the entire objective function, simply use the cell already developed in Excel for
the MAD. Remember, α must be between 0 and 1.
Solution
The MAD is a nonlinear function, so Solver in Excel can be used to solve this. There is only one con-
straint: The smoothing constant, α, must be less than or equal to 1. This can be entered directly into the
Add Constraint window in Solver, with a 1 entered on the right-hand side of the inequality. Program
10.13B provides the information and the final solution for this. The value for α that minimizes the MAD
is 0.3478, and this yields an MAD of 16.70.
Note: Solver can be used in a similar manner to find the best weights to use in a weighted moving aver-
age forecast.
418 CHAPTER 10 • INTEGER PROGRAMMING, GOAL PROGRAMMING, AND NONLINEAR PROGRAMMING

PROGRAM 10.13A
Excel 2010 Spreadsheet
for Solved Problem 10-3

PROGRAM 10.13B
Excel 2010 Spreadsheet
Solution for Solved
Problem 10-3

Solver Parameter Inputs and Selections Key Formulas

Set Objectives: E12

By Changing cells: B3
To: Min
Subject to the Constraints: Copy C6:E6 to C7:E11
B3 <= 1
Solving Method: GRG Nonlinear
! Make Variables Non-Negative
 DISCUSSION QUESTIONS AND PROBLEMS 419

Self-Test
! Before taking the self-test, refer to the learning objectives at the beginning of the chapter, the notes in the margins, and the
glossary at the end of the chapter.
! Use the key at the back of the book to correct your answers.
! Restudy pages that correspond to any questions that you answered incorrectly or material you feel uncertain about.
1. If all of the decision variables require integer solutions, 7. Nobel Laureate Herbert A. Simon of Carnegie-Mellon
the problem is University says that modern managers should always op-
a. a pure integer programming type of problem. timize, not satisfice.
b. a simplex method type of problem. a. True
c. a mixed-integer programming type of problem. b. False
d. a Gorsky type of problem. 8. The fixed charge problem is typically classified as
2. In a mixed-integer programming problem a. a goal programming problem.
a. some integers must be even and others must be odd. b. a 0–1 integer problem.
b. some decision variables must require integer results c. a quadratic programming problem.
only and some variables must allow for continuous d. an assignment problem.
results. 9. The 0–1 integer programming problem
c. different objectives are mixed together even a. requires the decision variables to have values between
though they sometimes have relative priorities 0 and 1.
established. b. requires that the constraints all have coefficients
3. A model containing a linear objective function and linear between 0 and 1.
constraints but requiring that one or more of the decision c. requires that the decision variables have coefficients
variables take on an integer value in the final solution is between 0 and 1.
called d. requires the decision variables to be equal to 0 or 1.
a. an integer programming problem. 10. Goal programming
b. a goal programming problem. a. requires only that you know whether the goal is direct
c. a nonlinear programming problem. profit maximization or cost minimization.
d. a multiple objective LP problem. b. allows you to have multiple goals.
4. An integer programming solution can never produce a c. is an algorithm with the goal of a quicker solution to
greater profit than the LP solution to the same problem. the pure integer programming problem.
a. True d. is an algorithm with the goal of a quicker solution to
b. False the mixed-integer programming problem.
5. In goal programming, if all the goals are achieved, the 11. Nonlinear programming includes problems
value of the objective function will always be zero. a. in which the objective function is linear but some
a. True constraints are not linear.
b. False b. in which the constraints are linear but the objective
6. The objective in a goal programming problem with one function is not linear.
priority level is to maximize the sum of the deviational c. in which both the objective function and all of the con-
variables. straints are not linear.
a. True d. solvable by quadratic programming.
b. False e. all of the above.

Discussion Questions and Problems

Discussion Questions 10-3 List the advantages and disadvantages of solving inte-
ger programming problems by (a) rounding off and
10-1 Compare the similarities and differences of linear
(b) enumeration.
and goal programming.
10-4 What is the difference between the three types of
10-2 A linear programming problem was developed, and
integer programming problems? Which do you think
the feasible region was found. If the additional re-
is most common, and why?
striction that all variables must be integers were
added to the problem, how would the size of the fea- 10-5 What is meant by “satisficing,” and why is the term
sible region change? How would the optimal value often used in conjunction with goal programming?
of the objective function change? 10-6 What are deviational variables? How do they differ
from decision variables in traditional LP problems?
420 CHAPTER 10 • INTEGER PROGRAMMING, GOAL PROGRAMMING, AND NONLINEAR PROGRAMMING

10-7 If you were the president of the college you are 10-11 A group of college students is planning a camping
attending and were employing goal programming to trip during the upcoming break. The group must
assist in decision making, what might your goals be? hike several miles through the woods to get to the
What kinds of constraints would you include in your campsite, and anything that is needed on this trip
model? must be packed in a knapsack and carried to the
10-8 What does it mean to rank goals in goal program- campsite. One particular student, Tina Shawl, has
ming? How does this affect the problem’s solution? identified eight items that she would like to take on
10-9 Which of the following are NLP problems, and the trip, but the combined weight is too great to take
why? all of them. She has decided to rate the utility of
each item on a scale of 1 to 100, with 100 being the
(a) Maximize profit = 3X1 + 5X2 + 99X3 most beneficial. The item weights in pounds and
subject to X1 Ú 10 their utility values are given below.
X2 … 5
X3 Ú 18 ITEM 1 2 3 4 5 6 7 8
WEIGHT 8 1 7 6 3 12 5 14
(b) Maximize cost = 25X1 + 30X2 + 8X1X2
UTILITY 80 20 50 55 50 75 30 70
subject to X1 Ú 8
X1 + X2 Ú 12
Recognizing that the hike to the campsite is a long
0.0005X1 - X2 = 11 one, a limit of 35 pounds has been set as the maxi-
(c) Maximize Z = P1d1- + P2d2+ + P3+ mum total weight of the items to be carried.
(a) Formulate this as a 0–1 programming problem to
subject to X1 + X2 + d1- - d1+ = 300
maximize the total utility of the items carried.
X2 + d2- - d2+ = 200 Solve this knapsack problem using a computer.
X1 + d3- - d3+ = 100 (b) Suppose item number 3 is an extra battery pack,
which may be used with several of the other
(d) Maximize profit = 3X1 + 4X2
items. Tina has decided that she will only take
subject to X12 - 5X2 Ú 8 item number 5, a CD player, if she also takes item
3X1 + 4X2 Ú 12 number 3. On the other hand, if she takes item
number 3, she may or may not take item number
(e) Minimize cost = 18X1 + 5X2 + X22
5. Modify this problem to reflect this and solve
subject to 4X1 - 3X2 Ú 8 the new problem.
X1 + X2 Ú 18 10-12 Student Enterprises sells two sizes of wall posters,
Are any of these quadratic programming problems? a large 3- by 4-foot poster and a smaller 2- by 3-foot
poster. The profit earned from the sale of each large
Problems* poster is $3; each smaller poster earns $2. The firm,
although profitable, is not large; it consists of one art
10-10 Elizabeth Bailey is the owner and general manager of
student, Jan Meising, at the University of Kentucky.
Princess Brides, which provides a wedding planning
Because of her classroom schedule, Jan has the fol-
service in Southwest Louisiana. She uses radio adver-
lowing weekly constraints: (1) up to three large
tising to market her business. Two types of ads are
posters can be sold, (2) up to five smaller posters can
available—those during prime time hours and those at
be sold, (3) up to 10 hours can be spent on posters
other times. Each prime time ad costs $390 and
during the week, with each large poster requiring
reaches 8,200 people, while the offpeak ads each cost
2 hours of work and each small one taking 1 hour.
$240 and reach 5,100 people. Bailey has budgeted
With the semester almost over, Jan plans on taking a
$1,800 per week for advertising. Based on comments
three-month summer vacation to England and doesn’t
from her customers, Bailey wants to have at least 2
want to leave any unfinished posters behind. Find the
prime time ads and no more than 6 off peak ads.
integer solution that will maximize her profit.
(a) Formulate this as a linear program.
(b) Find a good or optimal integer solution part (a) 10-13 An airline owns an aging fleet of Boeing 737 jet air-
by rounding off or making an educated guess at planes. It is considering a major purchase of up to 17
the answer. new Boeing model 757 and 767 jets. The decision
(c) Solve this as an integer programming problem must take into account numerous cost and capability
using a computer. factors, including the following: (1) the airline can

Note: means the problem may be solved with QM for Windows; means the problem may be
solved with Excel; and means the problem may be solved with QM for Windows and/or Excel.
 DISCUSSION QUESTIONS AND PROBLEMS 421

finance up to $1.6 billion in purchases; (2) each Innis has identified eight potential locations to con-
Boeing 757 will cost $80 million, and each Boeing struct new single-family dwellings, but he cannot
767 will cost $110 million; (3) at least one-third of put up homes on all of the sites because he has only
the planes purchased should be the longer-range $300,000 to invest in all projects. The accompany-
757; (4) the annual maintenance budget is to be no ing table shows the cost of constructing homes in
more than $8 million; (5) the annual maintenance each area and the expected profit to be made from
cost per 757 is estimated to be $800,000, and it is the sale of each home. Note that the home-building
$500,000 for each 767 purchased; and (6) each 757 costs differ considerably due to lot costs, site prepa-
can carry 125,000 passengers per year, whereas each ration, and differences in the models to be built.
767 can fly 81,000 passengers annually. Formulate Note also that a fraction of a home cannot be built.
this as an integer programming problem to maxi-
mize the annual passenger-carrying capability. What
COST OF BUILDING EXPECTED PROFIT
category of integer programming problem is this? LOCATION AT THIS SITE ($) ($)
Solve this problem.
Clifton 60,000 5,000
10-14 Trapeze Investments is a venture capital firm that is
currently evaluating six different investment oppor- Mt. Auburn 50,000 6,000
tunities. There is not sufficient capital to invest in Mt. Adams 82,000 10,000
all of these, but more than one will be selected. A Amberly 103,000 12,000
0–1 integer programming model is planned to help
Norwood 50,000 8,000
determine which of the six opportunities to choose.
Variables X1, X2, X3, X4, X5, and X6 represent the Covington 41,000 3,000
six choices. For each of the following situations, Roselawn 80,000 9,000
write a constraint (or several constraints) that would Eden Park 69,000 10,000
be used.
(a) At least 3 of these choices are to be selected.
(b) Either investment 1 or investment 4 must be (a) Formulate Innis’s problem using 0–1 integer
undertaken, but not both. programming.
(c) If investment 4 is selected, then investment (b) Solve with QM for Windows or Excel.
6 must also be selected. However, if investment 10-17 A real estate developer is considering three possible
4 is not selected, it is still possible to select num- projects: a small apartment complex, a small shop-
ber 6. ping center, and a mini-warehouse. Each of these re-
(d) Investment 5 cannot be selected unless both in- quires different funding over the next two years, and
vestments 2 and 3 are also selected. the net present value of the investments also varies.
(e) Investment 5 must be selected if both invest- The following table provides the required invest-
ments 2 and 3 are also selected. ment amounts (in $1,000s) and the net present value
10-15 Horizon Wireless, a cellular telephone company, is (NPV) of each (also expressed in $1,000s):
expanding into a new era. Relay towers are neces-
sary to provide wireless telephone coverage to the INVESTMENT
different areas of the city. A grid is superimposed on
a map of the city to help determine where the towers NPV YEAR 1 YEAR 2
should be located. The grid consists of 8 areas Apartment 18 40 30
labeled A through H. Six possible tower locations Shopping center 15 30 20
(numbered 1–6) have been identified, and each loca- Mini-warehouse 14 20 20
tion could serve several areas. The table below indi-
cates the areas served by each of the towers.
The company has $80,000 to invest in year 1 and
$50,000 to invest in year 2.
TOWER
(a) Develop an integer programming model to max-
LOCATION 1 2 3 4 5 6
imize the NPV in this situation.
AREAS (b) Solve the problem in part (a) using computer
SERVED A, B, D B, C, G C, D, E, F E, F, H E, G, H A, D, F
software. Which of the three projects would be
undertaken if NPV is maximized? How much
Formulate this as a 0–1 programming model to min- money would be used each year?
imize the total number of towers required to cover 10-18 Refer to the real estate investment situation in Prob-
all the areas. Solve this using a computer. lem 10-17.
10-16 Innis Construction Company specializes in building (a) Suppose that the shopping center and the apart-
moderately priced homes in Cincinnati, Ohio. Tom ment would be on adjacent properties, and the
422 CHAPTER 10 • INTEGER PROGRAMMING, GOAL PROGRAMMING, AND NONLINEAR PROGRAMMING

shopping center would only be considered if the 10-21 Geraldine Shawhan is president of Shawhan File
apartment were also built. Formulate the con- Works, a firm that manufactures two types of metal
straint that would stipulate this. file cabinets. The demand for her two-drawer model
(b) Formulate a constraint that would force exactly is up to 600 cabinets per week; demand for a three-
two of the three projects to be undertaken. drawer cabinet is limited to 400 per week. Shawhan
10-19 Triangle Utilities provides electricity for three cities. File Works has a weekly operating capacity of 1,300
The company has four electric generators that are hours, with the two-drawer cabinet taking 1 hour
used to provide electricity. The main generator oper- to produce and the three-drawer cabinet requiring
ates 24 hours per day, with an occasional shutdown 2 hours. Each two-drawer model sold yields a $10
for routine maintenance. Three other generators profit, and the profit for the large model is $15.
(1, 2, and 3) are available to provide additional Shawhan has listed the following goals in order of
power when needed. A startup cost is incurred each importance:
time one of these generators is started. The startup 1. Attain a profit as close to $11,000 as possi-
costs are $6,000 for 1, $5,000 for 2, and $4,000 for ble each week.
3. These generators are used in the following ways: 2. Avoid underutilization of the firm’s produc-
A generator may be started at 6:00 A.M. and run for tion capacity.
either 8 hours or 16 hours, or it may be started at 3. Sell as many two- and three-drawer cabinets
2:00 P.M. and run for 8 hours (until 10:00 P.M.). All as the demand indicates.
generators except the main generator are shut down Set this up as a goal programming problem.
at 10:00 P.M. Forecasts indicate the need for 3,200
10-22 Solve Problem 10-21. Are any goals unachieved in
megawatts more than provided by the main genera-
this solution? Explain.
tor before 2:00 P.M., and this need goes up to 5,700
megawatts between 2:00 and 10:00 P.M. Generator 1 10-23 Hilliard Electronics produces specially coded com-
may provide up to 2,400 megawatts, generator 2 puter chips for laser surgery in 64MB, 256MB, and
may provide up to 2,100 megawatts, and generator 3 512MB sizes. (1MB means that the chip holds 1 mil-
may provide up to 3,300 megawatts. The cost per lion bytes of information.) To produce a 64MB chip
megawatt used per eight hour period is $8 for 1, $9 requires 8 hours of labor, a 256MB chip takes 13
for 2, and $7 for 3. hours, and a 512MB chip requires 16 hours.
(a) Formulate this problem as an integer program- Hilliard’s monthly production capacity is 1,200
ming problem to determine the least-cost way to hours. Mr. Blank, the firm’s sales manager, esti-
meet the needs of the area. mates that the maximum monthly sales of the
(b) Solve using computer software. 64MB, 256MB, and 512MB chips are 40, 50, and
60, respectively. The company has the following
10-20 The campaign manager for a politician who is run- goals (ranked in order from most important to least
ning for reelection to a political office is planning important):
the campaign. Four ways to advertise have been se-
lected: TV ads, radio ads, billboards, and newspaper 1. Fill an order from the best customer for thirty
ads. The cost of these are $900 for each TV ad, $500 64MB chips and thirty-five 256MB chips.
for each radio ad, $600 for a billboard for one 2. Provide sufficient chips to at least equal the
month, and $180 for each newspaper ad. The audi- sales estimates set by Mr. Blank.
ence reached by each type of advertising has been 3. Avoid underutilization of the production
estimated to be 40,000 for each TV ad, 32,000 for capacity.
each radio ad, 34,000 for each billboard, and 17,000 Formulate this problem using goal programming.
for each newspaper ad. The total monthly advertis- 10-24 An Oklahoma manufacturer makes two products:
ing budget is $16,000. The following goals have speaker telephones 1X12 and pushbutton telephones
been established and ranked: 1X22. The following goal programming model has
1. The number of people reached should be at been formulated to find the number of each to pro-
least 1,500,000. duce each day to meet the firm’s goals:
2. The total monthly advertising budget should Minimize P1d1- + P2d2- + P3d3+ + P4d1+
not be exceeded.
subject to 2X1 + 4X2 + d1- - d1+ = 80
3. Together, the number of ads on either TV or
radio should be at least 6. 8X1 + 10X2 + d2- - d2+ = 320
4. No more than 10 ads of any one type of ad- 8X1 + 6X2 + d3- - d3+ = 240
vertising should be used. all Xi, di Ú 0
(a) Formulate this as a goal programming problem.
Find the optimal solution using a computer.
(b) Solve this using computer software.
(c) Which goals are completely met and which of 10-25 Major Bill Bligh, director of the Army War Col-
them are not? lege’s new 6-month attaché training program, is
 DISCUSSION QUESTIONS AND PROBLEMS 423

concerned about how the 20 officers taking the production-planning week Motorcross has 40 hours
course spend their precious time while in his charge. available in its final testing bay. Each XJ6 requires 1
Major Bligh recognizes that there are 168 hours per hour to test and each XJ8 takes 2 hours. The revenue
week and thinks that his students have been using (in $1,000s) for the firm is nonlinear and is stated as
them rather inefficiently. Bligh lets (Number of XJ6s)(4 # 0.1 number of XJ6s) +
(Number of XJ8s)(5 # 0.2 number of XJ8s).
X1 = number of hours of sleep needed per week
(a) Formulate this problem.
X2 = number of personal hours 1eating, personal (b) Solve using Excel.
hygiene, handling laundary, and so on2 10-29 During the busiest season of the year, Green-Gro
X3 = number of hours of class and studying Fertilizer produces two types of fertilizers. The stan-
X4 = number of hours of social time off base dard type (X) is just fertilizer, and the other type (Y)
1dating, sports, family visits, and so on2 is a special fertilizer and weed-killer combination.
The following model has been developed to deter-
He thinks that students should study 30 hours a week mine how much of each type should be produced to
to have time to absorb material. This is his most im- maximize profit subject to a labor constraint:
portant goal. Bligh feels that students need at most 7
hours sleep per night on average and that this goal is Maximize profit = 12X - 0.04X2 + 15Y - 0.06Y 2
number 2. He believes that goal number 3 is to pro- subject to 2X + 4Y … 160 hours
vide at least 20 hours per week of social time. X, Y Ú 0
(a) Formulate this as a goal programming problem.
(b) Solve the problem using computer software. Find the optimal solution to this problem.
10-26 Mick Garcia, a certified financial planner (CFP) has 10-30 Pat McCormack, a financial advisor for Investors R
been asked by a client to invest $250,000. This money Us, is evaluating two stocks in a particular industry.
may be placed in stocks, bonds, or a mutual fund in He wants to minimize the variance of a portfolio
real estate. The expected return on investment is 13% consisting of these two stocks, but he wants to have
for stocks, 8% for bonds, and 10% for real estate. an expected return of at least 9%. After obtaining
While the client would like a very high expected re- historical data on the variance and returns, he devel-
turn, she would be satisfied with a 10% expected re- ops the following nonlinear program:
turn on her money. Due to risk considerations, several Minimize portfolio variance = 0.16X2 + 0.2XY + 0.09Y 2
goals have been established to keep the risk at an ac-
subject to X + Y = 1 1all funds must
ceptable level. One goal is to put at least 30% of the
be invested2
money in bonds. Another goal is that the amount of
money in real estate should not exceed 50% of the 0.11X + 0.08Y Ú 0.09 1return on the
money invested in stocks and bonds combined. In ad- investment2
dition to these goals, there is one absolute restriction. X, Y Ú 0
Under no circumstances should more than $150,000
where
be invested in any one area.
(a) Formulate this as a goal programming problem. X = proportion of money invested in stock 1
Assume that all of the goals are equally important. Y = proportion of money invested in stock 2
(b) Use any available software to solve this problem. Solve this using Excel and determine how much to in-
How much money should be put in each of the vest in each of the two stocks. What is the return for
investment options? What is the total return? this portfolio? What is the variance of this portfolio?
Which of the goals are not met? 10-31 Summertime Tees sells two very popular styles of
10-27 Hinkel Rotary Engine, Ltd., produces four- and six- embroidered shirts in southern Florida: a tank top
cylinder models of automobile engines. The firm’s and a regular T-shirt. The cost of the tank top is $6,
profit for each four-cylinder engine sold during its and the cost of the T-shirt is $8. The demand for
quarterly production cycle is $1,800 - $50X1, these is sensitive to the price, and historical data in-
where X1 is the number sold. Hinkel makes dicate that the weekly demands are given by
$2,400 - $70X2 for each of the larger engines sold,
with X2 equal to the number of six-cylinder engines X1 = 500 - 12P1
sold. There are 5,000 hours of production time avail- X2 = 400 - 15P2
able during each production cycle. A four-cylinder where
engine requires 100 hours of production time,
X1 = demand for tank top
whereas six-cylinder engines take 130 hours to man-
ufacture. Formulate this production planning prob- P1 = price for tank top
lem for Hinkel.
X1 = demand for regular T-shirt
10-28 Motorcross of Wisconsin produces two models of
snowmobiles, the XJ6 and the XJ8. In any given P2 = price for regular T-shirt
424 CHAPTER 10 • INTEGER PROGRAMMING, GOAL PROGRAMMING, AND NONLINEAR PROGRAMMING

(a) Develop an equation for the total profit. (c) The expected return at Apopka was overesti-
(b) Use Excel to find the optimal solution to the fol- mated. The correct value is $160,000 per year
lowing nonlinear program. Use the profit func- (that is, 160). Using the original assumptions
tion developed in part (a). (namely, ignoring (b)), does your answer to part
(a) change?
Maximize profit
10-33 In Solved Problem 10-3, nonlinear programming
subject to X1 = 500 - 12P1 was used to find the best value for the smoothing
X2 = 400 - 15P2 constant, ', in an exponential smoothing forecasting
P1 … 20 problem. To see how much the MAD can vary due to
P2 … 25 the selection of the smoothing constant, use Excel
and the data in Program 10.13A to find the value of
X1, P1, X2, P2 Ú 0 the smoothing constant that would maximize the
10-32 The integer programming problem in the box below MAD. Compare this MAD to the minimum MAD
has been developed to help First National Bank found in the solved problem.
decide where, out of 10 possible sites, to locate four 10-34 Using the data in Solved Problem 10-3, develop a
new branch offices: spreadsheet for a 2-period weighted moving average
where Xi represents Winter Park, Maitland, Osceola, forecast with weights of 0.6 (w1) for the most recent
Downtown, South Orlando, Airport, Winter Garden, period and 0.4 (w2) for the other period. Note these
Apopka, Lake Mary, Cocoa Beach, for i equals 1 to weights sum to 1, so the forecast is simply
10, respectively. Forecast for next period = w11Value in
(a) Where should the four new sites be located, and current period2 + w21Value in last period2
what will be the expected return? Find the weights for this 2-period weighted moving
(b) If at least one new branch must be opened in Mait- average that would minimize the MAD. (Hint: The
land or Osceola, will this change the answers? weights must sum to 1.)
Add the new constraint and rerun.

IP for Problem 10-32

Maximize expected returns = 120X1 + 100X2 + 110X3 + 140X4 + 155X5 + 128X6 + 145X7 + 190X8 + 170X9 + 150X10
subject to
20X1 + 30X2 + 20X3 + 25X4 + 30X5 + 30X6 + 25X7 + 20X8 + 25X9 + 30X10 … 110
15X1 + 5X2 + 20X3 + 20X4 + 5X5 + 5X6 + 10X7 + 20X8 + 5X9 + 20X10 … 50
X2 + X6 + X7 + X9 + X10 … 3
X2 + X3 + X5 + X8 + X9 Ú 2
X1 + X3 + X10 Ú 1
X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 + X10 … 4
all Xi = 0 or 1

INTERNET HOMEWORK PROBLEMS

See our Internet home page, at www.pearsonhighered.com/render, for additional home-
work problems 10-35 to 10-40.
 CASE STUDY 425

Case Study
Schank Marketing Research
Schank Marketing Research has just signed contracts to con- feels that it is twice as important to assign a project manager im-
duct studies for four clients. At present, three project managers mediately to Hines’s task as it is to provide one to General
are free for assignment to the tasks. Although all are capable of Foundry, a brand-new client. Schank wants to minimize the to-
handling each assignment, the times and costs to complete the tal costs of all projects while considering each of these goals.
studies depend on the experience and knowledge of each man- He feels that all of these goals are important, but if he had to
ager. Using his judgment, John Schank, the president, has been rank them, he would put his concern about NASA first, his
able to establish a cost for each possible assignment. These worry about Gardener second, his need to keep Hines Corpora-
costs, which are really the salaries each manager would draw tion happy third, his promise to Ruth fourth, and his concern
on each task, are summarized in the following table. about minimizing all costs last.
Schank is very hesitant about neglecting NASA, which has Each project manager can handle, at most, one new client.
been an important customer in the past. (NASA has employed
the firm to study the public’s attitude toward the Space Shuttle Discussion Questions
and proposed Space Station.) In addition, Schank has promised
1. If Schank were not concerned about noncost goals, how
to try to provide Ruth a salary of at least $3,000 on his next as-
would he formulate this problem so that it could be solved
signment. From previous contracts, Schank also knows that
quantitatively?
Gardener does not get along well with the management at CBT
2. Develop a formulation that will incorporate all five
Television, so he hopes to avoid assigning her to CBT. Finally,
objectives.
as Hines Corporation is also an old and valued client, Schank

CLIENT
PROJECT MANAGER HINES CORP. NASA GENERAL FOUNDRY CBT TELEVISION
Gardener $3,200 $3,000 $2,800 $2,900
Ruth 2,700 3,200 3,000 3,100
Hardgraves 1,900 2,100 3,300 2,100

Case Study
Oakton River Bridge
The Oakton River had long been considered an impediment from the state prevailed upon the federal government to fund
to the development of a certain medium-sized metropolitan a major portion of a new toll bridge over the Oakton River
area in the southeast. Lying to the east of the city, the river and the state legislature appropriated the rest of the needed
made it difficult for people living on its eastern bank to com- monies for the project.
mute to jobs in and around the city and to take advantage of Progress in construction of the bridge has been in accor-
the shopping and cultural attractions that the city had to offer. dance with what was anticipated at the start of construction.
Similarly, the river inhibited those on its western bank from The state highway commission, which will have operational
access to the ocean resorts lying one hour to the east. The jurisdiction over the bridge, has concluded that the opening of
bridge over the Oakton River had been built prior to World the bridge for traffic is likely to take place at the beginning of
War II and was grossly inadequate to handle the existing traf- the next summer, as scheduled. A personnel task force has
fic, much less the increased traffic that would accompany the been established to recruit, train, and schedule the workers
forecasted growth in the area. A congressional delegation needed to operate the toll facility.
426 CHAPTER 10 • INTEGER PROGRAMMING, GOAL PROGRAMMING, AND NONLINEAR PROGRAMMING

The personnel task force is well aware of the budgetary Minimum Number of Toll Collectors Required per Shift
problems facing the state. They have taken as part of their man-
date the requirement that personnel costs be kept as low as pos- SHIFT SUN. MON. TUE. WED. THU. FRI. SAT.
sible. One particular area of concern is the number of toll A 8 13 12 12 13 13 15
collectors that will be needed. The bridge is scheduling three B 10 10 10 10 10 13 15
shifts of collectors: shift A from midnight to 8 A.M., shift B
C 15 13 13 12 12 13 8
from 8 A.M. to 4 P.M., and shift C from 4 P.M. to midnight.
Recently, the state employees union negotiated a contract with
the state which requires that all toll collectors be permanent,
full-time employees. In addition, all collectors must work a The numbers in the table include one or two extra collec-
five-on, two-off schedule on the same shift. Thus, for example, tors per shift to fill in for collectors who call in sick and to pro-
a worker could be assigned to work Tuesday, Wednesday, vide relief for collectors on their scheduled breaks. Note that
Thursday, Friday, and Saturday on shift A, followed by Sunday each of the eight collectors needed for shift A on Sunday, for
and Monday off. An employee could not be scheduled to work, example, could have come from any of the A shifts scheduled
say, Tuesday on shift A followed by Wednesday, Thursday, Fri- to begin on Wednesday, Thursday, Friday, Saturday, or Sunday.
day, and Saturday on shift B or on any other mixture of shifts
during a 5-day block. The employees would choose their as- Discussion Questions
signments in order of their seniority.
The task force has received projections of traffic flow on 1. Determine the minimum number of toll collectors that
the bridge by day and hour. These projections are based on ex- would have to be hired to meet the requirements ex-
trapolations of existing traffic patterns—the pattern of commut- pressed in the table.
ing, shopping, and beach traffic currently experienced with 2. The union had indicated that it might lift its opposition to
growth projections factored in. Standards data from other state- the mixing of shifts in a 5-day block in exchange for addi-
operated toll facilities have allowed the task force to convert tional compensation and benefits. By how much could the
these traffic flows into toll collector requirements, that is, the number of toll collectors required be reduced if this is done?
minimum number of collectors required per shift, per day, to
Source: Based on B. Render, R. M. Stair, and I. Greenberg. Cases and Readings in
handle the anticipated traffic load. These toll collector require- Management Science, 2nd ed., 1990, pp. 55–56. Reprinted by permission of Pren-
ments are summarized in the following table: tice Hall, Upper Saddle River, NJ.

Bibliography
Aköz, Onur, and Dobrila Petrovic. “A Fuzzy Goal Programming Method with Drees, Lawrence David, and Wilbert E. Wilhelm. “Scheduling
Imprecise Goal Hierarchy,” European Journal of Operational Research Experiments on a Nuclear Reactor Using Mixed Integer Programming,”
181, 3 (September 2007): 1427–1433. Computers and Operations Research 28, 10 (September 2001):
Araz, Ceyhun, Pinar Mizrak Ozfirat, and Irem Ozkarahan. “An Integrated 1013–1037.
Multicriteria Decision-Making Methodology for Outsourcing Manage- Hueter, Jackie, and William Swart. “An Integrated Labor-Management System
ment,” Computers & Operations Research 34, 12 (December 2007): for Taco Bell,” Interfaces 28, 1 (January–February 1998): 75–91.
3738–3756. Ignizio, James P. Introduction to Linear Goal Programming. Beverly Hills:
Bertsimas, Dimitris, Christopher Darnell, and Robert Soucy. “Portfolio Con- Sage Publications, 1985.
struction through Mixed-Integer Programming at Grantham, Mayo, Van Katok, Elena, and Dennis Ott. “Using Mixed-Integer Programming to Reduce
Otterloo and Company,” Interfaces 29, 1 (January 1999): 49–66. Label Changes in the Coors Aluminum Can Plant,” Interfaces 30, 2
Chang, Ching-Ter. “Multi-Choice Goal Programming,” Omega 25, 4 (August (March 2000): 1–12.
2007): 389–396. Land, Alisa, and Susan Powell. “A Survey of the Operational Use
Charnes, A., and W. W. Cooper. Management Models and Industrial Applica- of ILP Models,” Annals of Operations Research 149, 1 (2007):
tions of Linear Programming, Vols. I and II. New York: John Wiley & 147–156.
Sons, Inc.1961. Lee, Sang M., and Marc J. Schniederjans. “A Multicriterial Assignment Prob-
Dawid, Herbert, Johannes Konig, and Christine Strauss. “An Enhanced Ros- lem: A Goal Programming Approach, ”Interfaces 13, 4 (August 1983):
tering Model for Airline Crews,” Computers and Operations Research 75–79.
28, 7 (June 2001): 671–688.
 BIBLIOGRAPHY 427

Pati, Rupesh Kumar, Prem Vrat, and Pradeep Kumar. “A Goal Programming Wadhwa, Vijay, and A. Ravi Ravindran. “Vendor Selection in Outsourcing,”
Model for a Paper Recycling System,” Omega 36, 3 (June 2008): Computers & Operations Research 34, 12 (December 2007):
405–417. 3725–3737.
Reyes, Pedro M., and Gregory V. Frazier. “Goal Programming Model for Gro- Zangwill, W. I. Nonlinear Programming: A Unified Approach. Upper Saddle
cery Shelf Space Allocation,” European Journal of Operational River, NJ: Prentice Hall, 1969.
Research 181, 2 (September 2007): 634–644.