Final PDF to printer

Chapter Three

Linear Programming:
Formulation and
Applications
Learning Objectives
After completing this chapter, you should be able to

1. Recognize various kinds of managerial problems to which linear programming can be applied.
2. Describe the five major categories of linear programming problems, including their identify-
ing features.
3. Formulate a linear programming model from a description of a problem in any of these
categories.
4. Describe the difference between resource constraints and benefit constraints, including the
difference in how they arise.
5. Describe fixed-requirement constraints and where they arise.
6. Identify the kinds of Excel functions that linear programming spreadsheet models use for
the output cells, including the objective cell.
7. Identify the four components of any linear programming model and the kind of spreadsheet
cells used for each component.
8. Understand the flexibility that managers have in prescribing key considerations that can be
incorporated into a linear programming model.

Chapter 2 introduced the basic concepts of linear programming, but only considered two tiny
examples (the Wyndor case study introduced in Section 2.1 and the Profit & Gambit problem
considered in Section 2.7) that only begin to illustrate the great power and versatility of this
technique. This chapter will expand considerably further on the great application potential of
linear programming by describing and illustrating several different types of linear program-
ming problems. The many examples also will be supplemented by five solved problems pre-
sented at the end of the chapter.
Linear programming problems come in many guises. And their models take various forms.
This diversity can be confusing to both students and managers, making it difficult to recognize
when linear programming can be applied to address a managerial problem. Since managers
instigate management science studies, the ability to recognize the applicability of linear pro-
gramming is an important managerial skill. This chapter focuses largely on developing this skill.
The usual textbook approach to trying to teach this skill is to present a series of diverse
examples of linear programming applications. The weakness of this approach is that it empha-
sizes differences rather than the common threads between these applications. Our approach
will be to emphasize these common threads—the identifying features—that tie together lin-
ear programming problems even when they arise in very different contexts. We will describe
some broad categories of linear programming problems and the identifying features that char-
acterize them. Then we will use diverse examples, but with the purpose of illustrating and
emphasizing the common threads among them.
64

hiL18920_ch03_064-123.indd 64 12/08/17 06:54 AM

 Final PDF to printer

3.1  A Case Study: The Super Grain Corp. Advertising-Mix Problem   65

We will focus on the following five key categories of linear programming problems:
• resource-allocation problems
• cost-benefit-trade-off problems
• mixed problems
• transportation problems
• assignment problems
In each case, an important identifying feature is the nature of the restrictions on what deci-
sions can be made, and thus the nature of the resulting functional constraints in the linear
programming model. For each category, you will see how the basic data for a problem lead
directly to a linear programming model with a certain distinctive form. Thus, model formula-
tion becomes a by-product of proper problem formulation.
The chapter begins with a case study (Section 3.1) that will illustrate the systematic proce-
dure for formulating and solving the model for any linear programming problem. Section 3.2
will point out why the initial version of the case study turns out to be one example of a
resource-allocation problem. However, we then return to the case study in Section 3.4, where
additional managerial considerations turn the problem into a mixed problem.
Sections 3.2 to 3.6 focus on the five categories of linear programming problems in turn. (Sec-
tion 3.4 also briefly introduces still another category, fixed-requirements problems, where this
category provides the framework for the more specific categories considered in Sections 3.5
and 3.6.) Section 3.7 then takes a broader look at the formulation of linear programming
models from a managerial perspective. That section (along with Section 3.4) highlights the
importance of having the model accurately reflect the managerial view of the problem. These
(and other) sections also describe the flexibility available to managers for having the model
structured to best fit their view of the important considerations.

3.1 A CASE STUDY: THE SUPER GRAIN CORP. ADVERTISING-MIX PROBLEM

Claire Syverson, vice president for marketing of the Super Grain Corporation, is facing a
daunting challenge: how to break into an already overly crowded breakfast cereal market in
a big way. Fortunately, the company’s new breakfast cereal—Crunchy Start—has a lot going
for it: Great taste. Nutritious. Crunchy from start to finish. She can recite the litany in her
sleep now. It has the makings of a winning promotional campaign.
However, Claire knows that she has to avoid the mistakes she made in her last campaign for
a breakfast cereal. That had been her first big assignment since she won this promotion, and
what a disaster! She thought she had developed a really good campaign. But somehow it had
failed to connect with the most crucial segments of the market—young children and parents
of young children. She also has concluded that it was a mistake not to include cents-off cou-
pons in the magazine and newspaper advertising. Oh well. Live and learn.
But she had better get it right this time, especially after the big stumble last time. The com-
pany’s president, David Sloan, already has impressed on her how important the success of
Crunchy Start is to the future of the company. She remembers exactly how David concluded
the conversation. “The company’s shareholders are not happy. We need to get those earnings
headed in the right direction again.” Claire had heard this tune before, but she saw in David’s
eyes how deadly serious he was this time.
Claire often uses spreadsheets to help organize her planning. Her management science
course in business school impressed upon her how valuable spreadsheet modeling can be. She
regrets that she did not rely more heavily on spreadsheet modeling for the last campaign. That
was a mistake that she is determined not to repeat.
Now it is time for Claire to carefully review and formulate the problem in preparation for
formulating a spreadsheet model.

The Problem
Claire already has employed a leading advertising firm, Giacomi & Jackowitz, to help design
a nationwide promotional campaign that will achieve the largest possible exposure for Crunchy

hiL18920_ch03_064-123.indd 65 12/08/17 06:54 AM

 Final PDF to printer

66  Chapter Three   Linear Programming: Formulation and Applications

TABLE 3.1
Costs
Cost and Exposure Data
for the Super Grain Each TV
Corp. Advertising-Mix Cost Category Commercial Each Magazine Ad Each Sunday Ad
Problem
Ad budget $300,000 $150,000 $100,000
Planning budget 90,000 30,000 40,000
Expected number of exposures 1,300,000 600,000 500,000

Start. Super Grain will pay this firm a fee based on services performed (not to exceed $1 million)
and has allocated an additional $4 million for advertising expenses.
Giacomi & Jackowitz has identified the three most effective advertising media for this product:
Medium 1: Television commercials on Saturday morning programs for children.
Medium 2: Advertisements in food and family-oriented magazines.
Medium 3: Advertisements in Sunday supplements of major newspapers.
The problem now is to determine which levels should be chosen for these advertising activities
to obtain the most effective advertising mix.
To determine the best mix of activity levels for this particular advertising problem, it is
necessary (as always) to identify the overall measure of performance for the problem and
then the contribution of each activity toward this measure. An ultimate goal for Super Grain
is to maximize its profits, but it is difficult to make a direct connection between advertising
exposure and profits. Therefore, as a rough surrogate for profit, Claire decides to use expected
number of exposures as the overall measure of performance, where each viewing of an adver-
tisement by some individual counts as one exposure.
Giacomi & Jackowitz has made preliminary plans for advertisements in the three media.
The firm also has estimated the expected number of exposures for each advertisement in each
medium, as given in the bottom row of Table 3.1.
The number of advertisements that can be run in the different media are restricted by both
the advertising budget (a limit of $4 million) and the planning budget (a limit of $1 million
for the fee to Giacomi & Jackowitz). Another restriction is that there are only five commercial
spots available for running different commercials (one commercial per spot) on children’s
television programs Saturday morning (medium 1) during the time of the promotional cam-
paign. (The other two media have an ample number of spots available.)
Consequently, the three resources for this problem are:
Resource 1: Advertising budget ($4 million).
Resource 2: Planning budget ($1 million).
Resource 3: TV commercial spots available (5).
Table 3.1 shows how much of the advertising budget and the planning budget would be
used by each advertisement in the respective media.
• The first row gives the cost per advertisement in each medium.
• The second row shows Giacomi & Jackowitz’s estimates of its total cost (including over-
head and profit) for designing and developing each advertisement for the respective media.1
(This cost represents the billable fee from Super Grain.)
• The last row then gives the expected number of exposures per advertisement.
Analysis of the Problem
Claire decides to formulate and solve a linear programming model for this problem on a
spreadsheet. The formulation procedure summarized at the end of Section 2.2 guides this pro-
cess. Like any linear programming model, this model will have four components:
1. The data
2. The decisions
1
When presenting its estimates in this form, the firm is making two simplifying assumptions. One is that its
cost for designing and developing each additional advertisement in a medium is roughly the same as for the
first advertisement in that medium. The second is that its cost when working with one medium is unaffected
by how much work it is doing (if any) with the other media.

hiL18920_ch03_064-123.indd 66 12/08/17 06:54 AM

 Final PDF to printer

3.1  A Case Study: The Super Grain Corp. Advertising-Mix Problem   67

3. The constraints
4. The measure of performance
The spreadsheet needs to be formatted to provide the following kinds of cells for these
components:
Four kinds of cells are Data → data cells
needed for these four com-
Decisions → changing cells
ponents of a spreadsheet
model. Constraints → output cells
Measure of performance → objective cell
Figure 3.1 shows the spreadsheet model formulated by Claire. Let us see how she did this,
including why she structured the spreadsheet in this way, by considering each of the ­components
of the model individually.
FIGURE 3.1
The spreadsheet model for the Super Grain problem (Section 3.1), including the objective cell TotalExposures (H13) and the other
output cells BudgetSpent (F8:F9), as well as the specifications needed to set up Solver. The changing cells NumberOfAds (C13:E13)
show the optimal solution obtained by Solver.

A B C D E F G H
1 Super Grain Corp. Advertising-Mix Problem
2
3 TV Spots Magazine Ads SS Ads
4 Exposures per Ad 1,300 600 500
5 (thousands)
6 Budget Budget
7 Cost per Ad ($thousands) Spent Available
8 Ad Budget 300 150 100 4,000 ≤ 4,000
9 Planning Budget 90 30 40 1,000 ≤ 1,000
10
11 Total Exposures
12 TV Spots Magazine Ads SS Ads (thousands)
13 Number of Ads 0 20 10 17,000
14 ≤
15 Max TV Spots 5

Solver Parameters F
Set Objective Cell: TotalExposures
To: Max 6 Budget
By Changing Variable Cells: 7 Spent
NumberOfAds
Subject to the Constraints: 8 =SUMPRODUCT(C8:E8,NumberOfAds)
BudgetSpent <= BudgetAvailable 9 =SUMPRODUCT(C9:E9,NumberOfAds)
TVSpots <= MaxTVSpots
H
Solver Options:
Make Variables Nonnegative 11 Total Exposures
Solving Method: Simplex LP 12 (thousands)
13 =SUMPRODUCT(ExposuresPerAd,NumberOfAds)

Range Name Cells

BudgetAvailable H8: H9
BudgetSpent F8: F9
CostPerAd C8: E9
ExposuresPerAd C4: E4
MaxTVSpots C15
NumberOfAds C13: E13
TotalExposures H13
TVSpots C13

hiL18920_ch03_064-123.indd 67 12/08/17 06:54 AM

 Final PDF to printer

68  Chapter Three   Linear Programming: Formulation and Applications

The Data
It is best to begin structuring the spreadsheet by deciding where to locate the data cells. Most
of these cells normally are placed in the upper left-hand portion of the spreadsheet. The struc-
ture of the rest of the model will then follow the structure of the data cells. When some or
all of the data are being obtained from a well-constructed table, the clarity of the spread-
sheet usually is enhanced by retaining the format of the table when inserting the data into the
spreadsheet.
Claire followed these guidelines when she chose cells C4:E4 and C8:E9 to be the data cells
to hold the information in Table 3.1 when using units of thousands of dollars. These data cells
have been given the range names: ExposuresPerAd (C4:E4), CostPerAd (C8:E9), Budget-
Available (H8:H9), and MaxTVSpots (C15).
The other relevant kind of data here is the information given earlier about the amounts
available of the three resources for the problem (the advertising budget, the planning budget,
and the commercial spots available). Claire chose the locations of the data cells holding this
information (H8:H9 and C15) to fit with the constraints that will be inserted a little later.

The Decisions
The problem has been defined as determining the most effective advertising mix among the
three media selected by Giacomi & Jackowitz. Therefore, there are three decisions:
Decision 1: TV = Number of commercials for separate spots on television.
Decision 2:    M = Number of advertisements in magazines.
Decision 3:   SS = Number of advertisements in Sunday supplements.
The decisions already made for the locations of the data cells have designated columns C,
D, and E to be the columns for these media. Therefore, the changing cells to hold these three
decisions have been placed in row 13 in these columns:
​​T V → cell C13​  M → cell D13​  SS → cell E13​​
These changing cells are collectively referred to by the range name NumberOfAds (C13:E13).

The Constraints
These changing cells need to be nonnegative. In addition, constraints are needed for the three
resources. The first two resources are the ad budget and planning budget. Data for these two
budgets already have ben placed in data cells in rows 8 and 9, so the constraints involving
these budgets need to appear in these same two rows. The amounts spent from these two
budgets are shown in the range BudgetSpent (F8:F9) and the amounts available from these
budgets are shown in the range of data cells BudgetAvailable (H8:H9). As suggested by the ≤
signs entered into column G, the corresponding constraints are
Total spending on advertising ≤ 4,000   (Ad budget in $1,000s)
​​       ​  ​  ​ ​​
Total cost of planning ≤ 1,000 (Planning budget in $1,000s)
Using the data in columns C, D, and E for the resources, these totals are
Total spending on advertising = 300TV + 150M + 100SS
​​       ​  ​  ​  ​​​
Total cost of planning = 90TV + 30M + 40SS

These sums of products on the right-hand side are entered into the output cells BudgetSpent
(F8:F9) by using the SUMPRODUCT functions shown in the lower right-hand side of Figure
3.1. Although the ≤ signs entered in column G are only cosmetic (trial solutions still can be
Excel Tip: Range entered in the changing cells that violate these inequalities), they will serve as a reminder later
names may overlap. For to use these same ≤ signs when entering the constraints in Solver.
instance, we have used The third resource is TV spots for different commercials. Five such spots are available for
­NumberOfAds to refer to purchase. The number of spots used is one of the changing cells (C13). Since this cell will be
the whole range of chang-
ing cells, C13:E13, and
used in a constraint, we assign the cell its own range name: TVSpots (C13). The maximum
TVSpots to refer to the number of TV spots available is in the data cell MaxTVSpots (C15). Thus, the required con-
single cell, C13. straint is TVSpots ≤ MaxTVSpots.

hiL18920_ch03_064-123.indd 68 12/08/17 06:54 AM

 Final PDF to printer

3.1  A Case Study: The Super Grain Corp. Advertising-Mix Problem   69

The Measure of Performance

Claire Syverson is using expected number of exposures as the overall measure of perfor-
mance, so let
​Exposure = Expected number of exposures (in thousands) from all the advertising​
The data cells ExposuresPerAd (C4:E4) provide the expected number of exposures (in
­thousands) per advertisement in the respective media and the changing cells NumberOfAds
(C13:E13) give the number of each type of advertisement. Therefore,
Exposure = 1,300TV + 600M + 500SS
​​ ​    ​  ​  ​ ​​
​ = SUMPRODUCT (ExposuresPerAd, NumberOfAds)
is the formula that needs to be entered into the objective cell, TotalExposures (H13).
Summary of the Formulation
The above analysis of the four components of the model has formulated the following linear
programming model (in algebraic form) on the spreadsheet:
​Maximize   Exposure = 1,300TV + 600M + 500SS​
subject to
Ad spending: 300 TV + 150 M + 100 SS ≤ 4,000
​​ Planning
      costs:​ 
​  ​  90 TV ​  + 30 M + 40 SS​  ​  ≤​ ​  1,000​​​​
Number of television spots: TV ≤ 5
and
​​T V ≥ 0​  M ≥ 0​  SS ≥ 0​​
The difficult work of defining the problem and gathering all the relevant data, including the
information in Table 3.1, leads directly to this formulation.
Solving the Model
Excel Tip: With Excel’s To solve the spreadsheet model formulated above, some key information needs to be entered
Solver, the Solver dialog into Solver. The lower left-hand side of Figure 3.1 shows the needed entries: the objective cell
box is used to tell Solver the
(TotalExposures), the changing cells (NumberOfAds), the goal of maximizing the objective
location on the spreadsheet
of several of the elements cell, and the constraints BudgetSpent ≤ BudgetAvailable and TVSpots ≤ MaxTVSpots. Two
of the model: the changing options are also specified at the bottom of the Solver Parameters box on the lower left-hand
cells, the objective cell, and side of Figure 3.1. The changing cells need nonnegativity constraints because negative values of
the constraints. With Ana- advertising are not possible. Choose the Simplex LP (Excel’s Solver) or Standard LP/Quadratic
lytic Solver, the Decisions,
Engine (Analytic Solver) solving method, because this is a linear programming model. Running
Constraints, and Objec-
tive menu on the Analytic Solver then finds an optimal solution for the model and displays it in the changing cells.
Solver ribbon are used The optimal solution given in row 13 of the spreadsheet provides the following plan for the
along with the Model pane. promotional campaign:
Do not run any television commercials.
Run 20 advertisements in magazines.
Run 10 advertisements in Sunday supplements.
Since TotalExposures (H13) gives the expected number of exposures in thousands, this plan
would be expected to provide 17,000,000 exposures.
Evaluation of the Adequacy of the Model
When she chose to use a linear programming model to represent this advertising-mix prob-
lem, Claire recognized that this kind of model does not provide a perfect match to this prob-
lem. However, a mathematical model is intended to be only an approximate representation
of the real problem. Approximations and simplifying assumptions generally are required to
have a workable model. All that is really needed is that there be a reasonably high correlation
between the prediction of the model and what would actually happen in the real problem. The
t eam now needs to check whether this criterion is satisfied.

hiL18920_ch03_064-123.indd 69 12/08/17 06:54 AM

 Final PDF to printer

70  Chapter Three   Linear Programming: Formulation and Applications

Linear programming One assumption of linear programming is that fractional solutions are allowed. For the
models allow fractional current problem, this means that a fractional number (e.g., 3½) of television commercials
solutions.
(or of ads in magazines or Sunday supplements) should be allowed. This is technically true,
since a commercial can be aired for less than a normal run, or an ad can be run in just a frac-
tion of the usual magazines or Sunday supplements. However, one defect of the model is that
it assumes that Giacomi & Jackowitz’s cost for planning and developing a commercial or ad
that receives only a fraction of its usual run is only that fraction of its usual cost, even though
the actual cost would be the same as for a full run. Fortunately, the optimal solution obtained
was an integer solution (0 television commercials, 20 ads in magazines, and 10 ads in Sunday
supplements), so the assumption that fractional solutions are allowed was not even needed.
Although it is possible to have a fractional number of a normal run of commercials or
ads, a normal run tends to be much more effective than a fractional run. Therefore, it would
have been reasonable for Claire to drop the assumption that fractional solutions are allowed.
If Claire had done this and the optimal solution for the linear programming model had not
turned out to be integer, constraints can be added to require the changing cells to be integer.
(The TBA Airlines example in the next section provides an illustration of this type of con-
straint.) After adding such constraints, the model is called an integer programming model
instead of a linear programming model, but it still can be readily solved by Solver.
Linear programming mod- Another key assumption of linear programming is that the appropriate equation for each of
els should use SUM or the output cells, including the objective cell, is one that can be expressed as a SUMPRODUCT
SUMPRODUCT functions
of data cells and changing cells (or occasionally just a SUM of changing cells). For the objec-
for the output cells, includ-
ing the objective cell. tive cell (cell H13) in Figure 3.1, this implies that the expected number of exposures to be
obtained from each advertising medium is proportional to the number of advertisements in
that medium. This proportionality seems true, since each viewing of the advertisements by
some individual counts as another exposure. Another implication of using a SUMPRODUCT
function is that the expected number of exposures to be obtained from an advertising medium
is unaffected by the number of advertisements in the other media. Again, this implication
seems valid, since viewings of advertisements in different media count as separate exposures.
Although a SUMPRODUCT function is appropriate for calculating the expected number
of exposures, the choice of this number for the overall measure of performance is somewhat
questionable. Management’s real objective is to maximize the profit generated as a result of
the advertising campaign, but this is difficult to measure so expected number of exposures
was selected to be a surrogate for profit. This would be valid if profit were proportional to the
expected number of exposures. However, proportionality is only an approximation in this case
because too many exposures for the same individual reach a saturation level where the impact
(potential profit) from one more exposure is substantially less than for the first exposure.
(When proportionality is not a reasonable approximation, Chapter 8 will describe nonlinear
models that can be used instead.)
To check how reasonable it is to use expected number of exposures as a surrogate for
profit, Claire meets with Sid Jackowitz, one of the senior partners of Giacomi & Jackowitz.
Sid indicates that the contemplated promotional campaign (20 advertisements in magazines
and 10 in Sunday supplements) is a relatively modest one well below saturation levels. Most
readers will only notice these ads once or twice, and a second notice is very helpful for rein-
forcing the first one. Furthermore, the readership of magazines and Sunday supplements is
sufficiently different that the interaction of the advertising impact in these two media should
be small. Consequently, Claire concludes that using expected number of exposures for the
objective cell in Figure 3.1 provides a reasonable approximation. (A continuation of this case
study in Case 8-1 will delve into the more complicated analysis that is required in order to use
profit directly as the measure of performance to be recorded in the objective cell instead of
making this approximation.)
Next, Claire quizzes Sid about his firm’s costs for planning and developing advertisements
in these media. Is it reasonable to assume that the cost in a given medium is proportional
to the number of advertisements in that medium? Is it reasonable to assume that the cost of
developing advertisements in one medium would not be substantially reduced if the firm had
just finished developing advertisements in another medium that might have similar themes?
Sid acknowledges that there is some carryover in ad planning from one medium to another,

hiL18920_ch03_064-123.indd 70 12/08/17 06:54 AM

 Final PDF to printer

3.2  Resource-Allocation Problems   71

especially if both are print media (e.g., magazines and Sunday supplements), but that the
carryover is quite limited because of the distinct differences in these media. Furthermore,
he feels that the proportionality assumption is quite reasonable for any given medium since
the amount of work involved in planning and developing each additional advertisement in
the medium is nearly the same as for the first one in the medium. The total fee that Super
Grain will pay Giacomi & Jackowitz will eventually be based on a detailed accounting of the
amount of work done by the firm. Nevertheless, Sid feels that the cost estimates previously
provided by the firm (as entered in cells C9, D9, and E9 in units of thousands of dollars) give
a reasonable basis for roughly projecting what the fee will be for any given plan (the entries in
the changing cells) for the promotional campaign.
Based on this information, Claire concludes that using a SUMPRODUCT function for cell
F9 provides a reasonable approximation. Doing the same for cell F8 is clearly justified. Given
her earlier conclusions as well, Claire decides that the linear programming model incorpo-
rated into Figure 3.1 (plus any expansions of the model needed later for the detailed planning)
is a sufficiently accurate representation of the real advertising-mix problem. It will not be
necessary to refine the results from this model by turning next to a more complicated kind of
mathematical model (such as those to be described in Chapter 8).
Therefore, Claire sends a memorandum to the company’s president, David Sloan, describ-
ing a promotional campaign that corresponds to the optimal solution from the linear program-
ming model (no TV commercials, 20 ads in magazines, and 10 ads in Sunday supplements).
She also requests a meeting to evaluate this plan and discuss whether some modifications
should be made.
We will pick up this story again in Section 3.4.

Review 1. What is the problem being addressed in this case study?

Questions 2. What overall measure of performance is being used?
3. What are the assumptions of linear programming that need to be checked to evaluate the ade-
quacy of using a linear programming model to represent the problem under consideration?

3.2 RESOURCE-ALLOCATION PROBLEMS

In the opening paragraph of Chapter 2, we described managerial problems involving the allo-
cation of an organization’s resources to its various productive activities. Those were resource-
allocation problems.
Resource-allocation problems are linear programming problems involving the allocation of
resources to activities. The identifying feature for any such problem is that each functional con-
straint in the linear programming model is a resource constraint, which has the form
​Amount of resource used ≤ Amount of resource available​
for one of the resources.

The amount of a resource used depends on which activities are undertaken, the levels of
those activities, and how heavily those activities need to use the resource. Thus, the resource
constraints place limits on the levels of the activities. The objective is to choose the levels of
the activities so as to maximize some overall measure of performance (such as total profit)
from the activities while satisfying all the resource constraints.
Beginning with the Super Grain case study and then the Wyndor case study from ­Chapter 2,
we will look at four examples that illustrate the characteristics of resource-allocation prob-
lems. These examples also demonstrate how this type of problem can arise in a variety of
contexts.

The Super Grain Corp. Advertising-Mix Problem

The linear programming model formulated in Section 3.1 for the Super Grain case study is
one example of a resource-allocation problem. The three activities under consideration are the
advertising in the three types of media chosen by Giacomi & Jackowitz.

hiL18920_ch03_064-123.indd 71 12/08/17 06:54 AM

 Final PDF to printer

72  Chapter Three   Linear Programming: Formulation and Applications

Activity 1: TV commercials
Activity 2: Magazine ads
Activity 3: Sunday ads
The decisions being made are the levels of these activities, that is, the number of TV com-
mercials, magazine ads, and Sunday ads to run.
An initial step in formulat- The resources to be allocated to these activities are
ing any resource-allocation
problem is to identify the Resource 1: Advertising budget ($4 million).
activities and the resources. Resource 2: Planning budget ($1 million).
Resource 3: TV spots available for different commercials (5).
where the amounts available of these resources are given in parentheses. Thus, this problem
has three resource constraints:
1. Advertising budget used ≤ $4 million
2. Planning budget used ≤ $1 million
3. TV spots used ≤5
Rows 8–9 and cells C13:C15 in Figure 3.1 show these constraints in a spreadsheet. Cells
C8:E9 give the amount of the advertising budget and the planning budget used by each unit
of each activity, that is, the amount used by one TV spot, one magazine ad, and one Sunday
ad, respectively.
Cells C4, D4, and E4 on this spreadsheet give the contribution per unit of each activity to
the overall measure of performance (expected number of exposures).
Characteristics of Resource-Allocation Problems
For each proposed activ- Other resource-allocation problems have the same kinds of characteristics as the Super Grain
ity, a decision needs to be problem. In each case, there are activities where the decisions to be made are the levels of
made as to how much of
these activities. The contribution of each activity to the overall measure of performance is
the activity to do. In other
words, what should the proportional to the level of that activity. Commonly, this measure of performance is the total
level of the activity be? profit from the activities, but occasionally it is something else (as in the Super Grain problem).
Every problem of this type has a resource constraint for each resource. The amounts of the
resources used depend on the levels of the activities. For each resource, the amount used by
each activity is proportional to the level of that activity.
These three kinds of data The manager or management science team studying a resource allocation problem needs to
are needed for any resource- gather (with considerable help) three kinds of data:
allocation problem.
1. The amount available of each resource.
2. The amount of each resource needed by each activity. Specifically, for each combination of
resource and activity, the amount of the resource used per unit of the activity must be estimated.
3. The contribution per unit of each activity to the overall measure of performance.
Generally there is considerable work involved in developing these data. A substantial
amount of digging and consultation is needed to obtain the best estimates available in a timely
fashion. This step is critical. Well-informed estimates are needed to obtain a valid linear pro-
gramming model for guiding managerial decisions. The dangers involved in inaccurate esti-
mates are one reason why what-if analysis (Chapter 5) is such an important part of most linear
programming studies.

The Wyndor Glass Co. Product-Mix Problem

The product-mix problem facing the management of the Wyndor Glass Co. in Section 2.1 is
to determine the most profitable mix of production rates for the two new products, consider-
ing the limited availability of spare production capacity in the company’s three plants. This is
a resource-allocation problem.
The activities under consideration are
Activity 1: Produce the special new doors.
Activity 2: Produce the special new windows.

hiL18920_ch03_064-123.indd 72 12/08/17 06:54 AM

 Final PDF to printer

3.2  Resource-Allocation Problems   73

The decisions being made are the levels of these activities, that is, the production rates for the
doors and windows. Production rate is being measured as the number of units (doors or win-
dows) produced per week. Management’s objective is to maximize the total profit generated
by the two new products, so the overall measure of performance is total profit. The contribu-
tion of each product to profit is proportional to the production rate for that product.
The resources to be allocated to these activities are
Resource 1: Production capacity in Plant 1.
Resource 2: Production capacity in Plant 2.
Resource 3: Production capacity in Plant 3.
Each of the three functional constraints in the linear programming model formulated in
Section 2.2 (see rows 7–9 of the spreadsheet in Figure 2.3 or 2.4) is a resource constraint for
one of these three resources. Column E shows the amount of production capacity used in each
plant and column G gives the amount available.
Table 2.1 in Section 2.1 provides the data for the Wyndor problem. You already have seen
how the numbers in Table 2.1 become the parameters in the linear programming model in
either its spreadsheet formulation (Section 2.2) or its algebraic form (Section 2.3).

The TBA Airlines Problem

TBA Airlines is a small regional company that specializes in short flights in small passenger air-
planes. The company has been doing well and management has decided to expand its operations.
The Problem
The basic issue facing management now is whether to purchase more small airplanes to add
some new short flights or to start moving into the national market by purchasing some large
airplanes for new cross-country flights (or both). Many factors will go into management’s
final decision, but the most important one is which strategy is likely to be most profitable.
The first row of Table 3.2 shows the estimated net annual profit (inclusive of capital recov-
ery costs) from each type of airplane purchased. The second row gives the purchase cost per
airplane and also notes that the total amount of capital available for airplane purchases is
$250 million. The third row records the fact that management does not want to purchase more
than five small airplanes because of limited possibilities for adding lucrative short flights,
whereas they have not specified a maximum number for large airplanes (other than that
imposed by the limited capital available).
How many airplanes of each type should be purchased to maximize the total net annual profit?
Formulation
This is a resource-allocation problem. The activities under consideration are
Activity 1: Purchase small airplanes.
Activity 2: Purchase large airplanes.
The decisions to be made are the levels of these activities, that is,
S = Number of small airplanes to purchase
L = Number of large airplanes to purchase
The one resource to be allocated to these activities is
Resource: Investment capital ($250 million).
Thus, there is a single resource constraint:
​Investment capital spent ≤ $250 million​
TABLE 3.2
Small Airplane Large Airplane Capital Available
Data for the TBA
Airlines Problem Net annual profit per airplane $7 million $22 million
Purchase cost per airplane $25 million $75 million $250 million
Maximum purchase quantity 5 No maximum

hiL18920_ch03_064-123.indd 73 12/08/17 06:54 AM

 Final PDF to printer

74  Chapter Three   Linear Programming: Formulation and Applications

FIGURE 3.2
A spreadsheet model for the TBA Airlines integer programming problem where the changing cells, UnitsProduced (C12:D12), show the
optimal airplane purchases obtained by Solver, and the objective cell, TotalProfit (G12), gives the resulting total profit in millions of dollars.

A B C D E F G
1 TBA Airlines Airplane Purchasing Problem
2
3 Small Airplane Large Airplane
4 Unit Profit ($millions) 7 22
5
6 Capital Capital
7 Capital per Unit Purchased Spent Available
8 Capital ($millions) 25 75 250 <= 250
9
10 Total Profit
11 Small Airplane Large Airplane ($millions)
12 Number Purchased 1 3 73
13 <=
14 Maximum Small Airplanes 5

E
6 Capital
7 Spent
8 =SUMPRODUCT(CapitalPerUnitPurchased,NumberPurchased)

Solver Parameters G
Set Objective Cell: TotalProfit 10 Total Profit
To: Max 11 ($millions)
By Changing Variable Cells: 12 =SUMPRODUCT(UnitProfit,NumberPurchased)
NumberPurchased
Subject to the Constraints: Range Name Cells
CapitalSpent <= CapitalAvailable Capital Available G8
NumberPurchased = integer CapitalPerUnitPurchased C8:D8
SmallAirplanes <= MaxSmallAirplanes CapitalSpent E8
MaxSmallAirplanes C14
Solver Options:
NumberPurchased C12:D12
Make Variables Nonnegative
SmallAirplanes C12
Solving Method: Simplex LP
TotalProfit G12
UnitProfit C4:D4

In addition, management has specified one side constraint:

​Number of small airplanes purchased ≤ 5​
Figure 3.2 shows the formulation of a spreadsheet model for this problem, where the data
in Table 3.2 have been transferred into the data cells—UnitProfit (C4:D4), CapitalPerUnit-
Purchased (C8:D8), CapitalAvailable (G8), and MaxSmallAirplanes (C14). The resource con-
straint then appears in cells C8:G8 while C12:C14 shows the side constraint. The objective for
this problem is to maximize the total net annual profit, so the equation for the objective cell is
​TotalProfit ​​(​​G12​)​​​ = SUMPRODUCT (​​ ​​UnitProfit, UnitsPurchased​)​​​​
Since the TBA Airlines problem is a resource-allocation problem, this spreadsheet model
has essentially the same form as the Super Grain and Wyndor problems except for one small
difference. The changing cells in this case must have integer values since it is not feasible for
the company to purchase and operate a fraction of an airplane. Therefore, constraints that the
changing cells need to be integer are added. With Excel’s Solver, use the Add Constraint dia-
log box to choose the range of these cells (C12:D12) as the left-hand side and then choose int

hiL18920_ch03_064-123.indd 74 12/08/17 06:54 AM

 Final PDF to printer

3.2  Resource-Allocation Problems   75

from the pop-up menu between the left-hand and right-hand side. In Analytic Solver, choose
the changing cells (C12:D12), and then under the Constraint menu on the Analytic Solver rib-
bon, choose Integer under the Variable Type/Bound submenu.
These changing cells in Figure 3.2 show the optimal solution, (S, L) = (1, 3), obtained after
running Solver.
Excel Tip: To constrain a One of the assumptions of linear programming is that the changing cells are allowed to
range of changing cells to have any values, including fractional values, that satisfy the functional and nonnegativity
be integer in Excel’s Solver,
constraints. Therefore, technically speaking, the TBA problem is not a linear programming
choose the range of cells
in the left-hand side of the problem because of adding the constraints
Add Constraint dialog box ​NumberPurchased = integer​
and choose int from the
pop-up menu. Clicking OK that are displayed in the Solver Parameters box in Figure 3.2. Such a problem that fits linear
then enters the constraint programming except for adding such constraints is called an integer programming problem.
that these cells = integer in The method used by Solver to solve integer programming problems is quite different from
the Solver dialog box. In
that for solving linear programming problems. In fact, integer programming problems tend
Analytic Solver, select the
range of cells to be con- to be much more difficult to solve than linear programming problems so there is consider-
strained integer, and then ably more limitation on the size of the problem. However, this doesn’t matter to a spreadsheet
under the Constraint menu modeler dealing with small problems. From his or her viewpoint, there is virtually no distinc-
on the Analytic Solver rib- tion between linear programming and integer programming problems. They are formulated in
bon, choose Integer under
exactly the same way. Then, at the very end, a decision needs to be made as to whether any of
the Variable Type/Bound
submenu. the changing cells need to be restricted to integer values. If so, those constraints are added as
described above. Keep this option in mind as we continue to discuss the formulation of vari-
ous types of linear programming problems throughout the chapter.
Summary of the Formulation
Excel Tip: Even when a The above formulation of a model with one resource constraint and one side constraint for the
changing cell is constrained TBA Airlines problem now can be summarized (in algebraic form) as follows:
to be integer, rounding
errors occasionally will ​​Maximize ​  Profit = 7S + 22L​​
cause Excel to return a non-
subject to
integer value very close to
an integer (e.g., 1.23E-10, 25S + 75L ≤ 250
​​   ​  ​  ​  ​​​
meaning 0.000000000123). S ≤ 5
To make the spreadsheet and
cleaner, you may replace
​​S ≥ 0​  L ≥ 0​​
these “ugly” representations
by their proper integer val-
ues in the changing cells. Capital Budgeting
Excel Tip: In Excel’s Solver Financial planning is one of the most important areas of application for resource-allocation
Options, the Integer Opti- problems. The resources being allocated in this area are quite different from those for applica-
mality (%) setting tions in the production planning area (such as the Wyndor Glass Co. product-mix problem),
(1 percent by default) causes
where the resources tend to be production facilities of various kinds. For financial planning,
Solver to stop solving an
integer programming prob- the resources tend to be financial assets such as cash, securities, accounts receivable, lines of
lem when it finds a feasible credit, and so forth. Our specific example involves capital budgeting, where the resources are
solution whose objective amounts of investment capital available at different points in time.
function value is within
the specified percentage of The Problem
being optimal. In Analytic The Think-Big Development Co. is a major investor in commercial real-estate development
Solver, the equivalent setting
is Integer Tolerance under projects. It currently has the opportunity to share in three large construction projects:
the Engine tab of the Solver Project 1: Construct a high-rise office building.
model (see Figure 2.19).
This is useful to speed up Project 2: Construct a hotel.
solving large problems. Project 3: Construct a shopping center.
For smaller problems (e.g.,
homework problems), this Each project requires each partner to make investments at four different points in time: a
option should be set to 0 to down payment now, and additional capital after one, two, and three years. Table 3.3 shows for
guarantee finding an optimal each project the total amount of investment capital required from all the partners at these four
solution.
points in time. Thus, a partner taking a certain percentage share of a project is obligated to
invest that percentage of each of the amounts shown in the table for the project.

hiL18920_ch03_064-123.indd 75 12/08/17 06:54 AM

 Final PDF to printer

An Application Vignette
A key part of a country’s financial infrastructure is its securities Following more than 12,000 man-hours devoted to this rede-
markets. By allowing a variety of financial institutions and their sign, the new system was successfully launched in N ­ ovember
clients to trade stocks, bonds, and other financial securities, 2008. The core of the new system is a large linear program-
they help fund both public and private initiatives. Therefore, the ming model that is applied many times daily to choose which
efficient operation of its securities markets plays a crucial role pending transactions should be settled immediately with the
in providing a platform for the economic growth of the country. depositor’s available balances. Linear programming is ideally
Each central securities depository and its system for quickly suited for this application because it can maximize the value of
settling security transactions are part of the operational back- the transactions settled while taking into account the various
bone of securities markets and a key component of financial relevant constraints.
system stability. In Mexico, an institution called INDEVAL pro- This application of linear programming has substantially
vides both the central securities depository and its security enhanced and strengthened the Mexican financial infrastruc-
settlement system for the entire country. This security settle- ture by reducing its daily liquidity requirements by $130 billion.
ment system uses electronic book entries, modifying cash and It also reduces the intraday financing costs for market par-
securities balances, for the various parties in the transactions. ticipants by more than $150 million annually. This application
The total value of the securities transactions the INDEVAL also led to INDEVAL winning the prestigious first prize in the
settles averages over $250 billion daily. This makes INDEVAL 2010 international competition for the Franz Edelman Award
the main liquidity conduit for Mexico’s entire financial sector. for Achievement in Operations Research and the Management
Therefore, it is extremely important that INDEVAL’s system for Sciences.
clearing securities transactions be an exceptionally efficient one
Source: D. Muñoz, M. de Lascurain, O. Romeo-Hernandez, F. Solis, L.
that maximizes the amount of cash that can be delivered almost
de los Santoz, A. Palacios-Brun, F. Herrería, and J. Villaseñor, “INDEVAL
instantaneously after the transactions. Because of past dissatis- Develops a New Operating and Settlement System Using Operations
faction with this system, INDEVAL’s board of directors ordered a Research,” Interfaces 41, no. 1 (January–February 2011), pp. 8–17. (A link
major study in 2005 to completely redesign the system. to this article is available at www.mhhe.com/Hillier6e.)

TABLE 3.3
Investment Capital Requirements
Financial Data for the
Projects Being Consid- Year Office Building Hotel Shopping Center
ered for Partial Invest-
0 $40 million $80 million $90 million
ment by the Think-Big
1 60 million 80 million 50 million
Development Co.
2 90 million 80 million 20 million
3 10 million 70 million 60 million
Net present value $45 million $70 million $50 million

All three projects are expected to be very profitable in the long run. So the management of
Think-Big wants to invest as much as possible in some or all of them. Management is willing
to commit all the company’s investment capital currently available, as well as all additional
investment capital expected to become available over the next three years. The objective is
to determine the investment mix that will be most profitable, based on current estimates of
profitability.
Since it will be several years before each project begins to generate income, which will
continue for many years thereafter, we need to take into account the time value of money
in evaluating how profitable it might be. This is done by discounting future cash outflows
(capital invested) and cash inflows (income), and then adding discounted net cash flows, to
calculate a project’s net present value.
Based on current estimates of future cash flows (not included here except for outflows),
the estimated net present value for each project is shown in the bottom row of Table 3.3. All
the investors, including Think-Big, then will split this net present value in proportion to their
share of the total investment.
For each project, participation shares are being sold to major investors, such as Think-
Big, who become the partners for the project by investing their proportional shares at the four
specified points in time. For example, if Think-Big takes a 10 percent share of the office build-
ing, it will need to provide $4 million now, and then $6 million, $9 million, and $1 ­million in
1 year, 2 years, and 3 years, respectively.
76

hiL18920_ch03_064-123.indd 76 12/08/17 06:54 AM

 Final PDF to printer

3.2  Resource-Allocation Problems   77

The company currently has $25 million available for capital investment. Projections are
that another $20 million will become available after one year, $20 million more after two
years, and another $15 million after three years. What share should Think-Big take in the
respective projects to maximize the total net present value of these investments?
Formulation
This is a resource-allocation problem. The activities under consideration are
Activity 1: Invest in the construction of an office building.
Activity 2: Invest in the construction of a hotel.
Activity 3: Invest in the construction of a shopping center.
Thus, the decisions to be made are the levels of these activities, that is, what participation
share to take in investing in each of these projects. A participation share can be expressed as
either a fraction or a percentage of the entire project, so the entire project is considered to be
one “unit” of that activity.
The resources to be allocated to these activities are the funds available at the four invest-
ment points. Funds not used at one point are available at the next point. (For simplicity, we
will ignore any interest earned on these funds.) Therefore, the resource constraint for each
point must reflect the cumulative funds to that point.
Resource 1: Total investment capital available now.
Resource 2: Cumulative investment capital available by the end of one year.
Resource 3: Cumulative investment capital available by the end of two years.
Resource 4: Cumulative investment capital available by the end of three years.
Since the amount of investment capital available is $25 million now, another $20 million
in one year, another $20 million in two years, and another $15 million in three years, the
amounts available of the resources are the following:
Amount of resource 1 available = $25million
Amount of resource 2 available = $(25 + 20) million = $45 million
Amount of resource 3 available = $(25 + 20 + 20) million = $65 million
Amount of resource 4 available = $(25 + 20 + 20 + 15) million = $80 million
Table 3.4 shows all the data involving these resources. The rightmost column gives the
amounts of resources available calculated above. The middle columns show the cumulative
amounts of the investment capital requirements listed in Table 3.3. For example, in the Office
Building column of Table 3.4, the second number ($100 million) is obtained by adding the
first two numbers ($40 million and $60 million) in the Office Building column of Table 3.3.
The Data   As with any resource-allocation problem, three kinds of data need to be gathered.
One is the amounts available of the resources, as given in the rightmost column of Table 3.4.
A second is the amount of each resource needed by each project, which is given in the middle
columns of this table. A third is the contribution of each project to the overall measure of
performance (net present value), as given in the bottom row of Table 3.3.
The first step in formulating the spreadsheet model is to enter these data into data cells
in the spreadsheet. In Figure 3.3, the data cells (and their range names) are NetPresentValue

TABLE 3.4
Cumulative Investment Capital Required for an Entire Project
Resource Data for the
Think-Big Development Office Shopping Amount of
Co. Investment-Mix Resource Building Hotel Center Resource Available
Problem
1 (Now) $ 40 million $ 80 million $ 90 million $25 million
2 (End of year 1)   100 million     160 million     140 million 45 million
3 (End of year 2)    190 million     240 million     160 million 65 million
4 (End of year 3)    200 million     310 million     220 million 80 million

hiL18920_ch03_064-123.indd 77 12/08/17 06:54 AM

 Final PDF to printer

78  Chapter Three   Linear Programming: Formulation and Applications

FIGURE 3.3
The spreadsheet model for the Think-Big problem, including the formulas for the objective cell TotalNPV (H16) and the other output
cells CapitalSpent (F9:F12), as well as the specifications needed to set up Solver. The changing cells ParticipationShare (C16:E16)
show the optimal solution obtained by Solver.

A B C D E F G H
1 Think-Big Development Co. Capital Budgeting Program
2
3 Office Shopping
4 Building Hotel Center
5 Net Present Value 45 70 50
6 ($millions) Cumulative Cumulative
7 Capital Capital
8 Cumulative Capital Required ($millions) Spent Available
9 Now 40 80 90 25 ≤ 25
10 End of Year 1 100 160 140 44.76 ≤ 45
11 End of Year 2 190 240 160 60.58 ≤ 65
12 End of Year 3 200 310 220 80 ≤ 80
13
14 Office Shopping Total NPV
15 Building Hotel Center ($millions)
16 Participation Share 0.00% 16.50% 13.11% 18.11

Solver Parameters F
Set Objective Cell: TotalNPV Range Name Cells
To: Max 6 Cumulative
By Changing Variable Cells: CapitalAvailable H9:H12
ParticipationShare CapitalRequired C9:E12 7 Capital
Subject to the Constraints: CapitalSpent F9:F12 8 Spent
CapitalSpent <= CapitalAvailable ParticipationShare C16:E16
9 =SUMPRODUCT(C9:E9,ParticipationShare)
NetPresentValue C5:E5
Solver Options: 10 =SUMPRODUCT(C10:E10,ParticipationShare)
Make Variables Nonnegative
TotalNPV H16
Solving Method: Simplex LP 11 =SUMPRODUCT(C11:E11,ParticipationShare)
12 =SUMPRODUCT C12:E12,ParticipationShare)

H
14 Total NPV
15 ($millions)
16 =SUMPRODUCT(NetPresentValue,ParticipationShare)

(C5:E5), CapitalRequired (C9:E12), and CapitalAvailable (H9:H12). To save space on the

spreadsheet, these numbers are entered in units of millions of dollars.
The Decisions   With three activities under consideration, there are three decisions to be made.
Decision 1: OB = Participation share in the office building.
Decision 2: H = Participation share in the hotel.
Decision 3: SC = Participation share in the shopping center.
For example, if Think-Big management were to decide to take a one-tenth participation share
(i.e., a 10 percent participation share) in each of these projects, then
OB = 0.1 = 10%
H = 0.1 = 10%
SC = 0.1 = 10%
However, it may not be desirable to take the same participation share (expressed as either a
fraction or a percentage) in each of the projects, so the idea is to choose the best combination

hiL18920_ch03_064-123.indd 78 12/08/17 06:54 AM

 Final PDF to printer

3.2  Resource-Allocation Problems   79

of values of OB, H, and SC. In Figure 3.3, the participation shares (expressed as percentages)
have been placed in changing cells under the data cells (row 16) in the columns for the three
projects, so
​​OB → cell C16​  H → D16​  SC → cell E16​​
where these cells are collectively referred to by the range name ParticipationShare (C16:E16).

The Constraints   The numbers in these changing cells make sense only if they are non-
negative, so the Make Variables Nonnegative option will need to be selected in the Excel’s
Solver dialog box (or equivalently in Analytic Solver, set the Assume Non-Negative option
to True in the Engine tab of the Model pane). In addition, the four resources require resource
constraints:
Total invested now ≤ 25 (millions of dollars available)
Total invested within 1 year ≤ 45 (millions of dollars available)
Total invested within 2 years ≤ 65 (millions of dollars available)
Total invested within 3 years ≤ 80 (millions of dollars available)
The data in columns C, D, and E indicate that (in millions of dollars)
Total invested now = 40 OB + 80 H + 90 SC
Total invested within 1 year = 100 OB + 160 H + 140 SC
Total invested within 2 years = 190 OB + 240 H + 160 SC
Total invested within 3 years = 200 OB + 310 H + 220 SC
These totals are calculated in the output cells CapitalSpent (F9:F12) using the SUMPRODUCT
function, as shown below the spreadsheet in Figure 3.3. Finally, ≤ signs are entered into col-
umn G to indicate the resource constraints that will need to be entered in Solver.

The Measure of Performance   The objective is to

​​Maximize ​  NPV = total net present value of the investments​​

NetPresentValue (C5:E5) shows the net present value of each entire project, while
­ParticipationShare (C16:E16) shows the participation share for each of the projects. There-
fore, the total net present value of all the participation shares purchased in all three projects is
(in millions of dollars)

NPV = 45 OB + 70 H + 50 SC
​  =​ ​  ​  SUMPRODUCT (​​ ​​NetPresentValue,
​ ​​​   
      ​ ParticipationShare​)​​​​​​
​ → cell H16

Summary of the Formulation   This completes the formulation of the linear programming
model on the spreadsheet, as summarized below (in algebraic form).
​​Maximize ​  NPV = 45 OB + 70 H + 50 SC​​
subject to
Total invested now:    40 OB +     80 H +     90 SC ≤ 25
Total invested within 1 year:      100 OB + 160 H + 140 SC ≤ 45
Total invested within 2 years: 190 OB + 240 H + 160 SC ≤ 65
Total invested within 3 years: 200 OB + 310 H + 220 SC ≤ 80
and
​​OB ≥ 0​  H ≥ 0​  SC ≥ 0​​
where all these numbers are in units of millions of dollars.

hiL18920_ch03_064-123.indd 79 12/08/17 06:54 AM

 Final PDF to printer

80  Chapter Three   Linear Programming: Formulation and Applications

Note that this model possesses the key identifying feature for resource-allocation prob-
lems, namely, each functional constraint is a resource constraint that has the form
​Amount of resource used ≤ Amount of resource available​

Solving the Model  The lower left-hand side of Figure 3.3 shows the entries needed in
Solver to specify the model, along with the selection of the usual two options. The spread-
sheet shows the resulting optimal solution in row 16, namely,
Invest nothing in the office building.
Invest in 16.50 percent of the hotel.
Invest in 13.11 percent of the shopping center.
TotalNPV (H16) indicates that this investment program would provide a total net present
value of $18.11 million.
This amount actually is only an estimate of what the total net present value would turn out
to be, depending on the accuracy of the financial data given in Table 3.3. There is some uncer-
tainty about the construction costs for the three real estate projects, so the actual investment
capital requirements for years 1, 2, and 3 may deviate somewhat from the amounts specified
in this table. Because of the risk involved in these projects, the net present value for each one
also might deviate from the amounts given at the bottom of the table. Chapter 5 describes
one approach to analyzing the effect of such deviations. Chapters 12 and 13 will present
another technique, called computer simulation, for systematically taking future uncertainties
into account. Section 13.4 will focus on further analysis of this same example.

Another Look at Resource Constraints

These examples of resource-allocation problems illustrate a variety of resources: financial
allocations for advertising and planning purposes, TV commercial spots available for pur-
chase, available production capacities of different plants, the total amount of capital avail-
able for investment, and cumulative investment capital available by certain times. However,
these illustrations only scratch the surface of the realm of possible resources that need to be
allocated to activities in resource-allocation problems. In fact, by interpreting resource suf-
ficiently broadly, any restriction on the decisions to be made that has the form
​Amount used ≤ Amount available​
can be thought of as a resource constraint, where the thing whose amount is being measured
is the corresponding “resource.” Since any functional constraint with a ≤ sign in a linear
programming model (including the side constraint limiting the number of small airplanes
that can be purchased in the TBA Airlines example) can be verbalized in this form, any such
constraint can be thought of as a resource constraint.
Hereafter, we will use resource constraint to refer to any functional constraint with a
≤ sign in a linear programming model. The constant on the right-hand side represents the
amount available of a resource. Therefore, the left-hand side represents the amount used of this
resource. In the algebraic form of the constraint, the coefficient (positive or negative) of each
decision variable is the resource usage per unit of the corresponding activity.

Summary of the Formulation Procedure for

Resource-Allocation Problems
The four examples illustrate that the following steps are used for any resource-allocation prob-
lem to define the specific problem, gather the relevant data, and then formulate the linear
programming model.
1. Since any linear programming problem involves finding the best mix of levels of various
activities, identify these activities for the problem at hand. The decisions to be made are
the levels of these activities.
2. From the viewpoint of management, identify an appropriate overall measure of perfor-
mance (commonly profit, or a surrogate for profit) for solutions of the problem.

hiL18920_ch03_064-123.indd 80 12/08/17 06:54 AM

 Final PDF to printer

3.3  Cost–Benefit–Trade-Off Problems   81

FIGURE 3.4
A template of a spreadsheet model for pure resource-allocation problems.
Activities

Unit Profit Profit per unit of activity

Resources Resources
Used Available
Constraints

SUMPRODUCT
Resource used per unit of activity (resource used per unit,
changing cells)
≤

Total Profit
Level of Activity Changing cells SUMPRODUCT(profit per unit, changing cells)

3. For each activity, estimate the contribution per unit of the activity to this overall measure
of performance.
4. Identify the resources that must be allocated to the activities.
5. For each resource, identify the amount available and then the amount used per unit of each
activity.
6. Enter the data gathered in steps 3 and 5 into data cells in a spreadsheet. A convenient for-
mat is to have the data associated with each activity in a separate column, the data for the
unit profit and each constraint in a separate row, and to leave two blank columns between
the activity columns and the amount of resource available column. Figure 3.4 shows a
template of the overall format of a spreadsheet model for resource-allocation problems.
7. Designate changing cells for displaying the decisions on activity levels.
8. For the two blank columns created in step 6, use the left one as a Totals column for output
cells and enter ≤ signs into the right one for all the resources. In the row for each resource,
use the SUMPRODUCT function to enter the total amount used in the Totals column.
9. Designate an objective cell for displaying the overall measure of performance. Use a
SUMPRODUCT function to enter this measure of performance.
All the functional constraints in this linear programming model in a spreadsheet are
resource constraints, that is, constraints with a ≤ sign. This is the identifying feature that clas-
sifies the problem as being a resource-allocation problem.

Review 1. What is the identifying feature for a resource-allocation problem?

Questions 2. What is the form of a resource constraint?
3. What are the three kinds of data that need to be gathered for a resource-allocation problem?
4. Compare the types of activities for the four examples of resource-allocation problems.
5. Compare the types of resources for the four examples of resource-allocation problems.

3.3 COST–BENEFIT–TRADE-OFF PROBLEMS

Cost–benefit–trade-off problems have a form that is very different from resource-allocation
problems. The difference arises from managerial objectives that are very different for the two
kinds of problems.
For resource-allocation problems, limits are set on the use of various resources (including
financial resources), and then the objective is to make the most effective use (according to
some overall measure of performance) of these given resources.
For cost–benefit–trade-off problems, management takes a more aggressive stance, pre-
scribing what benefits must be achieved by the activities under consideration (regardless of

hiL18920_ch03_064-123.indd 81 12/08/17 06:54 AM

 Final PDF to printer

82  Chapter Three   Linear Programming: Formulation and Applications

A cost–benefit–trade-off the resulting resource usage), and then the objective is to achieve all these benefits with mini-
formulation enables man- mum cost. By prescribing a minimum acceptable level for each kind of benefit, and then mini-
agement to specify mini-
mizing the cost needed to achieve these levels, management hopes to obtain an appropriate
mum goals for the benefits
that need to be achieved by trade-off between cost and benefits. (You will see in Chapter 5 that what-if analysis plays a
the activities. key role in providing the additional information needed for management to choose the best
trade-off between cost and benefits.)
Cost–benefit–trade-off problems are linear programming problems where the mix of levels
of various activities is chosen to achieve minimum acceptable levels for various benefits at a
minimum cost. The identifying feature is that each functional constraint is a benefit constraint,
which has the form
​Level achieved ≥ Minimum acceptable level​

for one of the benefits.

Interpreting benefit broadly, we can think of any functional constraint with a ≥ sign as
a benefit constraint. In most cases, the minimum acceptable level will be prescribed by
management as a policy decision, but occasionally this number will be dictated by other
circumstances.
For any cost–benefit–trade-off problem, a major part of the study involves identifying all
the activities and benefits that should be considered and then gathering the data relevant to
these activities and benefits.
These three kinds of data Three kinds of data are needed:
are needed for any cost–
benefit–trade-off problem. 1. The minimum acceptable level for each benefit (a managerial policy decision).
2. For each benefit, the contribution of each activity to that benefit (per unit of the activity).
3. The cost per unit of each activity.
Let’s examine two examples of cost–benefit–trade-off problems.

The Profit & Gambit Co. Advertising-Mix Problem

As described in Section 2.7, the Profit & Gambit Co. will be undertaking a major new adver-
tising campaign focusing on three cleaning products. The two kinds of advertising to be used
are television and the print media. Management has established minimum goals—the mini-
mum acceptable increase in sales for each product—to be gained by the campaign.
The problem is to determine how much to advertise in each medium to meet all the sales
goals at a minimum total cost.
An initial step in formu- The activities in this cost–benefit–trade-off problem are
lating any cost–benefit–­
tradeoff problem is to Activity 1: Advertise on television.
identify the activities and Activity 2: Advertise in the print media.
the benefits.
The benefits being sought from these activities are
Benefit 1: Increased sales for a spray prewash stain remover.
Benefit 2: Increased sales for a liquid laundry detergent.
Benefit 3: Increased sales for a powder laundry detergent.
Management wants these increased sales to be at least 3 percent, 18 percent, and 4 percent,
respectively. As shown in Section 2.7, each benefit leads to a benefit constraint that incor-
porates the managerial goal for the minimum acceptable level of increase in the sales for the
corresponding product, namely,
Level of benefit 1 achieved ≥ 3%
Level of benefit 2 achieved ≥ 18%
Level of benefit 3 achieved ≥ 4%
The data for this problem are given in Table 2.2 (Section 2.7). Section 2.7 describes how
the linear programming model is formulated directly from the numbers in this table.
This example provides an interesting contrast with the Super Grain Corp. case study in ­Section
3.1, which led to a formulation as a resource-allocation problem. Both are advertising-mix

hiL18920_ch03_064-123.indd 82 12/08/17 06:54 AM

 Final PDF to printer

3.3  Cost–Benefit–Trade-Off Problems   83

problems, yet they lead to entirely different linear programming models. They differ because of
the differences in the managerial view of the key issues in each case:
• As the vice president for marketing of Super Grain, Claire Syverson focused first on how
much to spend on the advertising campaign and then set limits (an advertising budget of
$4 million and a planning budget of $1 million) that led to resource constraints.
• The management of Profit & Gambit instead focused on what it wanted the advertising
campaign to accomplish and then set goals (minimum required increases in sales) that led
to benefit constraints.
From this comparison, we see that it is not the nature of the application that determines the
classification of the resulting linear programming formulation. Rather, it is the nature of the
restrictions imposed on the decisions regarding the mix of activity levels. If the restrictions
involve limits on the usage of resources, that identifies a resource-allocation problem. If the
restrictions involve goals on the levels of benefits, that characterizes a cost–benefit–trade-off
problem. Frequently, the nature of the restrictions arise from the way management frames the
problem.
However, we don’t want you to get the idea that every linear programming problem falls
entirely and neatly into either one type or the other. In the preceding section and this one, we
are looking at pure resource-allocation problems and pure cost–benefit–trade-off problems.
Although many real problems tend to be either one type or the other, it is fairly common to
have both resource constraints and benefit constraints, even though one may predominate.
(In the next section, you will see an example of how both types of constraints can arise in the
same problem when the management of the Super Grain Corp. introduces additional con-
siderations into the analysis of their advertising-mix problem.) Furthermore, we still need to
consider additional categories of linear programming problems in the remaining sections of
this chapter.
Now, another example of a pure cost–benefit–trade-off problem.

Personnel Scheduling
One of the common applications of cost–benefit–trade-off analysis involves personnel sched-
uling for a company that provides some kind of service, where the objective is to schedule
the work times of the company’s employees so as to minimize the cost of providing the level
of service specified by management. The following example illustrates how this can be done.
The Problem
Union Airways is adding more flights to and from its hub airport and so needs to hire addi-
tional customer service agents. However, it is not clear just how many more should be hired.
Management recognizes the need for cost control while also consistently providing a satisfac-
tory level of service to the company’s customers, so a desirable trade-off between these two
factors is being sought. Therefore, a management science team is studying how to schedule
the agents to provide satisfactory service with the smallest personnel cost.
Based on the new schedule of flights, an analysis has been made of the minimum number
of customer service agents that need to be on duty at different times of the day to provide a
satisfactory level of service. (The queueing models presented in Chapter 11 can be used to
determine the minimum numbers of agents needed to reduce the waiting times incurred by
customers down to reasonable levels.) These numbers are shown in the last column of Table
3.5 for the time periods given in the first column. The other entries in this table reflect one of
the provisions in the company’s current contract with the union that represents the customer
service agents. The provision is that each agent works an eight-hour shift. The authorized
shifts are
Shift 1: 6:00 AM to 2:00 PM.
Shift 2: 8:00 AM to 4:00 PM.
Shift 3: Noon to 8:00 PM.
Shift 4: 4:00 PM to midnight.
Shift 5: 10:00 PM to 6:00 AM.

hiL18920_ch03_064-123.indd 83 12/08/17 06:54 AM

 Final PDF to printer

84  Chapter Three   Linear Programming: Formulation and Applications

TABLE 3.5
Time Periods Covered by Shift
Data for the Union Air-
ways Personnel Schedul- Minimum Number
ing Problem Time Period 1 2 3 4 5 of Agents Needed
6:00 AM to 8:00 AM ✓ 48
8:00 AM to 10:00 AM ✓ ✓ 79
10:00 AM to noon ✓ ✓ 65
Noon to 2:00 PM ✓ ✓ ✓ 87
2:00 PM to 4:00 PM ✓ ✓ 64
4:00 PM to 6:00 PM ✓ ✓ 73
6:00 PM to 8:00 PM ✓ ✓ 82
8:00 PM to 10:00 PM ✓ 43
10:00 PM to midnight ✓ ✓ 52
Midnight to 6:00 AM ✓ 15
Daily cost per agent $170 $160 $175 $180 $195

Check marks in the main body of Table 3.5 show the time periods covered by the respec-
tive shifts. Because some shifts are less desirable than others, the wages specified in the
contract differ by shift. For each shift, the daily compensation (including benefits) for each
agent is shown in the bottom row. The problem is to determine how many agents should be
assigned to the respective shifts each day to minimize the total personnel cost for agents,
based on this bottom row, while meeting (or surpassing) the service requirements given in
the last column.
Formulation
This problem is, in fact, a pure cost–benefit–trade-off problem. To formulate the problem, we
need to identify the activities and benefits involved.
Activities correspond to shifts.
The level of each activity is the number of agents assigned to that shift.
A unit of each activity is one agent assigned to that shift.
Thus, the general description of a linear programming problem as finding the best mix of
activity levels can be expressed for this specific application as finding the best mix of shift
sizes.
Benefits correspond to time periods.
For each time period, the benefit provided by the activities is the service that agents pro-
vide customers during that period.
The level of a benefit is measured by the number of agents on duty during that time period.
Once again, a careful formulation of the problem, including gathering all the relevant data,
leads rather directly to a spreadsheet model. This model is shown in Figure 3.5, and we out-
line its formulation below.
The Data   As indicated in this figure, all the data in Table 3.5 have been entered directly
into the data cells CostPerShift (C5:G5), ShiftWorksTimePeriod (C8:G17), and Minimum-
Needed (J8:J17). For the ShiftWorksTimePeriod (C8:G17) data, an entry of 1 indicates that
the corresponding shift includes that time period whereas 0 indicates not. Like any cost–­
benefit–tradeoff problem, these numbers indicate the contribution of each activity to each
benefit. Each agent working a shift contributes either 0 or 1 toward the minimum number of
agents needed in a time period.
The Decisions   Since the activities in this case correspond to the five shifts, the decisions to
be made are
S1 = Number of agents to assign to Shift 1 (starts at 6 AM)
S2 = Number of agents to assign to Shift 2 (starts at 8 AM)

hiL18920_ch03_064-123.indd 84 12/08/17 06:54 AM

 Final PDF to printer

3.3  Cost–Benefit–Trade-Off Problems   85

FIGURE 3.5
The spreadsheet model for the Union Airways problem, including the formulas for the objective cell TotalCost (J21) and the other
output cells TotalWorking (H8:H17), as well as the specifications needed to set up Solver. The changing cells NumberWorking
(C21:G21) show the optimal solution obtained by Solver.

A B C D E F G H I J
1 Union Airways Personnel Scheduling Problem
2
3 6AM–2PM 8AM–4PM Noon–8PM 4PM–Midnight 10PM–6AM
4 Shift Shift Shift Shift Shift
5 Cost per Shift $170 $160 $175 $180 $195
6 Total Minimum
7 Time Period Shift Works Time Period? (1=yes, 0=no) Working Needed
8 6AM–8AM 1 0 0 0 0 48 ≥ 48
9 8AM–10AM 1 1 0 0 0 79 ≥ 79
10 10AM–12PM 1 1 0 0 0 79 ≥ 65
11 12PM–2PM 1 1 1 0 0 118 ≥ 87
12 2PM–4PM 0 1 1 0 0 70 ≥ 64
13 4PM–6PM 0 0 1 1 0 82 ≥ 73
14 6PM–8PM 0 0 1 1 0 82 ≥ 82
15 8PM–10PM 0 0 0 1 0 43 ≥ 43
16 10PM–12AM 0 0 0 1 1 58 ≥ 52
17 12AM–6AM 0 0 0 0 1 15 ≥ 15
18
19 6AM–2PM 8AM–4PM Noon–8PM 4PM–Midnight 10PM–6AM
20 Shift Shift Shift Shift Shift Total Cost
21 Number Working 48 31 39 43 15 $30,610

Solver Parameters H
Set Objective Cell: TotalCost Range Name Cells
To: Min 6 Total
CostPerShift C5:G5
By Changing Variable Cells:
MinimumNeeded J8:J17 7 Working
NumberWorking
Subject to the Constraints: NumberWorking C21:G21 8 =SUMPRODUCT(C8:G8,NumberWorking)
NumberWorking = integer ShiftWorksTimePeriod C8:G17
9 =SUMPRODUCT(C9:G9,NumberWorking)
TotalWorking >= MinimumNeeded TotalCost J21
TotalWorking H8:H17 10 =SUMPRODUCT(C10:G10,NumberWorking)
Solver Options: 11 =SUMPRODUCT(C11:G11,NumberWorking)
Make Variables Nonnegative
Solving Method: Simplex LP 12 =SUMPRODUCT(C12:G12,NumberWorking)
13 =SUMPRODUCT(C13:G13,NumberWorking)
14 =SUMPRODUCT(C14:G14,NumberWorking)
15 =SUMPRODUCT(C15:G15,NumberWorking)
16 =SUMPRODUCT(C16:G16,NumberWorking)
17 =SUMPRODUCT(C17:G17,NumberWorking)

J
20 Total Cost
21 =SUMPRODUCT(CostPerShift,NumberWorking)

S3 = Number of agents to assign to Shift 3 (starts at noon)

S4 = Number of agents to assign to Shift 4 (starts at 4 PM)
S5 = Number of agents to assign to Shift 5 (starts at 10 PM)
The changing cells to hold these numbers have been placed in the activity columns in row 21, so
​​​S​  1​​ → cell C21​  ​S​  2​​ → cell D21​  . . .​  ​S​  5​​ → cell G21​​

hiL18920_ch03_064-123.indd 85 12/08/17 06:54 AM

 Final PDF to printer

86  Chapter Three   Linear Programming: Formulation and Applications

where these cells are collectively referred to by the range name NumberWorking (C21:G21).
The Constraints   These changing cells need to be nonnegative. In addition, we need 10
benefit constraints, where each one specifies that the total number of agents serving in the
corresponding time period listed in column B must be no less than the minimum acceptable
number given in column J. Thus, these constraints are

Total number of agents serving 6 − 8 am ≥ 48 (min. acceptable)

Total number of agents serving 8 − 10 am ≥ 79 (min. acceptable)
​​​​      
      ​  ​  ​  .​  ​  ​ ​ ​  ​​ ​​​
​ . ​ ​
​ . ​ ​
Total number of agents serving midnight − 6 am ≥ 15 (min. acceptable)

Since columns C to G indicate which of the shifts serve each of the time periods, these totals are

Total number of agents serving 6 − 8 am = ​S​  1​​

Total number of agents serving 8 − 10 am = ​S​  1​​ + ​S​  2​​
​​​​     
     
      ​  ​  ​  .​  ​  ​​  ​​
​ ​
​ . ​
​ . ​
Total number of agents serving midnight − 6 am = ​S​  5​​

These totals are calculated in the output cells TotalWorking (H8:H17) using the SUMPROD-
UCT functions shown below the spreadsheet in Figure 3.5.
One other type of constraint is that the number of agents assigned to each shift must have
an integer value. These constraints for the five shifts should be added in the same way as
described for the TBA Airlines problem in Section 3.2. In particular, with Excel’s Solver they
are added in the Add Constraint dialog box by entering NumberWorking on the left-hand side
and then choosing int from the pop-up menu between the left-hand side and the right-hand
side. The set of constraints, NumberWorking = integer, then appears in the Solver Parameters,
as shown in Figure 3.5. In Analytic Solver, select the range of cells to be constrained integer,
and then under the Constraint menu on the Analytic Solver ribbon, choose Integer under the
Variable Type/Bound submenu.
The Measure of Performance   The objective is to
​​Minimize ​  Cost = Total daily personnel cost for all agents​​
Since CostPerShift (C5:G5) gives the daily cost per agent on each shift and NumberWorking
(C21:G21) gives the number of agents working each shift,

Cost = 170 ​S​  1​​  + 160 ​S​  2​​  + 175 ​S​  3​​  + 180 ​S​  4​​  + 195 ​S​  5​​  (in dollars)
      
​ ​     
​​ ​  =​  ​  SUMPRODUCT (CostPerShift, NumberWorking)​ ​​​
​ → cell J21

Summary of the Formulation   The above steps provide the complete formulation of the
linear programming model on a spreadsheet, as summarized below (in algebraic form).
​​Minimize​  Cost = 170 ​S​  1​​  + 160 ​S​  2​​  + 175 ​S​  3​​  + 180 ​S​  4​​  + 195 ​S​  5​​​ (in dollars)​
subject to
Total agents 6 − 8 am ​ ​S​  1​​ ≥ 48
Total agents 8 − 10 am ​ ​S​  1​​ + ​S​  2​​ ≥ 79
      .
    
​​     ​  ​  ​  ​  ​​  ​​  ​  ​  ​  ​​  ​  ​​ ​​​
      . ​ ​ ​ ​
      . ​ ​ ​ ​
Total agents midnight − 6 am ​   ​   S​  5​​ ≥ 15

hiL18920_ch03_064-123.indd 86 12/08/17 06:54 AM

 Final PDF to printer

3.4  Mixed Problems   87

FIGURE 3.6
A template of a spreadsheet model for pure cost–benefit–trade-off problems.
Activities

Unit Cost Cost per unit of activity

Benefit Benefit
Achieved Needed
Constraints

SUMPRODUCT
Benefit achieved per unit of activity (benefit per unit,
changing cells)
≥

Total Cost
Level of Activity Changing cells SUMPRODUCT(cost per unit, changing cells)

and
​​​S​  1​​ ≥ 0​  ​S​  2​​ ≥ 0​  ​S​  3​​ ≥ 0​  ​S​  4​​ ≥ 0​  ​S​  5​​ ≥ 0​​

Solving the Model   The lower left-hand corner of Figure 3.5 shows the entries needed in Solver,
along with the selection of the usual two options. After solving, NumberWorking (C21:G21) in
the spreadsheet shows the resulting optimal solution for the number of agents that should be
assigned to each shift. TotalCost (J21) indicates that this plan would cost $30,610 per day.

Summary of the Formulation Procedure for Cost–Benefit–Trade-Off

Problems
The nine steps in formulating any cost–benefit–trade-off problem follow the same pattern as
presented at the end of the preceding section for resource-allocation problems, so we will not
repeat them here. The main differences are that the overall measure of performance now is
the total cost of the activities (or some surrogate of total cost chosen by management) in steps
2 and 3, benefits now replace resources in steps 4 and 5, and ≥ signs now are entered to the
right of the output cells for benefits in step 8. Figure 3.6 shows a template of the format of a
spreadsheet model for cost–benefit–trade-off problems.
All the functional constraints in the resulting model are benefit constraints, that is, con-
straints with a ≥ sign. This is the identifying feature of a pure cost–benefit–trade-off problem.

Review 1. What is the difference in managerial objectives between resource-allocation problems and
Questions cost–benefit–trade-off problems?
2. What is the identifying feature of a cost–benefit–trade-off problem?
3. What is the form of a benefit constraint?
4. What are the three kinds of data that need to be gathered for a cost–benefit–trade-off problem?
5. Compare the types of activities for the two examples of cost–benefit–trade-off problems.
6. Compare the types of benefits for the two examples of cost–benefit–trade-off problems.

3.4 MIXED PROBLEMS

Sections 3.2 and 3.3 each described a broad category of linear programming problems—
resource-allocation and cost–benefit–trade-off problems. As summarized in Table 3.6, each
features one of the first two types of functional constraints shown there. In fact, the identifying
feature of a pure resource-allocation problem is that all its functional constraints are resource
constraints. The identifying feature of a pure cost–benefit–trade-off problem is that all its
functional constraints are benefit constraints. (Keep in mind that the functional constraints
include all the constraints of a problem except its nonnegativity constraints.)

hiL18920_ch03_064-123.indd 87 12/08/17 06:54 AM

 Final PDF to printer

88  Chapter Three   Linear Programming: Formulation and Applications

The bottom row of Table 3.6 shows the last of the three types of functional constraints,
namely, fixed-requirement constraints, which require that the left-hand side of each such
constraint must exactly equal some fixed amount. Thus, since the left-hand side represents the
amount provided of some quantity, the form of a fixed-requirement constraint is
​Amount provided = Required amount​
The identifying feature of a pure fixed-requirements problem is that it is a linear program-
ming problem where all its functional constraints are fixed-requirement constraints. The next
two sections will describe two particularly prominent types of fixed-requirement problems
called transportation problems and assignment problems.
However, before turning to these types of problems, we first will use a continuation of the
Super Grain case study from Section 3.1 to illustrate how many linear programming problems
fall into another broad category called mixed problems.
Many linear programming problems do not fit completely into any of the previously discussed
categories (pure resource-allocation problems, cost–benefit–trade-off problems, and fixed-
requirement problems) because the problem’s functional constraints include more than one of
the types shown in Table 3.6. Such problems are called mixed problems.

Now let us see how a more careful analysis of the Super Grain case study turns this resource-
allocation problem into a mixed problem that includes all three types of functional constraints
shown in Table 3.6.

Super Grain Management Discusses Its Advertising-Mix Problem

The description of the Super Grain case study in Section 3.1 ends with Clair Syverson (Super
Grain’s vice president for marketing) sending a memorandum to the company’s president,
David Sloan, requesting a meeting to evaluate her proposed promotional campaign for the
company’s new breakfast cereal.
Soon thereafter, Claire Syverson and David Sloan meet to discuss plans for the campaign.
David Sloan (president):   Thanks for your memo, Claire. The plan you outline for the
promotional campaign looks like a reasonable one. However, I am surprised that it does not
make any use of TV commercials. Why is that?
Claire Syverson (vice president for marketing):   Well, as I described in my memo, I used
a spreadsheet model to see how to maximize the number of exposures from the campaign
and this turned out to be the plan that does this. I also was surprised that it did not include
TV commercials, but the model indicated that introducing commercials would provide less
exposures on a dollar-for-dollar basis than magazine ads and Sunday supplement ads. Don’t
you think it makes sense to use the plan that maximizes the number of exposures?
David:   Not necessarily. Some exposures are a lot less important than others. For example,
we know that middle-aged adults are not big consumers of our cereals, so we don’t care
very much how many of those people see our ads. On the other hand, young children are big
consumers. Having TV commercials on the Saturday morning programs for children is our

TABLE 3.6
Types of Functional Constraints

Type Form* Typical Interpretation Main Usage

Resource constraint LHS ≤ RHS For some resource, Resource-allocation problems
Amount used ≤ Amount available and mixed problems
Benefit constraint LHS ≥ RHS For some benefit, Cost–benefit–trade-off prob-
Level achieved ≥ Minimum lems and mixed problems
acceptable level
Fixed-requirement LHS = RHS For some quantity, Fixed-requirements problems
constraint Amount provided = Required amount and mixed problems

*LHS = Left-hand side (a SUMPRODUCT function).

  RHS = Right-hand side (a constant).

hiL18920_ch03_064-123.indd 88 12/08/17 06:54 AM

 Final PDF to printer

3.4  Mixed Problems   89

primary method of reaching young children. You know how important it will be to get young
children to ask their parents for Crunchy Start. That is our best way of generating first-time
sales. Those commercials also get seen by a lot of parents who are watching the programs
with their kids. What we need is a commercial that is appealing to both parents and kids, and
that gets the kids immediately bugging their parents to go buy Crunchy Start. I think that is a
real key to a successful campaign.
Claire:   Yes, that makes a lot of sense. In fact, I already have set some goals regarding
the number of young children and the number of parents of young children that need to be
reached by this promotional campaign.
David:   Good. Did you include those goals in your spreadsheet model?
Claire:   No, I didn’t.
David:   Well, I suggest that you incorporate them directly into your model. I suspect that
maximizing exposures while also meeting your goals will give us a high impact plan that
includes some TV commercials.
Claire:   Good idea. I’ll try it.
David:   Are there any other factors that the plan in your memo doesn’t take into account as
well as you would like?
Claire:   Well, yes, one. The plan doesn’t take into account my budget for cents-off coupons
in magazines and newspapers.
David:   You should be able to add that to your model as well. Why don’t you go back and
see what happens when you incorporate these additional considerations?
Claire:   OK, will do. You seem to have had a lot of experience with spreadsheet
modeling.
David:   Yes. It is a great tool as long as you maintain some healthy skepticism about what
comes out of the model. No model can fully take into account everything that we must con-
sider when dealing with managerial problems. This is especially true the first time or two
you run the model. You need to keep asking, what are the missing quantitative consider-
ations that I still should add to the model? Then, after you have made the model as complete
as possible and obtained a solution, you still need to use your best managerial judgment to
weigh intangible considerations that cannot be incorporated into the model.

Incorporating Additional Managerial Considerations

into the Super Grain Model
Therefore, David and Claire conclude that the spreadsheet model needs to be expanded to
incorporate some additional considerations. In particular, since the promotional campaign
is for a breakfast cereal that should have special appeal to young children, they feel that two
audiences should be targeted—young children and parents of young children. (This is why
one of the three advertising media recommended by Giacomi & Jackowitz is commercials on
children’s television programs Saturday morning.) Consequently, Claire now has set two new
goals for the campaign.
Goal 1: The advertising should be seen by at least five million young children.
Goal 2: The advertising should be seen by at least five million parents of young children.
In effect, these two goals are minimum acceptable levels for two special benefits to be
achieved by the advertising activities.
Benefit 1: Promoting the new breakfast cereal to young children.
Benefit 2: Promoting the new breakfast cereal to parents of young children.
Because of the way the goals have been articulated, the level of each of these benefits is mea-
sured by the number of people in the specified category that are reached by the advertising.
To enable constructing the corresponding benefit constraints (as described in Section 3.3),
Claire asks Giacomi & Jackowitz to estimate how much each advertisement in each of the
media will contribute to each benefit, as measured by the number of people reached in the
specified category. These estimates are given in Table 3.7.

hiL18920_ch03_064-123.indd 89 12/08/17 06:54 AM

 Final PDF to printer

90  Chapter Three   Linear Programming: Formulation and Applications

TABLE 3.7
Benefit Data for the Revised Super Grain Corp. Advertising-Mix Problem

Number Reached in Target Category (in millions)

Each TV Each Each Minimum
Target Category Commercial Magazine Ad Sunday Ad Acceptable Level
Young children 1.2 0.1 0 5
Parents of young 0.5 0.2 0.2 5
children

It is interesting to observe that management wants special consideration given to these two
kinds of benefits even though the original spreadsheet model (Figure 3.1) already takes them
into account to some extent. As described in Section 3.1, the expected number of exposures
is the overall measure of performance to be maximized. This measure counts up all the times
that an advertisement is seen by any individual, including all those individuals in the target
audiences. However, maximizing this general measure of performance does not ensure that
Benefit constraints are use- the two specific goals prescribed by management (Claire Syverson) will be achieved. Claire
ful for incorporating mana- feels that achieving these goals is essential to a successful promotional campaign. Therefore,
gerial goals into the model. she complements the general objective with specific benefit constraints that do ensure that the
goals will be achieved. Having benefit constraints added to incorporate managerial goals into
the model is a prerogative of management.
Claire has one more consideration she wants to incorporate into the model. She is a strong
believer in the promotional value of cents-off coupons (coupons that shoppers can clip from
printed advertisements to obtain a refund of a designated amount when purchasing the adver-
tised item). Consequently, she always earmarks a major portion of her annual marketing bud-
get for the redemption of these coupons. She still has $1,490,000 left from this year’s allotment
for coupon redemptions. Because of the importance of Crunchy Start to the company, she has
decided to use this entire remaining allotment in the campaign promoting this cereal.
This fixed amount for coupon redemptions is a fixed requirement that needs to be expressed
as a fixed-requirement constraint. As described at the beginning of this section, the form of a
fixed-requirement constraint is that, for some type of quantity,
​Amount provided = Required amount​
In this case, the quantity involved is the amount of money provided for the redemption of
cents-off coupons. To specify this constraint in the spreadsheet, we need to estimate how
much each advertisement in each of the media will contribute toward fulfilling the required
amount for the quantity. Both medium 2 (advertisements in food and family-oriented maga-
zines) and medium 3 (advertisements in Sunday supplements of major newspapers) will fea-
ture cents-off coupons. The estimates of the amount of coupon redemption per advertisement
in each of these media is given in Table 3.8.

Formulation of the Revised Spreadsheet Model

Figure 3.7 shows one way of formatting the spreadsheet to expand the original spreadsheet
model in Figure 3.1 to incorporate the additional managerial considerations. We then outline
the four components of the revised model next.

TABLE 3.8
Data for the Fixed-Requirement Constraint for the Revised Super Grain Corp. Advertising-Mix Problem

Contribution toward Required Amount

Each Each Each Required

Requirement TV Spot Magazine Ad Sunday Ad Amount
Coupon redemption 0 $40,000 $120,000 $1,490,000

hiL18920_ch03_064-123.indd 90 12/08/17 06:54 AM

 Final PDF to printer

3.4  Mixed Problems   91

FIGURE 3.7
The spreadsheet model for the revised Super Grain problem, including the formulas for the objective cell TotalExposures (H19) and
the other output cells in column F, as well as the specifications needed to set up Solver. The changing cells NumberOfAds (C19:E19)
show the optimal solution obtained by Solver.

A B C D E F G H
1 Super Grain Corp. Advertising-Mix Problem
2
3 TV Spots Magazine Ads SS Ads
4 Exposures per Ad 1,300 600 500
5 (thousands)
6 Cost per Ad ($thousands) Budget Spent Budget Available
7 Ad Budget 300 150 100 3,775 ≤ 4,000
8 Planning Budget 90 30 40 1,000 ≤ 1,000
9
10 Number Reached per Ad (millions) Total Reached Minimum Acceptable
11 Young Children 1.2 0.1 0 5 ≥ 5
TotalExposures
12 Parents of Young Children 0.5 0.2 0.2 5.85 ≥ 5
($thousands)
13
14 TV Spots Magazine Ads SS Ads Total Redeemed Required Amount
15 Coupon Redemption 0 40 120 1,490 = 1,490
16 per Ad ($thousands)
17 Total Exposures
18 TV Spots Magazine Ads SS Ads (thousands)
19 Number of Ads 3 14 7.75 16,175
20 ≤
21 Maximum TV Spots 5

Solver Parameters F
Set Objective Cell: TotalExposures Range Name Cells
To: Max 6 Budget Spent
By Changing Variable Cells: BudgetAvailable H7:H8
NumberOfAds BudgetSpent F7:F8 7 =SUMPRODUCT(C7:E7,NumberOfAds)
Subject to the Constraints: CostPerAd C7:E8
BudgetSpent <= Budget Available 8 =SUMPRODUCT(C8:E8,NumberOfAds)
TVSpots <= MaxTVSpots CouponRedemptionPerAd C15:E15
9
TotalReached >= MinimumAcceptable ExposuresPerAd C4:E4
TotalRedeemed = RequiredAmount MaxTVSpots C21 10 Total Reached
MinimumAcceptable H11:H12 11 =SUMPRODUCT(C11:E11,NumberOfAds)
Solver Options:
Make Variables Nonnegative NumberOfAds C19:E19
12 =SUMPRODUCT(C12:E12,NumberOfAds)
Solving Method: Simplex LP NumberReachedPerAd C11:E12
RequiredAmount H15 13
TotalExposures H19 14 Total Redeemed
TotalReached F11:F12 15 =SUMPRODUCT(CouponRedemptionPerAd,
TotalRedeemed F15
NumberOfAds)
TVSpots C19

H
17 Total Exposures
18 (thousands)
19 =SUMPRODUCT(ExposuresPerAd,NumberOfAds)

The Data
Additional data cells in NumberReachedPerAd (C11:E12), MinimumAcceptable (H11:H12),
CouponRedemptionPerAd (C15:E15), and RequiredAmount (H15) give the data in Tables 3.7
and 3.8.

hiL18920_ch03_064-123.indd 91 12/08/17 06:54 AM

 Final PDF to printer

92  Chapter Three   Linear Programming: Formulation and Applications

The Decisions
Recall that, as before, the decisions to be made are
TV = Number of commercials on television
M = Number of advertisements in magazines
SS = Number of advertisements in Sunday supplements
The changing cells to hold these numbers continue to be in NumberOfAds (C19:E19).
The Constraints
In addition to the original constraints, we now have two benefit constraints and one fixed-
requirement constraint. As specified in rows 11 and 12, columns F to H, the benefit con-
straints are
Total number of young children reached ≥ 5 (goal 1 in millions)
​​       ​  ​  ​  ​​​
Total number of parents reached ≥5 (goal 2 in millions)
Using the data in columns C to E of these rows,

Total number of young children reached = 1.2TV + 0.1M + 0SS

​ = SUMPRODUCT (C11:E11, NumberOfAds)
​​​     
​       
      
       ​  ​  →​  ​  cell F11​ ​ ​ ​​​
  Total number of parents reached = 0.5TV + 0.2M + 0.2SS
​ = SUMPRODUCT (C12:E12, NumberOfAds)
​ → cell F12

These output cells are given the range name TotalReached (F11:F12).
The fixed-requirement constraint indicated in row 15 is that
​​Total coupon redemption = 1,490​  ​​(​​allotment in $1,000s​)​​​​​
CouponRedemptionPerAd (C15:E15) gives the number of coupons redeemed per ad, so
Total coupon redemption = 0TV + 40M + 120SS
​​     
​        ​  =​  ​  SUMPRODUCT (CouponRedemptionPerAd,
​ NumberOfAds)​​​
​ → cell F15
These same constraints are specified in Solver, along with the original constraints, in
Figure 3.7.
The Measure of Performance
The measure of performance continues to be
Exposure = 1,300TV + 600M + 500SS
​​      
​     
​ ​  =​  ​  SUMPRODUCT (ExposuresPerAd,
​ NumberOfAds)​​​
​ → cell H19
so the objective cell is again TotalExposures (H19).
Summary of the Formulation
The above steps have resulted in formulating the following linear programming model (in
algebraic form) on a spreadsheet.
​​Maximize ​  Exposure = 1,300TV + 600M + 500SS​​
subject to the following constraints:
1. Resource constraints:
300TV + 150M + 100SS ≤ 4,000 ​​(​​ad budget in $1,000s​)​​​
​​      
      
90TV ​  +   30M + ​  ≤​ ​  1,000​ ​ 
40SS​  (​​ ​​planning budget in $1,000s​
​  ​​)​​​​
TV ≤    5 (​​ ​​television spots available​)​​​

hiL18920_ch03_064-123.indd 92 12/08/17 06:54 AM

 Final PDF to printer

An Application Vignette
The Chevron Corporation is one of the world’s leading inte- programming. One application involves optimizing the com-
grated energy companies. It explores extensively for crude bination of refined products (gasoline, jet, diesel fuels) to pro-
oil and natural gas throughout the world. Thanks to its vast duce to maximize the total profit. Another application involves
reserves, it produces nearly 2 million barrels of crude oil per periodically determining the optimal way to run the refining
day and similar amounts of natural gas. It then uses its refiner- processing units when changes occur in crude oil prices, raw
ies to refine and market nearly 3 million barrels per day of trans- material availability, product prices, product specifications, and
portation fuels, chemicals, and lubricants. equipment capabilities. Still another application of linear pro-
Dating nearly back to the invention of linear programming gramming (along with also using decision analysis, the subject
in 1947, Chevron quickly became one of the heaviest users of of Chapter 9) involves optimizing the use of capital for new proj-
this exciting new technique. The earliest applications involved ects to improve its refining system on an ongoing basis.
the following blending problem. Any single grade of gasoline The combination of all these applications of linear program-
needs to be blended from about three to ten components (dif- ming to minimize the total cost or maximize the total profit in
ferent forms of processed crude oil), where no single compo- various ways has had a dramatic impact on Chevron’s bot-
nent meets the quality specifications of the grade of gasoline tom line. The estimated cumulative value to Chevron now
but various combinations of the components can accomplish approaches $1 billion annually. In recognition of this and
this. A typical refinery might have 20 different components related work, the Institute for Operations Research and the
to be blended into four or more grades of gasoline differing Management Sciences (INFORMS) awarded Chevron the presti-
in octane and other properties by marketing area. Linear pro- gious 2015 INFORMS Prize for its long and innovative history in
gramming of the mixed type can achieve huge savings by solv- applying advanced analytics and management science across
ing for how to minimize the total cost of accomplishing all this the company.
blending.
Source: T. Kutz, M. Davis, R. Creek, N. Kenaston, C. Stenstrom, and
As time went on, the exponential growth in computing
M. Connor, Interfaces 44, no. 1 (January-February 2014), pp. 39–54.
power enabled Chevron to greatly expand its use of linear (A link to this article is available at www.mhhe.com/Hillier6e.)

2. Benefit constraints:
1.2TV + 0.1M ≥5 ​​(​​millions of young children​)​​​
​​       ​  ​  ​   ​​​
0.5TV + 0.2M + 0.2SS ≥ 5 (​​ ​​millions of parents​)​​​
3. Fixed-requirement constraint:
​​40M + 120SS = 1,490​  ​​(​​coupon budget in $1,000s​)​​​​​
4. Nonnegativity constraints:
​​ V ≥ 0​ 
T M ≥ 0​  SS ≥ 0​​

Solving the Model

The lower left-hand corner of Figure 3.7 shows the entries needed in Solver, along with the
selection of the usual two options. Solver then finds the optimal solution given in row 19. This
optimal solution provides the following plan for the promotional campaign:
Run 3 television commercials.
Run 14 advertisements in magazines.
Run 7.75 advertisements in Sunday supplements (so the eighth advertisement would
appear in only 75 percent of the newspapers).
Although the expected number of exposures with this plan is only 16,175,000, versus the
17,000,000 with the first plan shown in Figure 3.1, both Claire Syverson and David Sloan feel
that the new plan does a much better job of meeting all of management’s goals for this cam-
paign. They decide to adopt the new plan.
A model may need to be This case study illustrates a common theme in real applications of linear programming—
modified a number of times the continuing evolution of the linear programming model. It is common to make later adjust-
before it adequately incor-
ments in the initial version of the model, perhaps even many times, as experience is gained
porates all the important
considerations. in using the model. Frequently, these adjustments are made to more adequately reflect some
important managerial considerations. This may result in a mixed problem because the new
functional constraints needed to incorporate the managerial considerations may be of a differ-
ent type from those in the original model.
93

hiL18920_ch03_064-123.indd 93 12/08/17 06:54 AM

 Final PDF to printer

94  Chapter Three   Linear Programming: Formulation and Applications

Other Examples
This case study provides a relatively simple example of a small mixed problem. Most of
the mixed problems that arise in practice are much larger, sometimes involving hundreds
or thousands of activities and hundreds or thousands of constraints. At first glance, these
larger problems may seem considerably more complicated than the case study. However,
the important thing to remember is that any linear programming problem can have only
three types of functional constraints—resource constraints, benefit constraints, and fixed-
requirement ­constraints—where each type is formulated just as illustrated above for the
case study.
There are numerous kinds of managerial problems to which linear programming can be
applied. We don’t have nearly enough space available to give examples of all the most impor-
tant kinds of applications. However, if you would like to explore this further, we suggest that
you go through the five solved problems that are summarized in front of the Problems section
for this chapter. Reading the seven cases that follow the Problems section, as well as the appli-
cation vignettes in both this chapter and the preceding chapter, also will further illustrate the
unusually wide applicability of linear programming.
Meanwhile, we soon will turn to two more categories of linear programming problems in
the next two sections.

Summary of the Formulation Procedure for Mixed

Linear Programming Problems
The procedure for formulating mixed problems is similar to the one outlined at the end of
Section 3.2 for resource-allocation problems. However, pure resource-allocation prob-
­
lems only have resource constraints whereas mixed problems can include all three types of
functional constraints (resource constraints, benefit constraints, and fixed-requirement con-
straints). Therefore, the following summary for formulating mixed problems includes separate
steps for dealing with these different types of constraints. Also see Figure 3.8 for a template
of the format for a spreadsheet model of mixed problems. (This format works well for most
mixed problems, including those encountered in this chapter, but more flexibility is occasion-
ally needed, as will be illustrated in the next chapter.)

FIGURE 3.8
A template of a spreadsheet model for mixed problems.
Activities

Unit Profit or Cost Profit/cost per unit of activity

Resources Resources
Used Available

SUMPRODUCT
Resource used per unit of activity (resource used per unit,
changing cells)
≤
Constraints

Benefit Benefit
Achieved Needed

Benefit achieved per unit of activity

SUMPRODUCT
(benefit per unit,
≥
changing cells)
=
Total Profit or Cost
Level of Activity Changing cells SUMPRODUCT(profit/cost per unit, changing cells)

hiL18920_ch03_064-123.indd 94 12/08/17 06:54 AM

 Final PDF to printer

3.5  Transportation Problems   95

1. Since any linear programming problem involves finding the best mix of levels of various
activities, identify these activities for the problem at hand. The decisions to be made are
the levels of these activities.
2. From the viewpoint of management, identify an appropriate overall measure of perfor-
mance for solutions of the problem.
3. For each activity, estimate the contribution per unit of the activity to this overall measure
of performance.
4. Identify any resources that must be allocated to the activities (as described in Section 3.2). For
each one, identify the amount available and then the amount used per unit of each activity.
5. Identify any benefits to be obtained from the activities (as described in Section 3.3). For
each one, identify the minimum acceptable level prescribed by management and then the
benefit contribution per unit of each activity.
6. Identify any fixed requirements that, for some type of quantity, the amount provided must
equal a required amount (as described in Section 3.4). For each fixed requirement, iden-
tify the required amount and then the contribution toward this required amount per unit
of each activity.
7. Enter the data gathered in steps 3–6 into data cells in a spreadsheet.
8. Designate changing cells for displaying the decisions on activity levels.
9. Use output cells to specify the constraints on resources, benefits, and fixed requirements.
10. Designate an objective cell for displaying the overall measure of performance.

Review 1. What types of functional constraints can appear in a mixed linear programming problem?
Questions 2. What managerial goals needed to be incorporated into the expanded linear programming
model for the Super Grain Corp. problem?
3. Which categories of functional constraints are included in the new linear programming model?
4. Why did management adopt the new plan even though it provides a smaller expected number
of exposures than the original plan recommended by the original linear programming model?

3.5 TRANSPORTATION PROBLEMS

One of the most common applications of linear programming involves optimizing a shipping
plan for transporting goods. In a typical application, a company has several plants producing
a certain product that needs to be shipped to the company’s customers (or perhaps to distribu-
tion centers). How much should each plant ship to each customer in order to minimize the
total cost? Linear programming can provide the answer. This type of linear programming
problem is called a transportation problem.
This kind of application normally needs two kinds of functional constraints. One kind
specifies that the amount of the product produced at each plant must equal the total amount
shipped to customers. The other kind specifies that the total amount received from the plants
by each customer must equal the amount ordered. These are fixed-requirement constraints,
which makes the problem a fixed-requirements problem. However, there also are variations of
this problem where resource constraints or benefit constraints are needed.
Transportation problems and assignment problems (described in the next section) are
such important types of linear programming problems that the entire Chapter 15 (available at
www.mhhe.com/Hillier6e) is devoted to further describing these two related types of prob-
lems and providing examples of a wide variety of applications.
We provide below an example of a typical transportation problem.

The Big M Company Transportation Problem

The Big M Company produces a variety of heavy duty machines at two factories. One of its
products is a large turret lathe. Orders have been received from three customers to purchase
some of these turret lathes next month. These lathes will be shipped individually, and Table 3.9
shows what the cost will be for shipping each lathe from each factory to each customer. This

hiL18920_ch03_064-123.indd 95 12/08/17 06:54 AM

 Final PDF to printer

96  Chapter Three   Linear Programming: Formulation and Applications

TABLE 3.9
Shipping Cost for Each Lathe
Some Data for the Big M
Company Distribution- To Customer 1 Customer 2 Customer 3 Output
Network Problem
From
Factory 1 $700 $900 $800 12 lathes
Factory 2 800 900 700 15 lathes
Order size 10 lathes 8 lathes 9 lathes

table also shows how many lathes have been ordered by each customer and how many will be
produced by each factory. The company’s distribution manager now wants to determine how
many machines to ship from each factory to each customer to minimize the total shipping cost.
Figure 3.9 depicts the distribution network for this problem. This network ignores the geo-
graphical layout of the factories and customers and instead lines up the two factories in one
column on the left and the three customers in one column on the right. Each arrow shows one
of the shipping lanes through this distribution network.
Formulation of the Problem in Linear Programming Terms
We need to identify the activities and requirements of this transportation problem to for-
mulate it as a linear programming problem. In this case, two kinds of activities have been
­mentioned—the production of the turret lathes at the two factories and the shipping of these
lathes along the various shipping lanes. However, we know the specific amounts to be pro-
duced at each factory, so no decisions need to be made about the production activities. The
decisions to be made concern the levels of the shipping activities—how many lathes to ship
through each shipping lane. Therefore, we need to focus on the shipping activities for the lin-
ear programming formulation.
The activities correspond to shipping lanes, depicted by arrows in Figure 3.9.
The level of each activity is the number of lathes shipped through the corresponding ship-
ping lane.
Just as any linear programming problem can be described as finding the best mix of activ-
ity levels, this one involves finding the best mix of shipping amounts for the various shipping
lanes. The decisions to be made are
SF1-C1 = Number of lathes shipped from Factory 1 to Customer 1
SF1-C2 = Number of lathes shipped from Factory 1 to Customer 2

FIGURE 3.9 10 lathes

The distribution network C1
needed
for the Big M Company
problem. the
0/la
$70

12 lathes $900/la
F1 the
produced $8
00
/la
the 8 lathes
C2
needed
e
th
/la
00

e
/lath
$8

15 lathes $900
F2
produced

$70
0/la
the

9 lathes
C3
needed

hiL18920_ch03_064-123.indd 96 12/08/17 06:54 AM

 Final PDF to printer

3.5  Transportation Problems   97

SF1-C3 = Number of lathes shipped from Factory 1 to Customer 3

SF2-C1 = Number of lathes shipped from Factory 2 to Customer 1
SF2-C2 = Number of lathes shipped from Factory 2 to Customer 2
SF2-C3 = Number of lathes shipped from Factory 2 to Customer 3
so six changing cells will be needed in the spreadsheet.
The objective is to
​​Minimize ​  Cost = Total cost for shipping the lathes​​
Using the shipping costs given in Table 3.9,
​Cost = 700 ​S​  F1-C1​​ + 900 ​S​  F1-C2​​ + 800 ​S​  F1-C3​​ + 800 ​S​  F2-C1​​ + 900 ​S​  F2-C2​​ + 700 ​S​  F2-C3​​​
is the quantity in dollars to be entered into the objective cell. (We will use a SUMPRODUCT
function to do this a little later.)
The spreadsheet model also will need five constraints involving fixed requirements. Both
Table 3.9 and Figure 3.9 show these requirements.
Requirement 1: Factory 1 must ship 12 lathes.
Requirement 2: Factory 2 must ship 15 lathes.
Requirement 3: Customer 1 must receive 10 lathes.
Requirement 4: Customer 2 must receive 8 lathes.
Requirement 5: Customer 3 must receive 9 lathes.
Thus, there is a specific requirement associated with each of the five locations in the distribu-
tion network shown in Figure 3.9.
All five of these requirements can be expressed in constraint form as
​Amount provided = Required amount​
For example, Requirement 1 can be expressed algebraically as
​​S​  F1-C1​​ + ​S​  F1-C2​​ + S​ ​  F1-C3​​ = 12​
where the left-hand side gives the total number of lathes shipped from Factory 1, and 12 is
the required amount to be shipped from Factory 1. Therefore, this constraint restricts SF1-C1,
SF1-C2, and SF1-C3 to values that sum to the required amount of 12. In contrast to the ≤ form
for resource constraints and the ≥ form for benefit constraints, the constraints express fixed
requirements that must hold with equality, so this transportation problem falls into the category
of fixed-requirements problems introduced in the preceding section. However, Chapter 15
(available at www.mhhe.com/Hillier6e) gives several examples that illustrate how variants of
transportation problems can have resource constraints or benefit constraints as well. For exam-
ple, if 12 lathes represent the manufacturing capacity of Factory 1 (the maximum number that
can be shipped) rather than a requirement for how many must be shipped, the constraint just
given for Requirement 1 would become a ≤ resource constraint instead. Such variations can
be incorporated readily into the spreadsheet model.
Formulation of the Spreadsheet Model
Careful problem formulation In preparation for formulating the model, the problem has been formulated above by identifying
needs to precede model the decisions to be made, the constraints on these decisions, and the overall measure of perfor-
formulation. mance, as well as gathering all the important data displayed in Table 3.9. All this information
leads to the spreadsheet model shown in Figure 3.10. The data cells include ShippingCost
(C5:E6), Output (H11:H12), and OrderSize (C15:E15), incorporating all the data from Table
3.9. The changing cells are UnitsShipped (C11:E12), which give the decisions on the amounts
to be shipped through the respective shipping lanes. The output cells are TotalShippedOut
Here is an example where (F11:F12) and TotalToCustomer (C13:E13), where the SUM functions entered into these
SUM functions are used cells are shown below the spreadsheet in Figure 3.10. The constraints are that TotalShipped-
for output cells instead of Out is required to equal Output and TotalToCustomer is required to equal OrderSize. These
SUMPRODUCT functions.
constraints have been specified on the spreadsheet and entered into Solver. The objective

hiL18920_ch03_064-123.indd 97 12/08/17 06:54 AM

 Final PDF to printer

98  Chapter Three   Linear Programming: Formulation and Applications

cell is TotalCost (H15), where its SUMPRODUCT function gives the total shipping cost.
The lower left-hand corner of Figure 3.10 shows the entries needed in Solver, along with the
selection of the usual two options.
The layout of the spreadsheet is different than for all the prior linear programming examples
in the book. Rather than a separate column for each activity and a separate row for each con-
straint, the cost data and changing cells are laid out in a table format. This format provides a
more natural and compact way of displaying the constraints and results.
UnitsShipped (C11:E12) in the spreadsheet in Figure 3.10 shows the result of applying
Solver to obtain an optimal solution for the number of lathes to ship through each shipping
lane. TotalCost (H15) indicates that the total shipping cost for this shipping plan is $20,500.
Since any transportation problem is a special type of linear programming problem, it makes
the standard assumption that fractional solutions are allowed. However, we actually don’t
want this assumption for this particular application since only integer numbers of lathes can
be shipped from a factory to a customer. Fortunately, even while making the standard assump-
tion, the numbers in the optimal solution shown in UnitsShipped (C11:E12) only have integer
values. This is no coincidence. Because of the form of its model, almost any transportation

FIGURE 3.10
The spreadsheet model for the Big M Company problem, including the formulas for the objective cell TotalCost (H15) and the other
output cells TotalShippedOut (F11:F12) and TotalToCustomer (C13:E13), as well as the specifications needed to set up Solver.
The changing cells UnitsShipped (C11:E12) show the optimal solution obtained by Solver.

A B C D E F G H
1 Big M Company Distribution Problem
2
3 Shipping Cost
4 (per Lathe) Customer 1 Customer 2 Customer 3
5 Factory 1 $700 $900 $800
6 Factory 2 $800 $900 $700
7
8 Total
9 Shipped
10 Units Shipped Customer 1 Customer 2 Customer 3 Out Output
11 Factory 1 10 2 0 12 = 12
12 Factory 2 0 6 9 15 = 15
13 Total to Customer 10 8 9
14 = = = Total Cost
15 Order Size 10 8 9 $20,500

Solver Parameters F
Range Name Cells
Set Objective Cell: TotalCost
To: Min 8 Total
OrderSize C15:E15
By Changing Variable Cells: 9
Output H11:H12 Shipped
UnitsShipped
Subject to the Constraints: ShippingCost C5:E6 10 Out
TotalShippedOut = Output TotalCost H15
11 =SUM(C11:E11)
TotalToCustomer = OrderSize TotalShippedOut F11:F12
TotalToCustomer C13:E13 12 =SUM(C12:E12)
Solver Options: UnitsShipped C11:E12
Make Variables Nonnegative
Solving Method: Simplex LP B C D E
13 Total to Customer =SUM(C11:C12) =SUM(D11:D12) =SUM(E11:E12)

H
14 Total Cost
15 =SUMPRODUCT(ShippingCost,UnitsShipped)

hiL18920_ch03_064-123.indd 98 12/08/17 06:54 AM

 Final PDF to printer

3.6  Assignment Problems   99

problem (including this one) is guaranteed in advance to have an optimal solution that has
only integer values despite the fact that fractional solutions also are allowed. In particular, as
long as the data for the problem includes only integer values for all the supplies and demands
(which are the outputs and order sizes in the Big M Company problem), any transportation
problem with feasible solutions is guaranteed to have an optimal solution with integer values
for all its decision variables. Therefore, it is not necessary to add constraints to the model that
require these variables to have only integer values.
To summarize, here is the algebraic form of the linear programming model that has been
formulated in the spreadsheet:
Minimize   Cost = 700 SF1-C1 + 900 SF1-C2 + 800 SF1-C3 + 800 SF2-C1
   + 900 SF2-C2 + 700 SF2-C3
subject to the following constraints:

1. Fixed-requirement constraints:

SF1-C1 + SF1-C2 + SF1-C3 = 12   (Factory 1)

SF2-C1 + SF2-C2 + SF2-C3 = 15   (Factory 2)
SF1-C1 + SF2-C1 = 10   (Customer 1)
SF1-C2 + SF2-C2 = 8   (Customer 2)
SF1-C3 + SF2-C3 = 9   (Customer 3)

2. Nonnegativity constraints:

SF1-C1 ≥ 0   SF1-C2 ≥ 0   SF1-C3 ≥ 0   SF2-C1 ≥ 0   SF2-C2 ≥ 0   SF2-C3 ≥ 0

Review 1. Why are transportation problems given this name?

Questions 2. What is an identifying feature of transportation problems?
3. How does the form of a fixed-requirement constraint differ from that of a resource constraint?
A benefit constraint?
4. What are the quantities with fixed requirements in the Big M Company problem?

3.6 ASSIGNMENT PROBLEMS

We now turn to another special type of linear programming problem called assignment
problems. As the name suggests, this kind of problem involves making assignments. Fre-
quently, these are assignments of people to jobs. Thus, many applications of the assignment
problem involve aiding managers in matching up their personnel with tasks to be performed.
Other applications might instead involve assigning machines, vehicles, or plants to tasks.
Here is a typical example.

An Example: The Sellmore Company Problem

The marketing manager of the Sellmore Company will be holding the company’s annual
sales conference soon for sales regional managers and personnel. To assist in the administra-
tion of the conference, he is hiring four temporary employees (Ann, Ian, Joan, and Sean),
where each will handle one of the following four tasks:
1. Word processing of written presentations.
2. Computer graphics for both oral and written presentations.
3. Preparation of conference packets, including copying and organizing written materials.
4. Handling of advance and on-site registrations for the conference.
He now needs to decide which person to assign to each task.

hiL18920_ch03_064-123.indd 99 12/08/17 06:54 AM

 Final PDF to printer

100  Chapter Three   Linear Programming: Formulation and Applications

TABLE 3.10
Required Time per Task (Hours)
Data for the Sellmore
Co. Problem Temporary Word Hourly
Employee Processing Graphics Packets Registrations Wage
Ann 35 41 27 40 $14
Ian 47 45 32 51 12
Joan 39 56 36 43 13
Sean 32 51 25 46 15

Decisions need to be made Although each temporary employee has at least the minimal background necessary to perform
regarding which person to any of the four tasks, they differ considerably in how efficiently they can handle the different types
assign to each task.
of work. Table 3.10 shows how many hours each would need for each task. The rightmost
column gives the hourly wage based on the background of each employee.
Formulation of a Spreadsheet Model
Using Cost (D15:G18), Figure 3.11 shows a spreadsheet model for this problem. Table 3.10 is entered at the top. Combin-
the objective is to mini- ing these required times and wages gives the cost (cells D15:G18) for each possible assignment
mize the total cost of the
of a temporary employee to a task, using equations shown at the bottom of Figure 3.11. This
assignments.
cost table is just the way that any assignment problem is displayed. The objective is to determine
which assignments should be made to minimize the sum of the associated costs.
A value of 1 in a chang- The values of 1 in Supply (J24:J27) indicate that each person (assignee) listed in column
ing cell indicates that the C must perform exactly one task. The values of 1 in Demand (D30:G30) indicate that each
corresponding assignment
is being made, whereas 0 task must be performed by exactly one person. These requirements then are specified in the
means that the assignment constraints given in Solver.
is not being made. Each of the changing cells Assignment (D24:G27) is given a value of 1 when the corre-
sponding assignment is being made, and a value of 0 otherwise. Therefore, the Excel equation
Excel Tip: When solving an
for the objective cell, TotalCost = SUMPRODUCT(Cost, Assignment), gives the total cost for
assignment problem, round- the assignments being made. The Solver Parameters box specifies that the goal is to minimize
ing errors occasionally will this objective cell.
cause Excel to return a non- The changing cells in Figure 3.11 show the optimal solution obtained after running Solver.
integer value very close to This solution is
0 (e.g., 1.23 E-10, meaning
0.000000000123) or very Assign Ann to prepare conference packets.
close to 1 (e.g., 0.9999912).
To make the spreadsheet
Assign Ian to do the computer graphics.
cleaner, you may replace Assign Joan to handle registrations.
these “ugly” representations Assign Sean to do the word processing.
by their proper value of 0 or
1 in the changing cells. The total cost given in cell J30 is $1,957.

Characteristics of Assignment Problems

Note that all the functional constraints of the Sellmore Co. problem (as shown in cells H24:J27
and D28:G30 of Figure 3.11) are fixed-requirement constraints which require each person to
perform exactly one task and require each task to be performed by exactly one person. Thus,
like the Big M Company transportation problem, the Sellmore Co. is a fixed-requirements
problem. This is a characteristic of all pure assignment problems. However, Chapter 15 (avail-
able at www.mhhe.com/Hillier6e) gives some examples of variants of assignment problems
where this is not the case.
Like the changing cells Assignment (D24:G27) in Figure 3.11, the changing cells in
the spreadsheet model for any pure assignment problem gives a value of 1 when the corre-
sponding assignment is being made and a value of 0 otherwise. Since the fixed-requirement
constraints require only each row or column of changing cells to add up to 1 (which could
happen, e.g., if two of the changing cells in the same row or column had a value of 0.5 and
the rest 0), this would seem to necessitate adding the constraints that each of the changing
cells must be integer. After choosing the Solver option to make the changing cells nonnega-
tive, this then would force each of the changing cells to be 0 or 1. However, it turned out to
be unnecessary to add the constraints that require the changing cells to have values of 0 or 1

hiL18920_ch03_064-123.indd 100 12/08/17 06:54 AM

 Final PDF to printer

3.6  Assignment Problems   101

FIGURE 3.11
A spreadsheet formulation of the Sellmore Co. problem as an assignment problem, including the objective cell TotalCost (J30) and
the other output cells Cost (D15:G18), TotalAssignments (H24:H27), and TotalAssigned (D28:G28), as well as the specifications
needed to set up the model. The values of 1 in the changing cells Assignment (D24:G27) show the optimal plan obtained by Solver
for assigning the people to the tasks.
A B C D E F G H I J
1 Sellmore Co. Assignment Problem
2
3 Task
4 Required Time Word Hourly
5 (Hours) Processing Graphics Packets Registrations Wage
6 Ann 35 41 27 40 $14
7 Assignee Ian 47 45 32 51 $12
8 Joan 39 56 36 43 $13
9 Sean 32 51 25 46 $15
10
11
12 Task
13 Word
14 Cost Processing Graphics Packets Registrations
15 Ann $490 $574 $378 $560
16 Assignee Ian $564 $540 $384 $612
17 Joan $507 $728 $468 $559
18 Sean $480 $765 $375 $690
19
20
21 Task
22 Assignment Word Total
23 Processing Graphics Packets Registrations Assignments Supply
24 Ann 0 0 1 0 1 = 1
25 Assignee Ian 0 1 0 0 1 = 1
26 Joan 0 0 0 1 1 = 1
27 Sean 1 0 0 0 1 = 1
28 Total Assigned 1 1 1 1
29 = = = = Total Cost
30 Demand 1 1 1 1 $1,957
B C D E F G H
13 Word 22 Total
14 Cost Processing Graphics Packets Registrations 23 Assignments

15 Ann =D6*I6 =E6*I6 =F6*I6 =G6*I6 24 =SUM(D24:G24)

16 Assignee Ian =D7*I7 =E7*I7 =F7*I7 =G7*I7 25 =SUM(D25:G25)

17 Joan =D8*I8 =E8*I8 =F8*I8 =G8*I8 26 =SUM(D26:G26)

18 Sean =D9*I9 =E9*I9 =F9*I9 =G9*I9 27 =SUM(D27:G27)

Solver Parameters
Range Name Cells
Set Objective Cell: TotalCost
To: Min Assignment D24:G27
By Changing Variable Cells: Cost D15:G18
Assignment J Demand D30:G30
Subject to the Constraints: 29 Total Cost HourlyWage I6:I9
TotalAssigned = Demand RequiredTime D6:G9
TotalAssignments = Supply 30 =SUMPRODUCT(Cost,Assignment) Supply J24:J27
TotalAssigned D28:G28
Solver Options: TotalAssignments H24:H27
Make Variables Nonnegative TotalCost J30
Solving Method: Simplex LP

C D E F G
28 Total Assigned =SUM(D24:D27) =SUM(E24:E27) =SUM(F24:F27) =SUM(G24:G27)

hiL18920_ch03_064-123.indd 101 12/08/17 06:54 AM

 Final PDF to printer

An Application Vignette
Taylor Communications (formerly the Standard Operations to presses. The cost of each individual assignment is the sum of
Company before its acquisition and rebranding by the Taylor the cost of producing that job on that press and the cost of trans-
Corporation in 2016) is one of the largest printing conglomer- porting the finished job to the customer. The objective is to select
ates in the world. It provides high volume print production of all of the assignments of jobs to presses so as to minimize the total
forms and print stationery for major firms in and near the United cost of all the assignments. This is in fact an assignment problem
States. It operates dozens of manufacturing facilities across the as described in the current section except for the modification
United States and Mexico. that a given press might be able to handle multiple jobs. There-
The print market is highly competitive, so company manage- fore, a constraint is added for each press that the total process-
ment must continually address the strategic challenge of minimiz- ing time for all of the jobs being assigned to that press cannot
ing the total cost of producing and distributing its products in order exceed the total processing time that is available at that press.
to offer competitive pricing. A wide variety of printing presses are Periodically solving this very large assignment problem and
available in the various manufacturing facilities and each press has implementing this solution has provided an estimated annual
the capacity to perform multiple print jobs. However, the cost of savings of over $10 million.
performing any given job can differ greatly from one press to the
next. Furthermore, the cost of transporting the finished product to
the customer can vary greatly depending on which manufacturing Source: S. L. Ahire, M. F. Gorman, D. Dwiggins, and O. Mudry, “Opera-
tions Research Helps Reshape Operations Strategy at Standard
facility is used and the resulting distance to the customer. Register Company,” Interfaces 37, no. 6 (November-December
Given a large number of print jobs that need to be performed 2007), pp.553–565. (A link to this article is available at
over the next time period, the problem is to assign the print jobs www.mhhe.com/Hillier6e.)

in Figure 3.11 because Solver gave an optimal solution that had only values of 0 or 1 anyway.
In fact, a general characteristic of pure assignment problems is that Solver always provides
such an optimal solution without needing to add these additional constraints.
As described further in Chapter 15, available at www.mhhe.com/Hillier6e, another inter-
esting characteristic of any pure assignment problem is that it can be viewed as a special type
of pure transportation problem. In particular, every fixed-requirement constraint in the corre-
sponding transportation problem would require that either a row or column of changing cells
add up to 1. This would result in Solver giving an optimal solution where every changing cell
has a value of either 0 or 1, just as for the original assignment problem.

Review 1. Why are assignment problems given this name?

Questions 2. Pure assignment problems have what type of functional constraints?
3. What is the interpretation of the changing cells in the spreadsheet model of a pure assign-
ment problem?

3.7 MODEL FORMULATION FROM A BROADER PERSPECTIVE

Formulating and analyzing a linear programming model provides information to help manag-
ers make their decisions. That means the model must accurately reflect the managerial view
of the problem:
Both the measure of perfor- • The overall measure of performance must capture what management wants accomplished.
mance and the constraints • When management limits the amounts of resources that will be made available to the activ-
in a model need to reflect
the managerial view of the ities under consideration, these limitations should be expressed as resource constraints.
problem. • When management establishes minimum acceptable levels for benefits to be gained from
the activities, these managerial goals should be incorporated into the model as benefit
constraints.
• If management has fixed requirements for certain quantities, then fixed-requirement con-
straints are needed.
With the help of spreadsheets, some managers now are able to formulate and solve small
linear programming models themselves. However, larger linear programming models may be
formulated by management science teams, not managers. When this is done, the management
102

hiL18920_ch03_064-123.indd 102 12/08/17 06:54 AM

 Final PDF to printer

3.7  Model Formulation from a Broader Perspective   103

Linear programming stud- science team must thoroughly understand the managerial view of the problem. This requires clear
ies need strong managerial communication with management from the very beginning of the study and maintaining effective
input and support.
communication as new issues requiring managerial guidance are identified. Management needs
to clearly convey its view of the problem and the important issues involved. A manager cannot
expect to obtain a helpful linear programming study without making clear just what help is
wanted.
As is necessary in any textbook, the examples in this chapter are far smaller, simpler, and
more clearly spelled out than is typical of real applications. Many real studies require formu-
lating complicated linear programming models involving hundreds or thousands (or possibly
even millions) of decisions and constraints. In these cases, there usually are many ambiguities
about just what should be incorporated into the model. Strong managerial input and support
are vital to the success of a linear programming study for such complex problems.
When dealing with huge real problems, there is no such thing as “the” correct linear pro-
gramming model for the problem. The model continually evolves throughout the course of the
study. Early in the study, various techniques are used to test initial versions of the model to
identify the errors and omissions that inevitably occur when constructing such a large model.
This testing process is referred to as model validation.
Once the basic formulation has been validated, there are many reasonable variations of the
model that might be used. Which variation to use depends on such factors as the assumptions
about the problem that seem most reasonable, the estimates of the parameters of the model
that seem most reliable, and the degree of detail desired in the model.
In large linear programming studies, a good approach is to begin with a relatively simple
version of the model and then use the experience gained with this model to evolve toward
more elaborate models that more nearly reflect the complexity of the real problem. This pro-
cess of model enrichment continues only as long as the model remains reasonably easy to
solve. It must be curtailed when the study’s results are needed by management. Managers
often need to curb the natural instinct of management science teams to continue adding “bells
and whistles” to the model rather than winding up the study in a timely fashion with a less
elegant but adequate model.
When managers study the output of the current model, they often detect some undesirable
characteristics that point toward needed model enrichments. These enrichments frequently
take the form of new benefit constraints to satisfy some managerial goals not previously artic-
ulated. (Recall that this is what happened in the Super Grain case study.)
What-if analysis addresses Even though many reasonable variations of the model could be used, an optimal solution can
some key questions that be solved for only with respect to one specific version of the model at a time. This is why what-
remain after formulating if analysis is such an important part of a linear programming study. After obtaining an optimal
and solving a model.
solution with respect to one specific model, management will have many what-if questions:
• What if the estimates of the parameters in the model are incorrect?
• How do the conclusions change if different plausible assumptions are made about the problem?
• What happens when certain managerial options are pursued that are not incorporated into
the current model?
Chapter 5 is devoted primarily to describing how what-if analysis addresses these and related
issues, as well as how managers use this information.
Because managers instigate management science studies, they need to know enough about
linear programming models and their formulation to be able to recognize managerial problems to
which linear programming can be applied. Furthermore, since managerial input is so impor-
tant for linear programming studies, managers need to understand the kinds of managerial
concerns that can be incorporated into the model. Developing these two skills have been the
most important goals of this chapter.

Review 1. A linear programming model needs to reflect accurately whose view of the problem?
Questions 2. What is meant by model validation?
3. What is meant by the process of model enrichment?
4. Why is what-if analysis an important part of a linear programming study?

hiL18920_ch03_064-123.indd 103 12/08/17 06:54 AM

 Final PDF to printer

104  Chapter Three   Linear Programming: Formulation and Applications

3.8 Summary Functional constraints with a ≤ sign are called resource constraints, because they require that the amount
used of some resource must be less than or equal to the amount available of that resource. The identify-
ing feature of resource-allocation problems is that all their functional constraints are resource constraints.
Functional constraints with a ≥ sign are called benefit constraints, since their form is that the level
achieved for some benefit must be greater than or equal to the minimum acceptable level for that ben-
efit. Frequently, benefit constraints express goals prescribed by management. If every functional con-
straint is a benefit constraint, then the problem is a cost–benefit–trade-off problem.
Functional constraints with an = sign are called fixed-requirement constraints, because they express the
fixed requirement that, for some quantity, the amount provided must be equal to the required amount. The
identifying feature of fixed-requirements problems is that their functional constraints are fixed-requirement
constraints. One prominent type of fixed-requirements problem is transportation problems, which typically
involve finding a shipping plan that minimizes the total cost of transporting a product from a number of
plants to a number of customers. Another prominent type is assignment problems, which typically involves
assigning people to tasks so as to minimize the total cost of performing these tasks.
Linear programming problems that do not fit into any of these three categories are called mixed problems.
In many real applications, management science teams formulate and analyze large linear program-
ming models to help guide managerial decision making. Such teams need strong managerial input and
support to help ensure that their work really meets management’s needs.

Glossary assignment problem   A type of linear program- integer programming problem   A variation of a
ming problem that typically involves assigning linear programming problem that has the additional
people to tasks so as to minimize the total cost of restriction that some or all of the decision variables
performing these tasks. (Section 3.6), 99 must have integer values. (Section 3.2), 75
benefit constraint   A functional constraint mixed problem   Any linear programming prob-
with a ≥ sign. The left-hand side is interpreted as lem that includes at least two of the three types
the level of some benefit that is achieved by the of functional constraints (resource constraints,
activities under consideration, and the right-hand benefit constraints, and fixed-requirement con-
side is the minimum acceptable level for that ben- straints). (Section 3.4), 88
efit. (Section 3.3), 82 model enrichment   The process of using expe-
cost–benefit–trade-off problem   A type of rience with a model to identify and add important
linear programming problem involving the details that will provide a better representation of
trade-off between the total cost of the activi- the real problem. (Section 3.7), 103
ties under consideration and the benefits to be model validation   The process of checking and
achieved by these activities. Its identifying fea- testing a model to develop a valid model. (Section
ture is that each functional constraint in the lin- 3.7), 103
ear programming model is a benefit constraint. resource-allocation problem   A type of linear
(Section 3.3), 82 programming problem concerned with allocating
fixed-requirement constraint  A functional resources to activities. Its identifying feature is
constraint with an = sign. The left-hand side rep- that each functional constraint in its model is a
resents the amount provided of some type of quan- resource constraint. (Section 3.2), 71
tity, and the right-hand side represents the required resource constraint   A functional constraint
amount for that quantity. (Section 3.4), 88 with a ≤ sign. The left-hand side represents the
fixed-requirements problem   A type of linear amount of some resource that is used by the
programming problem concerned with optimiz- activities under consideration, and the right-
ing how to meet a number of fixed requirements. hand side represents the amount available of that
Its identifying feature is that each functional resource. (Section 3.2), 71
constraint in its model is a fixed-requirement con- transportation problem   A type of linear pro-
straint. (Section 3.4), 88 gramming problem that typically involves finding
identifying feature   A feature of a model that a shipping plan that minimizes the total cost of
identifies the category of linear programming transporting a product from a number of plants to
problem it represents. (Chapter introduction), 64 a number of customers. (Section 3.5), 95

Learning Aids for This Chapter

All learning aids are available at www.mhhe.com/Hillier6e.
Excel Files: Big M Example
Super Grain Example Revised Super Grain Example
TBA Airlines Example Sellmore Example
Think-Big Example Excel Add-in:
Union Airways Example
Analytic Solver

hiL18920_ch03_064-123.indd 104 12/08/17 06:54 AM

 Final PDF to printer

Chapter 3   Solved Problems   105

Solved Problems
The solutions are available at www.mhhe.com/Hillier6e.

3.S1. Farm Management 12-inch rods (with no waste). For the next production period,
Dwight and Hattie have run the family farm for over 30 years. Decora needs twenty-five 12-inch rods, fifty-two 18-inch rods,
They are currently planning the mix of crops to plant on their forty-five 24-inch rods, thirty 40-inch rods, and twelve 60-inch
120-acre farm for the upcoming season. The table gives the rods. What is the fewest number of 60-inch rods that can be
labor-hours and fertilizer required per acre, as well as the total purchased to meet their production needs? Formulate and solve
expected profit per acre for each of the potential crops under an integer programming model in a spreadsheet.
consideration. Dwight, Hattie, and their children can work at
3.S4. Producing and Distributing AEDs at Heart Start
most 6,500 total hours during the upcoming season. They have
Heart Start produces automated external defibrillators in each
200 tons of fertilizer available. What mix of crops should be
of two different plants (A and B). The unit production costs
planted to maximize the family’s total profit?
and monthly production capacity of the two plants are indi-
a. Formulate and solve a linear programming model for this cated in the table below. The automated external defibrillators
problem in a spreadsheet. are sold through three wholesalers. The shipping cost from
b. Formulate this same model algebraically. each plant to the warehouse of each wholesaler along with the
monthly demand from each wholesaler are also indicated in
Labor Fertilizer the table. The management of Heart Start now has asked their
Required Required Expected top management scientist (you) to address the following two
(hours per (tons per Profit (per questions. How many automated external defibrillators should
Crop acre) acre) acre) be produced in each plant, and how should they be distributed
to each of the three wholesaler warehouses so as to minimize
Oats 50 1.5 $500 the combined cost of production and shipping? Formulate and
Wheat 60 2 $600 solve a linear programming model in a spreadsheet.
Corn 105 4 $950

Unit Shipping Cost

3.S2. Diet Problem Unit Monthly
The kitchen manager for Sing Sing prison is trying to decide Ware- Ware- Ware- Produc- Produc-
what to feed its prisoners. She would like to offer some com- house house house tion tion
bination of milk, beans, and oranges. The goal is to minimize 1 2 3 Cost Capacity
cost, subject to meeting the minimum nutritional requirements
imposed by law. The cost and nutritional content of each food, Plant A $22 $14 $30 $600 100
along with the minimum nutritional requirements, are shown Plant B $16 $20 $24 $625 120
below. What diet should be fed to each prisoner? Monthly
Demand 80 60 70
a. Formulate and solve a linear programming model for this
problem in a spreadsheet.
b. Formulate this same model algebraically. 3.S5. Bidding for Classes
In the MBA program at a prestigious university in the Pacific
Northwest, students bid for electives in the second year of their
Oranges
program. Each student has 100 points to bid (total) and must take
Navy (large Minimum
two electives. There are four electives available: Management
Milk Beans Calif. Daily
Science (MS), Finance (Fin), Operations Management (OM), and
(gallons) (cups) Valencia) Requirement
Marketing (Mkt). Each class is limited to 5 students. The bids sub-
Niacin (mg) 3.2 4.9 0.8 13.0 mitted for each of the 10 students are shown in the table below.
Thiamin (mg) 1.12 1.3 0.19 1.5
Vitamin C
(mg) 32.0 0.0 93.0 45.0 Student Bids for Classes
Cost ($) 2.00 0.20 0.25 Student MS Fin OM Mkt

3.S3. Cutting Stock Problem George 60 10 10 20

Fred 20 20 40 20
Decora Accessories manufactures a variety of bathroom acces-
Ann 45 45 5 5
sories, including decorative towel rods and shower curtain
Eric 50 20 5 25
rods. Each of the accessories includes a rod made out of stain- Susan 30 30 30 10
less steel. However, many different lengths are needed: 12, 18, Liz 50 50 0 0
24, 40, and 60 inches. Decora purchases 60-inch rods from Ed 70 20 10 0
an outside supplier and then cuts the rods as needed for their David 25 25 35 15
products. Each 60-inch rod can be used to make a number of Tony 35 15 35 15
smaller rods. For example, a 60-inch rod could be used to make Jennifer 60 10 10 20
a 40-inch and an 18-inch rod (with 2 inches of waste), or five

hiL18920_ch03_064-123.indd 105 12/08/17 06:54 AM

 Final PDF to printer

106  Chapter Three   Linear Programming: Formulation and Applications

a. Formulate and solve a spreadsheet model to determine an as- b. Does the resulting solution seem like a fair assignment?
signment of students to classes so as to maximize the total bid c. Which alternative objectives might lead to a fairer assignment?
points of the assignments.

Problems
An asterisk on the problem number indicates that at least a par- b. Use the spreadsheet to check the following solu-
tial answer is given in the back of the book. tions: (x1, x2) = (2, 2), (3, 3), (2, 4), (4, 2), (3, 4),
(4, 3). Which of these solutions are feasible? Which
3.1. Reconsider the Super Grain Corp. case study as pre- of these feasible solutions has the best value of the
sented in Section 3.1. The advertising firm, Giacomi & Jackowitz, objective function?
now has suggested a fourth promising advertising medium—radio
c. Use Solver to find an optimal solution.
commercials—to promote the company’s new breakfast cereal,
Crunchy Start. Young children are potentially major consumers of d. Express this model in algebraic form.
this cereal, but parents of young children (the major potential pur- e. Use the graphical method to solve this model.
chasers) often are too busy to do much reading (so may miss the 3.4. Consider a resource-allocation problem having the fol-
company’s advertisements in magazines and Sunday supplements) lowing data.
or even to watch the Saturday morning programs for children
where the company’s television commercials are aired. However, Resource Usage
these parents do tend to listen to the radio during the commute to per Unit of
and from work. Therefore, to better reach these parents, Giacomi Each Activity
& Jackowitz suggests giving consideration to running commer-
cials for Crunchy Start on nationally syndicated radio programs Amount of
that appeal to young adults during typical commuting hours. Resource
Giacomi & Jackowitz estimates that the cost of developing each Resource 1 2 3 Available
new radio commercial would be $50,000, and that the expected A 30 20 0 500
number of exposures per commercial would be 900,000. The firm B 0 10 40 600
has determined that 10 spots are available for different radio com- C 20 20 30 1,000
mercials, and each one would cost $200,000 for a normal run. Contribution per unit $50 $40 $70
a. Formulate and solve a spreadsheet model for the Contribution per unit = profit per unit of the activity.
revised advertising-mix problem that includes this
fourth advertising medium. Identify the data cells, a. Formulate and solve a linear programming model
the changing cells, and the objective cell. Also show for this problem on a spreadsheet.
the Excel equation for each output cell expressed as
b. Express this model in algebraic form.
a SUMPRODUCT function.
b. Indicate why this spreadsheet model is a linear pro- 3.5. Consider a resource-allocation problem having the fol-
gramming model. lowing data.
c. Express this model in algebraic form.
3.2. Read the referenced article that fully describes the man- Resource Usage per
agement science study summarized in the application vignette Unit of Each Activity
presented in Section 3.2. Briefly describe how linear program-
Amount of
ming was applied in this study. Then list the various benefits that
Resource
resulted from this study.
Resource 1 2 3 4 Available
3.3.* Consider a resource-allocation problem having the fol-
lowing data. P 3 5 −2 4 400
Q 4 −1 3 2 300
Resource Usage R 6 3 2 −1 400
per Unit of S −2 2 5 3 300
Each Activity Contribution $11 $9 $8 $9
per unit
Amount of
Resource 1 2 Resource Available Contribution per unit = profit per unit of the activity.

1 2 1 10
a. Formulate a linear programming model for this
2 3 3 20
3 2 4 20 problem on a spreadsheet.
Contribution $20 $30 b. Make five guesses of your own choosing for the
per unit optimal solution. Use the spreadsheet to check each
one for feasibility and, if feasible, for the value of
Contribution per unit = profit per unit of the activity.
the objective function. Which feasible guess has the
a. Formulate a linear programming model for this best objective function value?
problem on a spreadsheet. c. Use Solver to find an optimal solution.

hiL18920_ch03_064-123.indd 106 12/08/17 06:54 AM

 Final PDF to printer

Chapter 3   Problems  107

3.6.* The Omega Manufacturing Company has discontinued Each machine is available 40 hours per month. Each part
the production of a certain unprofitable product line. This act manufactured will yield a unit profit as follows:
created considerable excess production capacity. Management
is considering devoting this excess capacity to one or more of Part
three products, Products 1, 2, and 3. The available capacity of A B C
the machines that might limit output is summarized in the fol-
lowing table. Profit $50 $40 $30

Ed wants to determine the mix of spare parts to produce to

Available Time maximize total profit.
Machine Type (in Machine-Hours per Week)
a. Identify both the activities and the resources for this
Milling machine 500 resource-allocation problem.
Lathe 350 b. Formulate a linear programming model for this
Grinder 150 problem on a spreadsheet.
c. Make three guesses of your own choosing for the
The number of machine-hours required for each unit of the optimal solution. Use the spreadsheet to check each
respective products are shown in the next table. one for feasibility and, if feasible, for the value of
the objective function. Which feasible guess has the
Productivity Coefficient (in Machine-Hours per Unit) best objective function value?
d. Use Solver to find an optimal solution.
Machine Type Product 1 Product 2 Product 3 e. Express the model in algebraic form.
Milling machine 9 3 5 3.8. Consider the following algebraic formulation of a
Lathe 5 4 0 resource-allocation problem with three resources, where the deci-
Grinder 3 0 2 sions to be made are the levels of three activities (A1, A2, and A3).
Maximize Profit = 20 A1 + 40 A2 + 30 A3

The Sales Department indicates that the sales potential for subject to
Products 1 and 2 exceeds the maximum production rate and that Resource 1: 3 A1 + 5 A2 + 4 A3 ≤ 400 (amount available)
the sales potential for product 3 is 20 units per week. The unit Resource 2: A1 + A2 + A3 ≤ 100 (amount available)
profit would be $50, $20, and $25, respectively, for Products 1, Resource 3: A1 + 3 A2 + 2 A3 ≤ 200 (amount available)
2, and 3. The objective is to determine how much of each prod-
uct Omega should produce to maximize profit. and
a. Indicate why this is a resource-allocation problem A1 ≥ 0 A2 ≥ 0 A3 ≥ 0
by identifying both the activities and the limited
Formulate and solve the spreadsheet model for this problem.
resources to be allocated to these activities.
3.9. Consider a cost–benefit–trade-off problem having the
b. Identify verbally the decisions to be made, the con-
following data.
straints on these decisions, and the overall measure
of performance for the decisions.
c. Convert these verbal descriptions of the constraints Benefit Contribu-
and the measure of performance into quantitative tion per Unit of Each
expressions in terms of the data and decisions. Activity
d. Formulate a spreadsheet model for this problem. Iden- Minimum
tify the data cells, the changing cells, the objective cell, Accept-
and the other output cells. Also show the Excel equa- Benefit 1 2 able Level
tion for each output cell expressed as a SUMPROD-
1 5 3 60
UCT function. Then use Solver to solve the model. 2 2 2 30
e. Summarize the model in algebraic form. 3 7 9 126
Unit cost $60 $50
3.7. Ed Butler is the production manager for the Bilco Cor-
poration, which produces three types of spare parts for auto-
mobiles. The manufacture of each part requires processing on a. Formulate a linear programming model for this
each of two machines, with the following processing times (in problem on a spreadsheet.
hours).
b. Use the spreadsheet to check the following solutions:
(x1, x2) = (7, 7), (7, 8), (8, 7), (8, 8), (8, 9), (9, 8). Which
Part of these solutions are feasible? Which of these feasible
solutions has the best value of the objective function?
Machine A B C c. Use Solver to find an optimal solution.
1 0.02 0.03 0.05 d. Express the model in algebraic form.
2 0.05 0.02 0.04 e. Use the graphical method to solve this model.

hiL18920_ch03_064-123.indd 107 12/08/17 06:54 AM

 Final PDF to printer

108  Chapter Three   Linear Programming: Formulation and Applications

3.10. Consider a cost–benefit–trade-off problem having the construction costs, Maureen needs to invest some of the compa-
following data. ny’s money now to meet these future cash flow needs. Maureen
may purchase only three kinds of financial assets, each of which
Benefit Contribution costs $1 million per unit. Fractional units may be purchased. The
per Unit of Each Activity assets produce income 5, 10, and 20 years from now, and that
Minimum income is needed to cover minimum cash flow requirements in
Acceptable those years, as shown in the following table.
Benefit 1 2 3 4 Level
P 2 −1 4 3 80 Income per Unit of Asset
Q 1 4 −1 2 60 Minimum
R 3 5 4 −1 110
Cash Flow
Unit cost $400 $600 $500 $300 Year Asset 1 Asset 2 Asset 3 Required
a. Formulate a linear programming model for this 5 $8 million $4 million $2 million $1.6 billion
problem on a spreadsheet. 10 2 million 2 million 4 million 400 million
20 0 6 million 8 million 1.2 billion
b. Make five guesses of your own choosing for the
optimal solution. Use the spreadsheet to check each
one for feasibility and, if feasible, for the value of Maureen wishes to determine the mix of investments in these
the objective function. Which feasible guess has the assets that will cover the cash flow requirements while minimiz-
best objective function value? ing the total amount invested.
c. Use Solver to find an optimal solution. a. Formulate a linear programming model for this
3.11.* Fred Jonasson manages a family-owned farm. To sup- problem on a spreadsheet.
plement several food products grown on the farm, Fred also raises b. Use the spreadsheet to check the possibility of pur-
pigs for market. He now wishes to determine the quantities of the chasing 100 units of asset 1, 100 units of asset 2, and
available types of feed (corn, tankage, and alfalfa) that should be 200 units of asset 3. How much cash flow would this
given to each pig. Since pigs will eat any mix of these feed types, mix of investments generate 5, 10, and 20 years from
the objective is to determine which mix will meet certain nutri- now? What would be the total amount invested?
tional requirements at a minimum cost. The number of units of c. Take a few minutes to use a trial-and-error approach
each type of basic nutritional ingredient contained within a kilo- with the spreadsheet to develop your best guess
gram of each feed type is given in the following table, along with for the optimal solution. What is the total amount
the daily nutritional requirements and feed costs. invested for your solution?

Minimum
Nutritional Kilogram Kilogram of Kilogram Daily
Ingredient of Corn Tankage of Alfalfa Requirement
Carbohydrates 90 20 40 200
Protein 30 80 60 180
Vitamins 10 20 60 150
Cost (¢) 84 72 60

a. Formulate a linear programming model for this

d. Use Solver to find an optimal solution.
problem on a spreadsheet.
e. Summarize the model in algebraic form.
b. Use the spreadsheet to check if (x1, x2, x3) = (1, 2, 2)
is a feasible solution and, if so, what the daily cost 3.13. Web Mercantile sells many household products
would be for this diet. How many units of each through an online catalog. The company needs substantial
nutritional ingredient would this diet provide daily? warehouse space for storing its goods. Plans now are being
made for leasing warehouse storage space over the next five
c. Take a few minutes to use a trial-and-error approach
months. Just how much space will be required in each of these
with the spreadsheet to develop your best guess for
months is known. However, since these space requirements are
the optimal solution. What is the daily cost for your
quite different, it may be most economical to lease only the
solution?
amount needed each month on a month-by-month basis. On the
d. Use Solver to find an optimal solution. other hand, the additional cost for leasing space for additional
e. Express the model in algebraic form. months is much less than for the first month, so it may be less
3.12. Maureen Laird is the chief financial officer for the Alva expensive to lease the maximum amount needed for the entire
Electric Co., a major public utility in the Midwest. The com- five months. Another option is the intermediate approach of
pany has scheduled the construction of new hydroelectric plants changing the total amount of space leased (by adding a new
5, 10, and 20 years from now to meet the needs of the growing lease and/or having an old lease expire) at least once but not
population in the region served by the company. To cover the every month.

hiL18920_ch03_064-123.indd 108 12/08/17 06:54 AM

 Final PDF to printer

Chapter 3   Problems  109

The space requirement and the leasing costs for the various 3.15. Larry Edison is the director of the Computer Center for
leasing periods are as follows. Buckly College. He now needs to schedule the staffing of the cen-
ter. It is open from 8 AM until midnight. Larry has monitored the
usage of the center at various times of the day and determined that
Required Space
the following number of computer consultants (qualified graduate
Month (Square Feet)
students) are required.
1 30,000
2 20,000 Minimum Number of
3 40,000 Consultants Required
4 10,000 Time of Day to Be on Duty
5 50,000
8 AM–noon 6
Noon–4 PM 8
4 PM–8 PM 12
Leasing Period Cost per 8 PM–midnight 6
(Months) Sq. Ft. Leased
Two types of computer consultants can be hired: full-time
1 $ 65 and part-time. The full-time consultants work for eight consecu-
2 100 tive hours in any of the following shifts: morning (8 AM–4 PM),
3 135 afternoon (noon–8 PM), and evening (4 PM–midnight). Full-time
4 160 consultants are paid $17.50 per hour.
5 190 Part-time consultants can be hired to work any of the four shifts
listed in the table. Part-time consultants are paid $15 per hour.
An additional requirement is that during every time period,
The objective is to minimize the total leasing cost for meet- there must be at least two full-time consultants on duty for every
ing the space requirements. part-time consultant on duty.
a. Indicate why this is a cost–benefit–trade-off prob- Larry would like to determine how many full-time and
lem by identifying both the activities and the ben- part-time consultants should work each shift to meet the above
efits being sought from these activities. requirements at the minimum possible cost.
b. Identify verbally the decisions to be made, the con- a. Which category of linear programming problem
straints on these decisions, and the overall measure does this problem fit? Why?
of performance for the decisions. b. Formulate and solve a linear programming model
c. Convert these verbal descriptions of the constraints for this problem on a spreadsheet.
and the measure of performance into quantitative c. Summarize the model in algebraic form.
expressions in terms of the data and decisions.
3.16.* The Medequip Company produces precision medical
d. Formulate a spreadsheet model for this problem. Iden- diagnostic equipment at two factories. Three medical centers
tify the data cells, the changing cells, the objective cell, have placed orders for this month’s production output. The fol-
and the other output cells. Also show the Excel equa- lowing table shows what the cost would be for shipping each unit
tion for each output cell expressed as a SUMPROD- from each factory to each of these customers. Also shown are
UCT function. Then use Solver to solve the model. the number of units that will be produced at each factory and the
e. Summarize the model in algebraic form. number of units ordered by each customer.
3.14. Consider the following algebraic formulation of a cost– A decision now needs to be made about the shipping plan
benefit–trade-off problem involving three benefits, where the for how many units to ship from each factory to each customer.
decisions to be made are the levels of four activities (A1, A2, A3, a. Which category of linear programming problem
and A4): does this problem fit? Why?
Minimize Cost = 2 A1 + A2 − A3 + 3 A4 b. Formulate and solve a linear programming model
for this problem on a spreadsheet.
subject to
c. Summarize this formulation in algebraic form.
Benefit 1: 3 A1 + 2 A2 − 2 A3 + 5 A4 ≥ 80 (minimum
acceptable level) Unit Shipping Cost
Benefit 2: A1 − A2 + A4 ≥ 10 (minimum
acceptable level)
To Customer Customer Customer
Benefit 3: A1 + A2 − A3 + 2 A4 ≥ 30 (minimum From 1 2 3 Output
acceptable
level) Factory 1 $600 $800 $700 400
and units
Factory 2 400 900 600 500
A1 ≥ 0   A2 ≥ 0   A3 ≥ 0   A4 ≥ 0 units
Order size 300 units 200 units 400 units
Formulate and solve the spreadsheet model for this problem.

hiL18920_ch03_064-123.indd 109 12/08/17 06:54 AM

 Final PDF to printer

110  Chapter Three   Linear Programming: Formulation and Applications

3.17. Read the referenced article that fully describes the man- invested at the beginning of year t plus Rt must
agement science study summarized in the application vignette equal the number of dollars available for investment
presented in Section 3.4. Briefly describe how linear program- at that time. Write such an equation in terms of the
ming was applied in this study. Then list the various benefits that relevant variables above for the beginning of each
resulted from this study. of the five years to obtain the five fixed-requirement
3.18. The Fagersta Steelworks currently is working two mines constraints for this problem.
to obtain its iron ore. This iron ore is shipped to either of two b. Formulate a complete linear programming model
storage facilities. When needed, it then is shipped on to the com- for this problem in algebraic form.
pany’s steel plant. The diagram below depicts this distribution c. Formulate and solve this model on a spreadsheet.
network, where M1 and M2 are the two mines, S1 and S2 are the 3.20. The Metalco Company desires to blend a new alloy of
two storage facilities, and P is the steel plant. The diagram also 40 percent tin, 35 percent zinc, and 25 percent lead from several
shows the monthly amounts produced at the mines and needed at available alloys having the following properties.
the plant, as well as the shipping cost and the maximum amount
that can be shipped per month through each shipping lane.
Alloy
40 tons $2,000/ton
produced M1 S1
30 tons max. $4 Property 1 2 3 4 5
70 00
ton /ton
$1 ons

sm 100 tons Percentage of tin 60 25 45 20 50

30

,7
00 ma

ax
t

. needed Percentage of zinc 10 15 45 50 40

/to x.
n

Percentage of lead 30 60 10 30 10
P
Cost ($/lb) 22 20 25 24 27
m n
o

ton
.
to 0/t
ax

/ x.
00
50 1,60

$8 s ma
ns
$

to n
70 The objective is to determine the proportions of these alloys
60 tons $1,100/ton that should be blended to produce the new alloy at a minimum cost.
M2 S2
produced 50 tons max.
a. Identify all the requirements that will need to be
expressed in fixed-requirement constraints.
Management now wants to determine the most economical b. Formulate and solve a linear programming model
plan for shipping the iron ore from the mines through the distri- for this problem on a spreadsheet.
bution network to the steel plant.
c. Express this model in algebraic form.
a. Identify all the requirements that will need to be
3.21. The Weigelt Corporation has three branch plants with
expressed in fixed-requirement constraints.
excess production capacity. Fortunately, the corporation has
b. Formulate and solve a linear programming model a new product ready to begin production, and all three plants
for this problem on a spreadsheet. have this capability, so some of the excess capacity can be
c. Express this model in algebraic form. used in this way. This product can be made in three sizes—
3.19.* Al Ferris has $60,000 that he wishes to invest now in large, medium, and small—that yield a net unit profit of $420,
order to use the accumulation for purchasing a retirement annu- $360, and $300, respectively. Plants 1, 2, and 3 have the excess
ity in five years. After consulting with his financial advisor, he capacity to produce 750, 900, and 450 units per day of this
has been offered four types of fixed-income investments, which product, respectively, regardless of the size or combination of
we will label as investments A, B, C, and D. sizes involved.
Investments A and B are available at the beginning of each of the The amount of available in-process storage space also
next five years (call them years 1 to 5). Each dollar invested in A at imposes a limitation on the production rates of the new product.
the beginning of a year returns $1.40 (a profit of $0.40) two years Plants 1, 2, and 3 have 13,000, 12,000, and 5,000 square feet,
later (in time for immediate reinvestment). Each dollar invested in B respectively, of in-process storage space available for a day’s
at the beginning of a year returns $1.70 three years later. production of this product. Each unit of the large, medium, and
Investments C and D will each be available at one time in small sizes produced per day requires 20, 15, and 12 square feet,
the future. Each dollar invested in C at the beginning of year 2 respectively.
returns $1.90 at the end of year 5. Each dollar invested in D at Sales forecasts indicate that if available, 900, 1,200, and 750
the beginning of year 5 returns $1.30 at the end of year 5. units of the large, medium, and small sizes, respectively, would
Al wishes to know which investment plan maximizes the amount be sold per day.
of money that can be accumulated by the beginning of year 6. At each plant, some employees will need to be laid off unless
a. For this problem, all its functional constraints can most of the plant’s excess production capacity can be used to
be expressed as fixed-requirement constraints. To produce the new product. To avoid layoffs if possible, manage-
do this, let At, Bt, Ct, and Dt be the amounts invested ment has decided that the plants should use the same percentage
in investments A, B, C, and D, respectively, at the of their excess capacity to produce the new product.
beginning of year t for each t where the investment Management wishes to know how much of each of the sizes
is available and will mature by the end of year 5. should be produced by each of the plants to maximize profit.
Also let Rt be the number of available dollars not a. Formulate and solve a linear programming model
invested at the beginning of year t (and so available for this mixed problem on a spreadsheet.
for investment in a later year). Thus, the amount b. Express the model in algebraic form.

hiL18920_ch03_064-123.indd 110 12/08/17 06:54 AM

 Final PDF to printer

Chapter 3   Problems  111

3.22.* A cargo plane has three compartments for storing cargo:

front, center, and back. These compartments have capacity limits Material Labor Gross
on both weight and space, as summarized below. Required Required Profit
Glove (Square Feet) (Minutes) (per Pair)

Men’s 2 30 $8
Weight Space Women’s 1.5 45 10
Capacity Capacity Children’s 1 40 6
Compartment (Tons) (Cubic Feet)

Front 12 7,000 Each full-time employee earns $13 per hour, while each
Center 18 9,000 part-time employee earns $10 per hour. Management wishes to
Back 10 5,000 know what mix of each of the three types of gloves to produce
per week, as well as how many full-time and part-time workers
to employ. They would like to maximize their net profit—their
gross profit from sales minus their labor costs.
Furthermore, the weight of the cargo in the respective com-
partments must be the same proportion of that compartment’s a. Formulate and solve a linear programming model
weight capacity to maintain the balance of the airplane. for this problem on a spreadsheet.
The following four cargoes have been offered for shipment b. Summarize this formulation in algebraic form.
on an upcoming flight as space is available. 3.24. Oxbridge University maintains a powerful mainframe
computer for research use by its faculty, Ph.D. students, and
research associates. During all working hours, an operator must
Weight Volume Profit be available to operate and maintain the computer, as well as to
Cargo (Tons) (Cubic Feet/Ton) ($/Ton) perform some programming services. Beryl Ingram, the director
of the computer facility, oversees the operation.
1 20 500 320 It is now the beginning of the fall semester and Beryl is con-
2 16 700 400 fronted with the problem of assigning different working hours to
3 25 600 360 her operators. Because all the operators are currently enrolled in
4 13 400 290 the university, they are available to work only a limited number
of hours each day.
There are six operators (four undergraduate students and two
graduate students). They all have different wage rates because of
Any portion of these cargoes can be accepted. The objec- differences in their experience with computers and in their pro-
tive is to determine how much (if any) of each cargo should be gramming ability. The following table shows their wage rates, along
accepted and how to distribute each among the compartments to with the maximum number of hours that each can work each day.
maximize the total profit for the flight.
a. Formulate and solve a linear programming model
Maximum Hours of Availability
for this mixed problem on a spreadsheet.
b. Express the model in algebraic form. Operators Wage Rate Mon. Tue. Wed. Thurs. Fri.
3.23. Comfortable Hands is a company that features a prod-
uct line of winter gloves for the entire family—men, women, and K. C. $20.00/hour 6 0 6 0 6
children. They are trying to decide what mix of these three types D. H. $20.20/hour 0 6 0 6 0
of gloves to produce.
H. B. $19.80/hour 4 8 4 0 4
Comfortable Hands’s manufacturing labor force is union-
ized. Each full-time employee works a 40-hour week. In addi- S. C. $19.60/hour 5 5 5 0 5
tion, by union contract, the number of full-time employees K. S. $21.60/hour 3 0 3 8 0
can never drop below 20. Nonunion, part-time workers also N. K. $22.60/hour 0 0 0 6 2
can be hired with the following union-imposed restrictions:
(1) each part-time worker works 20 hours per week and
(2) there must be at least two full-time employees for each part- Each operator is guaranteed a certain minimum number of
time employee. hours per week that will maintain an adequate knowledge of the
All three types of gloves are made out of the same 100 operation. This level is set arbitrarily at 8 hours per week for
percent genuine cowhide leather. Comfortable Hands has a the undergraduate students (K. C., D. H., H. B., and S. C.) and
long-term contract with a supplier of the leather and receives a 7 hours per week for the graduate students (K. S. and N. K.).
5,000-square-foot shipment of the material each week. The mate- The computer facility is to be open for operation from 8 AM
rial requirements and labor requirements, along with the gross to 10 PM Monday through Friday with exactly one operator on
profit per glove sold (not considering labor costs), are given in duty during these hours. On Saturdays and Sundays, the com-
the following table. puter is to be operated by other staff.

hiL18920_ch03_064-123.indd 111 12/08/17 06:54 AM

 Final PDF to printer

112  Chapter Three   Linear Programming: Formulation and Applications

Because of a tight budget, Beryl has to minimize cost. She 12 grams (g) of protein. Furthermore, for practical reasons, each
wishes to determine the number of hours she should assign to child needs exactly 2 slices of bread (to make the sandwich), at
each operator on each day. Formulate and solve a spreadsheet least twice as much peanut butter as jelly, and at least 1 cup of
model for this problem. liquid (milk and/or juice).
3.25. Slim-Down Manufacturing makes a line of nutritionally Joyce and Marvin would like to select the food choices
complete, weight-reduction beverages. One of its products is a for each child that minimize cost while meeting the above
strawberry shake that is designed to be a complete meal. The requirements.
strawberry shake consists of several ingredients. Some informa-
tion about each of these ingredients is given next.

Calories Total Calo- Vitamin

from Fat ries (per Content Thickeners Cost
Ingredient (per tbsp.) tbsp.) (mg/tbsp.) (mg/tbsp.) (¢/tbsp.)

Strawberry flavoring 1 50 20 3 10
Cream 75 100 0 8 8
Vitamin supplement 0   0 50 1 25
Artificial sweetener 0 120 0 2 15
Thickening agent 30 80 2 25 6

The nutritional requirements are as follows. The beverage a. Identify the requirements that lead to resource con-
must total between 380 and 420 calories (inclusive). No more straints, to benefit constraints, and to fixed-require-
than 20 percent of the total calories should come from fat. There ment constraints.
must be at least 50 milligrams (mg) of vitamin content. For taste b. Formulate and solve a linear programming model
reasons, there must be at least two tablespoons (tbsp.) of straw- for this problem on a spreadsheet.
berry flavoring for each tbsp. of artificial sweetener. Finally, to c. Express the model in algebraic form.
maintain proper thickness, there must be exactly 15 mg of thick-
3.27. The Cost-Less Corp. supplies its four retail outlets from
eners in the beverage.
its four plants. The shipping cost per shipment from each plant to
Management would like to select the quantity of each ingre-
each retail outlet is given below.
dient for the beverage that would minimize cost while meeting
the above requirements.
a. Identify the requirements that lead to resource con- Unit Shipping Cost
straints, to benefit constraints, and to fixed-require-
ment constraints. Retail Outlet: 1 2 3 4
b. Formulate and solve a linear programming model
for this problem on a spreadsheet. Plant
1 $500 $600 $400 $200
c. Summarize this formulation in algebraic form. 2 200 900 100 300
3.26. Joyce and Marvin run a day care for preschoolers. They 3 300 400 200 100
are trying to decide what to feed the children for lunches. They 4 200 100 300 200
would like to keep their costs down, but they also need to meet the
nutritional requirements of the children. They have already decided
to go with peanut butter and jelly sandwiches, and some combi- Plants 1, 2, 3, and 4 make 10, 20, 20, and 10 shipments per
nation of graham crackers, milk, and orange juice. The nutritional month, respectively. Retail outlets 1, 2, 3, and 4 need to receive
content of each food choice and its cost are given in the table below. 20, 10, 10, and 20 shipments per month, respectively.

Calories Total Vitamin C Protein Cost

Food Item from Fat Calories (mg) (g) (¢)

Bread (1 slice) 10 70 0 3 5
Peanut butter (1 tbsp.) 75 100 0 4 4
Strawberry jelly (1 tbsp.) 0 50 3 0 7
Graham cracker (1 cracker) 20 60 0 1 8
Milk (1 cup) 70 150 2 8 15
Juice (1 cup) 0 100 120 1 35

The nutritional requirements are as follows. Each child The distribution manager, Randy Smith, now wants to deter-
should receive between 400 and 600 calories. No more than 30 mine the best plan for how many shipments to send from each
percent of the total calories should come from fat. Each child plant to the respective retail outlets each month. Randy’s objec-
should consume at least 60 milligrams (mg) of vitamin C and tive is to minimize the total shipping cost.

hiL18920_ch03_064-123.indd 112 12/08/17 06:54 AM

 Final PDF to printer

Chapter 3   Problems  113

Formulate this problem as a transportation problem on a A total of 60 forklift trucks are produced and shipped per
spreadsheet and then use Solver to obtain an optimal solution. week. Each plant can produce and ship any amount up to a maxi-
3.28. The Childfair Company has three plants producing mum of 50 trucks per week, so there is considerable flexibility
child push chairs that are to be shipped to four distribution cen- on how to divide the total production between the two plants so
ters. Plants 1, 2, and 3 produce 12, 17, and 11 shipments per as to reduce shipping costs. However, each distribution center
month, respectively. Each distribution center needs to receive must receive exactly 20 trucks per week.
10 shipments per month. The distance from each plant to the Management’s objective is to determine how many forklift
respective distribution centers is given below. trucks should be produced at each plant, and then what the over-
all shipping pattern should be to minimize total shipping cost.
Formulate and solve a spreadsheet model for this problem.
Distance to Distribution Center (Miles)
3.31. Redo Problem 3.30 when any distribution center may
1 2 3 4 receive any quantity between 10 and 30 forklift trucks per week
in order to further reduce total shipping cost, provided only that
Plant the total shipped to all three distribution centers must still equal
1 800 1,300 400 700 60 trucks per week.
2 1,100 1,400 600 1,000
3.32. Read the referenced article that fully describes the man-
3 600 1,200 800 900
agement science study summarized in the application vignette
presented in Section 3.6. Briefly describe how the model for the
The freight cost for each shipment is $100 plus 50 cents/mile. assignment problem was applied in this study. Then list the various
How much should be shipped from each plant to each of the financial and nonfinancial benefits that resulted from this study.
distribution centers to minimize the total shipping cost? 3.33. Consider the assignment problem having the following
Formulate this problem as a transportation problem on a cost table.
spreadsheet and then use Solver to obtain an optimal solution.
3.29. The Onenote Co. produces a single product at three Job
plants for four customers. The three plants will produce 60, 80, 1 2 3
and 40 units, respectively, during the next week. The firm has
made a commitment to sell 40 units to customer 1, 60 units to Person
customer 2, and at least 20 units to customer 3. Both customers A $5 $7 $4
3 and 4 also want to buy as many of the remaining units as pos- B 3 6 5
sible. The net profit associated with shipping a unit from plant i C 2 3 4
for sale to customer j is given by the following table.
The optimal solution is A-3, B-1, C-2, with a total cost of $10.
Formulate this problem on a spreadsheet and then use Solver
Customer to obtain the optimal solution identified above.
1 2 3 4 3.34. Four cargo ships will be used for shipping goods from
one port to four other ports (labeled 1, 2, 3, 4). Any ship can be
Plant used for making any one of these four trips. However, because
1 $800 $700 $500 $200
of differences in the ships and cargoes, the total cost of loading,
2 500 200 100 300
transporting, and unloading the goods for the different ship–port
3 600 400 300 500
combinations varies considerably, as shown in the following table.

Management wishes to know how many units to sell to cus- Port

tomers 3 and 4 and how many units to ship from each of the
plants to each of the customers to maximize profit. Formulate 1 2 3 4
and solve a spreadsheet model for this problem.
Ship
3.30. The Move-It Company has two plants building forklift 1 $500 $400 $600 $700
trucks that then are shipped to three distribution centers. The 2 600 600 700 500
production costs are the same at the two plants, and the cost of 3 700 500 700 600
shipping each truck is shown below for each combination of 4 500 400 600 600
plant and distribution center.

The objective is to assign the four ships to four different ports

Distribution Center in such a way as to minimize the total cost for all four shipments.
1 2 3 a. Describe how this problem fits into the format for
an assignment problem.
Plant b. Formulate and solve this problem on a spreadsheet.
A $800 $700 $400
3.35. Reconsider Problem 3.30. Now distribution centers
B 600 800 500
1, 2, and 3 must receive exactly 10, 20, and 30 units per week,

hiL18920_ch03_064-123.indd 113 12/08/17 06:54 AM

 Final PDF to printer

114  Chapter Three   Linear Programming: Formulation and Applications

respectively. For administrative convenience, management has high-speed copier. Finally, the copiers need to be able to handle
decided that each distribution center will be supplied totally by a a capacity of at least 75,000 copies per day. The objective is to
single plant, so that one plant will supply one distribution center determine the mix of these two copiers that will handle all these
and the other plant will supply the other two distribution centers. requirements at minimum cost.
The choice of these assignments of plants to distribution centers is a. Formulate and solve a spreadsheet model for this
to be made solely on the basis of minimizing total shipping cost. problem.
Formulate and solve a spreadsheet model for this problem.
b. Formulate this same model in algebraic form.
3.36. Vincent Cardoza is the owner and manager of a machine
shop that does custom order work. This Wednesday afternoon, 3.38. Northeastern Airlines is considering the purchase of
he has received calls from two customers who would like to new long-, medium-, and short-range jet passenger airplanes.
place rush orders. One is a trailer hitch company that would like The purchase price would be $201 million for each long-
some custom-made heavy-duty tow bars. The other is a mini- range plane, $150 million for each medium-range plane, and
car-carrier company that needs some customized stabilizer bars. $105 million for each short-range plane. The board of direc-
Both customers would like as many as possible by the end of tors has authorized a maximum commitment of $4.5 billion for
the week (two working days). Since both products would require these purchases. Regardless of which airplanes are purchased,
the use of the same two machines, Vincent needs to decide and air travel of all distances is expected to be sufficiently large that
inform the customers this afternoon about how many of each these planes would be utilized at essentially maximum capac-
product he will agree to make over the next two days. ity. It is estimated that the net annual profit (after capital recov-
Each tow bar requires 3.2 hours on machine 1 and 2 hours ery costs are subtracted) would be $12.6 million per long-range
on machine 2. Each stabilizer bar requires 2.4 hours on machine plane, $9 million per medium-range plane, and $6.9 million per
1 and 3 hours on machine 2. Machine 1 will be available for short-range plane.
16 hours over the next two days and machine 2 will be available It is predicted that enough trained pilots will be available
for 15 hours. The profit for each tow bar produced would be $130 to the company to crew 30 new airplanes. If only short-range
and the profit for each stabilizer bar produced would be $150. planes were purchased, the maintenance facilities would be able
Vincent now wants to determine the mix of these production to handle 40 new planes. However, each medium-range plane is
quantities that will maximize the total profit. 11
equivalent to ___
​​   ​​short-range planes, and each long-range plane
a. Formulate an integer programming model in alge- 3
braic form for this problem. 12
is equivalent to ___ ​​   ​​short-range planes in terms of their use of the
3
b. Formulate and solve the model on a spreadsheet.
maintenance facilities.
3.37. Pawtucket University is planning to buy new copier The information given here was obtained by a preliminary
machines for its library. Three members of its Management analysis of the problem. A more detailed analysis will be con-
Science Department are analyzing what to buy. They are con- ducted subsequently. However, using the preceding data as a first
sidering two different models: Model A, a high-speed copier, approximation, management wishes to know how many planes
and Model B, a lower speed but less expensive copier. Model A of each type should be purchased to maximize profit.
can handle 20,000 copies a day and costs $6,000. Model B can
handle 10,000 copies a day but only costs $4,000. They would a. Formulate and solve a spreadsheet model for this
like to have at least six copiers so that they can spread them problem.
throughout the library. They also would like to have at least one b. Formulate this model in algebraic form.

Case 3-1
Shipping Wood to Market
Alabama Atlantic is a lumber company that has three sources of In the past, the company has shipped the wood by train.
wood and five markets to be supplied. The annual availability of However, because shipping costs have been increasing, the
wood at sources 1, 2, and 3 is 15, 20, and 15 million board feet, alternative of using ships to make some of the deliveries is
respectively. The amount that can be sold annually at markets 1, 2, being investigated. This alternative would require the company
3, 4, and 5 is 11, 12, 9, 10, and 8 million board feet, respectively. to invest in some ships. Except for these investment costs, the

Unit Cost by Rail ($1,000s) to Market Unit Cost by Ship ($1,000s) to Market

Source 1 2 3 4 5 1 2 3 4 5

1 61 72 45 55 66 31 38 24 — 35
2 69 78 60 49 56 36 43 28 24 31
3 59 66 63 61 47 — 33 36 32 26

hiL18920_ch03_064-123.indd 114 12/08/17 06:54 AM

 Final PDF to printer

Case 3-2   Capacity Concerns   115

shipping costs in thousands of dollars per million board feet by is to determine the overall shipping plan that minimizes the total
rail and by water (when feasible) is given by the above table for equivalent uniform annual cost (including shipping costs).
each route. You are the head of the management science team that has
The capital investment (in thousands of dollars) in ships been assigned the task of determining this shipping plan for each
required for each million board feet to be transported annually by of the three options listed next.
ship along each route is given next.
Option 1: Continue shipping exclusively by rail.

Unit Investment for Ships ($1,000s) to Option 2: Switch to shipping exclusively by water (except
Market where only rail is feasible).

Source 1 2 3 4 5 Option 3: Ship by either rail or water, depending on which is

less expensive for the particular route.
1 275 303 238 — 285
Present your results for each option. Compare.
2 293 318 270 250 265
3 — 283 275 268 240
Finally, consider the fact that these results are based on cur-
rent shipping and investment costs, so that the decision on the
option to adopt now should take into account management’s
Considering the expected useful life of the ships and the time projection of how these costs are likely to change in the future.
value of money, the equivalent uniform annual cost of these invest- For each option, describe a scenario of future cost changes that
ments is one-tenth the amount given in the table. The objective would justify adopting that option now.

Case 3-2
Capacity Concerns
Bentley Hamilton throws the business section of The New York missed or late orders have increased steadily and significantly
Times onto the conference room table and watches as his associ- over the past 12 months. I think this trend explains why we have
ates jolt upright in their overstuffed chairs. been losing market share, causing our stock to drop to its low-
Mr. Hamilton wants to make a point. est level in 52 weeks. We have angered and lost the business of
He throws the front page of the The Wall Street Journal on retailers, our customers who depend upon on-time deliveries to
top of The New York Times and watches as his associates widen meet the demand of consumers.”
their eyes once heavy with boredom. “Why have we missed our delivery dates when our productiv-
Mr. Hamilton wants to make a big point. ity level should have allowed us to fill all orders?” Mr. ­Hamilton
He then throws the front page of the Financial Times on top asks. “I called several departments to ask this question.”
of the newspaper pile and watches as his associates dab the fine “It turns out that we have been producing routers for the
beads of sweat off their brows. hell of it!” Mr. Hamilton says in disbelief. “The marketing and
Mr. Hamilton wants his point indelibly etched into his associ- sales departments do not communicate with the manufactur-
ates’ minds. ing department, so manufacturing executives do not know what
“I have just presented you with three leading financial routers to produce to fill orders. The manufacturing executives
newspapers carrying today’s top business story,” Mr. H ­ amilton want to keep the plant running, so they produce routers regard-
declares in a tight, angry voice. “My dear associates, our com- less of whether the routers have been ordered. Finished routers
pany is going to hell in a hand basket! Shall I read you the head- are sent to the warehouse, but marketing and sales executives
lines? From The New York Times, ‘CommuniCorp stock drops do not know the number and styles of routers in the warehouse.
to lowest in 52 weeks.’ From The Wall Street Journal, ‘Com- They try to communicate with warehouse executives to deter-
muniCorp loses 25 percent of the wireless router market in only mine if the routers in inventory can fill the orders, but they rarely
one year.’ Oh, and my favorite, from the Financial Times, ‘Com- receive answers to their questions.”
muniCorp cannot CommuniCate: CommuniCorp stock drops Mr. Hamilton pauses and looks directly at his associates.
because of internal communications disarray.’ How did our “Ladies and gentlemen, it seems to me that we have a serious
company fall into such dire straits?” internal communications problem. I intend to correct this prob-
Mr. Hamilton next points at a line sloping slightly upward on lem immediately. I want to begin by installing a companywide
the conference room display. “This is a graph of our productivity computer network to ensure that all departments have access to
over the last 12 months. As you can see from the graph, productiv- critical documents and are able to communicate with each other
ity in our router production facility has increased steadily over the more easily. Because this intranet will represent a large change
last year. Clearly, productivity is not the cause of our problem.” from the current communications infrastructure, I expect some
Mr. Hamilton next displays a second graph showing a line bugs in the system and some resistance from employees. I there-
sloping steeply upward. “This is a graph of our missed or late fore want to phase in the installation of the intranet.”
orders over the last 12 months.” Mr. Hamilton hears an audible Mr. Hamilton passes the following time line and require-
gasp from his associates. “As you can see from the graph, our ments chart to his associates (IN = intranet).

hiL18920_ch03_064-123.indd 115 12/08/17 06:54 AM

 Final PDF to printer

116  Chapter Three   Linear Programming: Formulation and Applications

Month 1 Month 2 Month 3 Month 4 Month 5

IN
education
Install IN in
sales
Install IN in
manufacturing
Install IN in
warehouse
Install IN in
marketing

month to support each new group of employees gaining access

Department Number of Employees to the intranet.”
“There are several factors that complicate your decision,”
Sales 60
Manufacturing 200
Mr. Hamilton continues. We can get a discount of 10 percent off
Warehouse 30 each workstation server purchased, but only if we purchase serv-
Marketing 75 ers in the first or second month. We can get a 25 percent discount
off all full rack servers purchased in the first two months. You are
also limited in the amount of money you can spend during the first
Mr. Hamilton proceeds to explain the time line and require- month. CommuniCorp has already allocated much of the budget
ments chart. “In the first month, I do not want to bring any for the next two months, so you only have a total of $9,500 avail-
department onto the intranet; I simply want to disseminate infor- able to purchase servers in months 1 and 2. Finally, the manufac-
mation about it and get buy-in from employees. In the second turing department requires at least one of the three more powerful
month, I want to bring the sales department onto the intranet servers. Have your decision on my desk at the end of the week.”
since the sales department receives all critical information from a. Emily first decides to evaluate the number and type of servers
customers. In the third month, I want to bring the manufactur- to purchase on a month-to-month basis. For each month, for-
ing department onto the intranet. In the fourth month, I want mulate a spreadsheet model to determine which servers Emily
to install the intranet at the warehouse, and in the fifth and should purchase in that month to minimize costs in that month
final month, I want to bring the marketing department onto and support the new users given your results for the preceding
the intranet. The requirements chart above lists the number of months. How many and which types of servers should she pur-
employees requiring access to the intranet in each department.” chase in each month? How much is the total cost of the plan?
Mr. Hamilton turns to Emily Jones, the head of Corporate
b. Emily realizes that she could perhaps achieve savings if she bought
Information Management. “I need your help in planning for the
a larger server in the initial months to support users in the final
installation of the intranet. Specifically, the company needs to
months. She therefore decides to evaluate the number and type of
purchase servers for the internal network. Employees will con-
servers to purchase over the entire planning period. Formulate a
nect to company servers and download information to their own
spreadsheet model to determine which servers Emily should pur-
desktop computers.”
chase in which months to minimize total cost and support all new
Mr. Hamilton passes Emily the following chart detailing the
users. How many and which types of servers should she purchase
types of servers available, the number of employees each server
in each month? How much is the total cost of the plan?
supports, and the cost of each server.

Number of Employees
Type of Server Server Supports Cost of Server

Mini Desktop Server Up to 30 employees $ 2,500

Desktop Server Up to 80 employees 5,000
Workstation Server Up to 200 employees 10,000
Full Rack Server Up to 2,000 employees 25,000

“Emily, I need you to decide what servers to purchase and c. Why is the answer using the first method different from that
when to purchase them to minimize cost and to ensure that the using the second method?
company possesses enough server capacity to follow the intranet d. Are there other costs for which Emily is not accounting in her
implementation timeline,” Mr. Hamilton says. “For example, you problem formulation? If so, what are they?
may decide to buy one large server during the first month to sup- e. What further concerns might the various departments of
port all employees, or buy several small servers during the first CommuniCorp have regarding the intranet?
month to support all employees, or buy one small server each

hiL18920_ch03_064-123.indd 116 12/08/17 06:54 AM

 Final PDF to printer

Case 3-3   Fabrics and Fall Fashions   117

Case 3-3
Fabrics and Fall Fashions
From the 10th floor of her office building, Katherine Rally The fall casual fashions include.
watches the swarms of New Yorkers fight their way through the
streets infested with yellow cabs and the sidewalks littered with hot Labor and
dog stands. On this sweltering July day, she pays particular atten- Clothing Material Machine
tion to the fashions worn by the various women and wonders what Item Requirements Price Cost
they will choose to wear in the fall. Her thoughts are not simply
Velvet pants 3 yards of velvet $350 $175
random musings; they are critical to her work since she owns and
2 yards of acetate
manages TrendLines, an elite women’s clothing company.
for lining
Today is an especially important day because she must meet
with Ted Lawson, the production manager, to decide upon next Cotton 1.5 yards of cotton 130 60
sweater
month’s production plan for the fall line. Specifically, she must
determine the quantity of each clothing item she should produce Cotton 0.5 yard of cotton 75 40
miniskirt
given the plant’s production capacity, limited resources, and
demand forecasts. Accurate planning for next month’s production Velvet shirt 1.5 yards of velvet 200 160
is critical to fall sales since the items produced next month will Button-down 1.5 yards of rayon 120 90
appear in stores during September and women generally buy the blouse
majority of the fall fashions when they first appear in September.
She turns back to her sprawling glass desk and looks at the
She knows that for the next month, she has ordered 45,000
numerous papers covering it. Her eyes roam across the clothing
yards of wool, 28,000 yards of acetate, 9,000 yards of cashmere,
patterns designed almost six months ago, the lists of material
18,000 yards of silk, 30,000 yards of rayon, 20,000 yards of vel-
requirements for each pattern, and the lists of demand forecasts
vet, and 30,000 yards of cotton for production. The prices of the
for each pattern determined by customer surveys at fashion
materials are listed below.
shows. She remembers the hectic and sometimes nightmarish
days of designing the fall line and presenting it at fashion shows
in New York, Milan, and Paris. Ultimately, she paid her team of Material Price per Yard
six designers a total of $860,000 for their work on her fall line.
With the cost of hiring runway models, hair stylists, and make-up Wool $ 9.00
Acetate 1.50
artists; sewing and fitting clothes; building the set; choreograph-
Cashmere 60.00
ing and rehearsing the show; and renting the conference hall, each
Silk 13.00
of the three fashion shows cost her an additional $2,700,000.
Rayon 2.25
She studies the clothing patterns and material requirements. Velvet 12.00
Her fall line consists of both professional and casual fashions. She Cotton 2.50
determined the price for each clothing item by taking into account
the quality and cost of material, the cost of labor and machining, the
demand for the item, and the prestige of the TrendLines brand name.
Any material that is not used in production can be sent back
The fall professional fashions include.
to the textile wholesaler for a full refund, although scrap mate-
rial cannot be sent back to the wholesaler.
She knows that the production of both the silk blouse and cot-
ton sweater leaves leftover scraps of material. Specifically, for
Clothing Material Labor and the production of one silk blouse or one cotton sweater, 2 yards
Item Requirements Price Machine Cost of silk and cotton, respectively, are needed. From these 2 yards,
1.5 yards are used for the silk blouse or the cotton sweater and
Tailored 3 yards of wool $300 $160 0.5 yard is left as scrap material. She does not want to waste the
wool slacks 2 yards of acetate material, so she plans to use the rectangular scrap of silk or cot-
for lining ton to produce a silk camisole or cotton miniskirt, respectively.
Cashmere 1.5 yards of 450 150 Therefore, whenever a silk blouse is produced, a silk camisole is
sweater cashmere also produced. Likewise, whenever a cotton sweater is produced,
Silk blouse 1.5 yards of silk 180 100 a cotton miniskirt is also produced. Note that it is possible to
Silk camisole 0.5 yard of silk 120 60 produce a silk camisole without producing a silk blouse and a
Tailored skirt 2 yards of rayon 270 120 cotton miniskirt without producing a cotton sweater.
1.5 yards of ace- The demand forecasts indicate that some items have lim-
tate for lining ited demand. Specifically, because the velvet pants and velvet
Wool blazer 2.5 yards of wool 320 140 shirts are fashion fads, TrendLines has forecasted that it can
1.5 yards of ace- sell only 5,500 pairs of velvet pants and 6,000 velvet shirts.
tate for lining TrendLines does not want to produce more than the forecasted
demand because once the pants and shirts go out of style, the

hiL18920_ch03_064-123.indd 117 12/08/17 06:54 AM

 Final PDF to printer

118  Chapter Three   Linear Programming: Formulation and Applications

company cannot sell them. TrendLines can produce less than the Before she makes her final decision, Katherine plans to
forecasted demand, however, since the company is not required explore the following questions independently, except where
to meet the demand. The cashmere sweater also has limited otherwise indicated.
demand because it is quite expensive, and TrendLines knows it
can sell at most 4,000 cashmere sweaters. The silk blouses and c. The textile wholesaler informs Katherine that the velvet can-
camisoles have limited demand because many women think silk not be sent back because the demand forecasts show that the
is too hard to care for, and TrendLines projects that it can sell at demand for velvet will decrease in the future. Katherine can
most 12,000 silk blouses and 15,000 silk camisoles. therefore get no refund for the velvet. How does this fact
The demand forecasts also indicate that the wool slacks, change the production plan?
tailored skirts, and wool blazers have a great demand because d. What is an intuitive economic explanation for the difference
they are basic items needed in every professional wardrobe. between the solutions found in parts b and c?
Specifically, the demand is 7,000 pairs of wool slacks and 5,000 e. The sewing staff encounters difficulties sewing the arms and
wool blazers. Katherine wants to meet at least 60 percent of the lining into the wool blazer since the blazer pattern has an
demand for these two items to maintain her loyal customer base awkward shape and the heavy wool material is difficult to
and not lose business in the future. Although the demand for cut and sew. The increased labor time to sew a wool blazer
tailored skirts could not be estimated, Katherine feels she should increases the labor and machine cost for each blazer by $80.
make at least 2,800 of them. Given this new cost, how many of each clothing item should
TrendLines produce to maximize profit?
a. Ted is trying to convince Katherine not to produce any velvet f. The textile wholesaler informs Katherine that since another
shirts since the demand for this fashion fad is quite low. He textile customer canceled his order, she can obtain an extra
argues that this fashion fad alone accounts for $500,000 of 10,000 yards of acetate. How many of each clothing item
the fixed design and other costs. The net contribution (price should TrendLines now produce to maximize profit?
of clothing item – materials cost – labor cost) from selling g. TrendLines assumes that it can sell every item that was not
the fashion fad should cover these fixed costs. Each velvet sold during September and October in a big sale in November
shirt generates a net contribution of $22. He argues that given at 60 percent of the original price. Therefore, it can sell all
the net contribution, even satisfying the maximum demand items in unlimited quantity during the November sale. (The
will not yield a profit. What do you think of Ted’s argument? previously mentioned upper limits on demand only concern
b. Formulate and solve a linear programming problem to maximize the sales during September and October.) What should the
profit given the production, resource, and demand constraints. new production plan be to maximize profit?

Case 3-4
New Frontiers
Rob Richman, president of AmeriBank, takes off his glasses, nonbank and foreign bank competitors do not readily provide:
rubs his eyes in exhaustion, and squints at the clock in his study. transaction services. He has decided that a more convenient trans-
It reads 3 AM. For the last several hours, Rob has been poring over action method must logically succeed the automatic teller machine,
­AmeriBank’s financial statements from the last three quarters and he believes that electronic banking over the Internet allows
of operation. AmeriBank, a medium-sized bank with branches this convenient transaction method. Over the Internet, customers
throughout the United States, is headed for dire economic straits. are able to perform transactions on their desktop computers either
The bank, which provides transaction, savings, investment, and at home or work. The explosion of the Internet means that most
loan services, has been experiencing a steady decline in its net potential customers understand and use it. He therefore feels that if
income over the past year, and trends show that the decline will AmeriBank offers Web banking (as the practice of Internet bank-
continue. The bank is simply losing customers to nonbank and ing is commonly called), the bank will attract many new customers.
foreign bank competitors. Before Rob undertakes the project to make Web banking
AmeriBank is not alone in its struggle to stay out of the possible, however, he needs to understand the market for Web
red. From his daily industry readings, Rob knows that many banking and the services AmeriBank should provide over the
­American banks have been suffering significant losses because Internet. For example, should the bank only allow customers to
of increasing competition from nonbank and foreign bank com- access account balances and historical transaction information
petitors offering services typically in the domain of American over the Internet, or should the bank develop a strategy to allow
banks. Because the nonbank and foreign bank competitors spe- customers to make deposits and withdrawals over the Internet?
cialize in particular services, they are able to better capture the Should the bank try to recapture a portion of the investment mar-
market for those services by offering less expensive, more effi- ket by continuously running stock prices and allowing customers
cient, more convenient services. For example, large corporations to make stock transactions over the Internet for a minimal fee?
now turn to foreign banks and commercial paper offerings for Therefore, Rob has concluded that a major survey project should
loans, and affluent Americans now turn to money-market funds be undertaken to learn what customers want.
for investment. Banks face the daunting challenge of distinguish- Because AmeriBank is not in the business of performing
ing themselves from nonbank and foreign bank competitors. surveys, Rob has decided to outsource the survey project to a
Rob has concluded that one strategy for distinguishing professional survey company. He has opened the project up for
AmeriBank from its competitors is to improve services that
­ bidding by several survey companies and will award the project

hiL18920_ch03_064-123.indd 118 12/08/17 06:54 AM

 Final PDF to printer

Case 3-5   Assigning Students to Schools   119

to the company that is willing to perform the survey for the least a. Formulate a linear programming model to minimize costs
cost. Rob provided each survey company with a list of survey while meeting all survey constraints imposed by AmeriBank.
requirements to ensure that AmeriBank receives the needed b. If the profit margin for Sophisticated Surveys is 15 percent of
information for planning the Web banking project. cost, what bid will it submit?
Because different age groups require different services, Amer- c. After submitting its bid, Sophisticated Surveys is informed
iBank is interested in surveying four different age groups. The first that it has the lowest cost but that AmeriBank does not like
group encompasses customers who are 18 to 25 years old. The the solution. Specifically, Rob feels that the selected survey
bank assumes that this age group has limited yearly income and population is not representative enough of the banking cus-
performs minimal transactions. The second group encompasses tomer population. Rob wants at least 50 people of each age
customers who are 26 to 40 years old. This age group has sig- group surveyed in each region. What is the new bid made by
nificant sources of income, performs many transactions, requires Sophisticated Surveys?
numerous loans for new houses and cars, and invests in various
d. Rob feels that Sophisticated Surveys oversampled the 18-to-
securities. The third group encompasses customers who are 41
25-year-old population and the Silicon Valley population. He
to 50 years old. These customers typically have the same level of
imposes a new constraint that no more than 600 individuals
income and perform the same number of transactions as the s­ econd
can be surveyed from the 18-to-25-year-old population and
age group, but the bank assumes that these customers are less
no more than 650 individuals can be surveyed from the Sili-
likely to use Web banking since they have not become as comfort-
con Valley population. What is the new bid?
able with the explosion of computers or the Internet. Finally, the
fourth group encompasses customers who are 51 years of age and e. When Sophisticated Surveys calculated the cost of reaching
over. These customers commonly crave security and require con- and surveying particular individuals, the company thought
tinuous information on retirement funds. The bank believes that it that reaching individuals in young populations would be
is highly unlikely that many customers in this age group will use easiest. In a recently completed survey, however, Sophisti-
Web banking, but the bank desires to learn the needs of this age cated Surveys learned that this assumption was wrong. The
group for the future. AmeriBank wants to interview 2,000 cus- new costs for surveying the 18-to-25-year-old population are
tomers with at least 20 percent from the first age group, at least listed below.
27.5 percent from the second age group, at least 15 percent from the
third age group, and at least 15 percent from the fourth age group. Region Cost per Person
Rob understands that some customers are uncomfortable
with using the Internet. He therefore wants to ensure that the Silicon Valley $6.50
survey includes a mix of customers who know the Internet well Big cities 6.75
and those that have less exposure to the Internet. To ensure Small towns 7.00
that AmeriBank obtains the correct mix, he wants to interview
at least 15 percent of customers from the Silicon Valley where Given the new costs, what is the new bid?
Internet use is high, at least 35 percent of customers from big f. To ensure the desired sampling of individuals, Rob imposes
cities where Internet use is medium, and at least 20 percent of even stricter requirements. He fixes the exact percentage of
customers from small towns where Internet use is low. people that should be surveyed from each population. The
Sophisticated Surveys is one of three survey companies requirements are listed next.
competing for the project. It has performed an initial analysis
of these survey requirements to determine the cost of surveying
different populations. The costs per person surveyed are listed in Population Percentage of People Surveyed
the following table.
18 to 25 25%
26 to 40 35
Age Group 41 to 50 20
51 and over 20
Region 18 to 25 26 to 40 41 to 50 51 and over Silicon Valley 20
Big cities 50
Silicon Valley $4.75 $6.50 $6.50 $5.00 Small towns 30
Big cities 5.25 5.75 6.25 6.25
Small towns 6.50 7.50 7.50 7.25
By how much would these new requirements increase the cost
of surveying for Sophisticated Surveys? Given the 15 ­percent
Sophisticated Surveys explores the following options profit margin, what would Sophisticated Surveys bid?
cumulatively.

Case 3-5
Assigning Students to Schools
The Springfield School Board has made the decision to close school students to the three remaining middle schools. The
one of its middle schools (sixth, seventh, and eighth grades) at school district provides busing for all middle school students
the end of this school year and reassign all of next year’s middle who must travel more than approximately a mile, so the school

hiL18920_ch03_064-123.indd 119 12/08/17 06:54 AM

 Final PDF to printer

120  Chapter Three   Linear Programming: Formulation and Applications

board wants a plan for reassigning the students that will mini- How much does this increase the total busing cost? (This line
mize the total busing cost. The annual cost per student for busing of analysis will be pursued more rigorously in Case 7-3.)
from each of the six residential areas of the city to each of the
schools is shown in the following table (along with other basic The school board is considering eliminating some busing to
data for next year), where 0 indicates that busing is not needed reduce costs. Option 1 is to only eliminate busing for students trav-
and a dash indicates an infeasible assignment. eling 1 to 1.5 miles, where the cost per student is given in the table

Busing Cost per Student

Number of Percentage Percentage Percentage

Area Students in 6th Grade in 7th Grade in 8th Grade School 1 School 2 School 3

1 450 32 38 30 $300 $ 0 $700

2 600 37 28 35 — 400 500
3 550 30 32 38 600 300 200
4 350 28 40 32 200 500 —
5 500 39 34 27    0 — 400
6 450 34 28 38 500 300    0
School capacity: 900 1,100 1,000

The school board also has imposed the restriction that as $200. Option 2 is to also eliminate busing for students traveling
each grade must constitute between 30 and 36 percent of each 1.5 to 2 miles, where the estimated cost per student is $300.
school’s population. The above table shows the percentage of
each area’s middle school population for next year that falls into d. Revise the model from part a to fit Option 1, and solve. Com-
each of the three grades. The school attendance zone boundar- pare these results with those from part b, including the reduc-
ies can be drawn so as to split any given area among more than tion in total busing cost.
one school, but assume that the percentages shown in the table e. Repeat part d for Option 2.
will continue to hold for any partial assignment of an area to a The school board now needs to choose among the three alter-
school. native busing plans (the current one or Option 1 or Option 2).
You have been hired as a management science consultant One important factor is busing costs. However, the school board
to assist the school board in determining how many students in also wants to place equal weight on a second factor: the inconve-
each area should be assigned to each school. nience and safety problems caused by forcing students to travel
by foot or bicycle a substantial distance (more than a mile, and
a. Formulate and solve a linear programming model for this
especially more than 1.5 miles). Therefore, they want to choose
problem.
a plan that provides the best trade-off between these two factors.
b. What is your resulting recommendation to the school board?
f. Use your results from parts b, d, and e to summarize the key
After seeing your recommendation, the school board information related to these two factors that the school board
expresses concern about all the splitting of residential areas needs to make this decision.
among multiple schools. They indicate that they “would like to g. Which decision do you think should be made? Why?
keep each neighborhood together.”
Note: This case will be continued in later chapters (Cases 5-4
c. Adjust your recommendation as well as you can to enable
and 7-3), so we suggest that you save your analysis, including
each area to be assigned to just one school. (Adding this
your basic spreadsheet model.
restriction may force you to fudge on some other constraints.)

Case 3-6
Reclaiming Solid Wastes
The Save-It Company operates a reclamation center that col- of the total weight for that product grade.) For each of the
lects four types of solid waste materials and then treats them two higher grades, a fixed percentage is specified for one of
so that they can be amalgamated (treating and amalgamating the materials. These specifications are given in the first table
are separate processes) into a salable product. Three differ- along with the cost of amalgamation and the selling price for
ent grades of this product can be made, depending on the mix each grade.
of the materials used. (See the first table.) Although there is The reclamation center collects its solid waste materials
some flexibility in the mix for each grade, quality standards from some regular sources and so is normally able to maintain a
specify the minimum or maximum amount of the materials steady rate for treating them. The second table gives the quanti-
allowed in that product grade. (This minimum or maximum ties available for collection and treatment each week, as well as
amount is the weight of the material expressed as a percentage the cost of treatment, for each type of material.

hiL18920_ch03_064-123.indd 120 12/08/17 06:54 AM

 Final PDF to printer

Case 3-7   Project Pickings   121

Amalgamation Selling Price

Grade Specification Cost per Pound per Pound

Material 1: Not more than 30% of the total

A Material 2: Not less than 40% of the total $3.00 $8.50
Material 3: Not more than 50% of the total
Material 4: Exactly 20% of the total
Material 1: Not more than 50% of the total
B Material 2: Not less than 10% of the total 2.50 7.00
Material 4: Exactly 10% of the total

C Material 1: Not more than 70% of the total 2.00 5.50

Available Treatment Cost

Material Pounds/Week per Pound Additional Restrictions

1 3,000 $3.00 1. For each material, at least half

2 2,000 6.00 of the pounds/week available
3 4,000 4.00 should be collected and
4 1,000 5.00 treated.
2. $30,000 per week should be
used to treat these materials.

The Save-It Company is solely owned by Green Earth, maximize the total weekly profit (total sales income minus total
an organization that is devoted to dealing with environmen- amalgamation cost).
tal issues; Save-It’s profits are all used to help support Green
Earth’s activities. Green Earth has raised contributions and a. Formulate this problem in linear programming terms by
grants, amounting to $30,000 per week, to be used exclusively to identifying all the activities, resources, benefits, and fixed
cover the entire treatment cost for the solid waste materials. The requirements that lurk within it.
board of directors of Green Earth has instructed the management b. Formulate and solve a spreadsheet model for this linear pro-
of Save-It to divide this money among the materials in such a gramming problem.
way that at least half of the amount available of each material is c. Express this linear programming model in the spreadsheet in
actually collected and treated. These additional restrictions are algebraic form.
listed in the second table.
Within the restrictions specified in the two tables, manage-
ment wants to allocate the materials to product grades so as to

Case 3-7
Project Pickings
Tazer, a pharmaceutical manufacturing company, entered the Now Tazer is beginning to fear the pressure of competition.
pharmaceutical market 15 years ago with the introduction of six Tazer knows that once the patent expires in five years, generic
new drugs. Five of the six drugs were simply permutations of drug manufacturing companies will swarm into the market like
existing drugs and therefore did not sell very heavily. The sixth vultures. Historical trends show that generic drugs decrease
drug, however, addressed hypertension and was a huge success. sales of branded drugs by 75 percent.
Since Tazer had a patent on the hypertension drug, it experienced Tazer is therefore looking to invest significant amounts of
no competition, and profits from the hypertension drug alone money in research and development this year to begin the search
kept Tazer in business. Pharmaceutical patents remain in force for a new breakthrough drug that will offer the company the
for 20 years, so this one has five more years before it expires. same success as the hypertension drug. Tazer believes that if
During the past 15 years, Tazer continued a moderate amount the company begins extensive research and development now,
of research and development, but it never stumbled upon a drug the probability of finding a successful drug shortly after the
as successful as the hypertension drug. One reason is that the expiration of the hypertension patent will be high.
company never had the motivation to invest heavily in innovative As head of research and development at Tazer, you are
research and development. The company was riding the profit wave responsible for choosing potential projects and assigning project
generated by its hypertension drug and did not feel the need to directors to lead each of the projects. After researching the needs
commit significant resources to finding new drug breakthroughs. of the market, analyzing the shortcomings of current drugs,

hiL18920_ch03_064-123.indd 121 12/08/17 06:54 AM

 Final PDF to printer

122  Chapter Three   Linear Programming: Formulation and Applications

and interviewing numerous scientists concerning the promising Project Up: 20

areas of medical research, you have decided that your depart- Project Stable: 450
ment will pursue five separate projects, which are listed below: Project Choice: 451
Project Up: Develop a more effective antidepressant that Project Hope: 39
does not cause serious mood swings. Project Release: 40
Project Stable: Develop a drug that addresses
Under these new conditions with just four senior scientists,
manic-depression.
which scientists will lead which projects to maximize preferences?
Project Choice: Develop a less intrusive birth control
e. Do you support the assignments found in part d? Why or why
method for women.
not?
Project Hope: Develop a vaccine to prevent HIV infection.
f. Now you again consider all five scientists. You decide, how-
Project Release: Develop a more effective drug to lower ever, that several scientists cannot lead certain projects. In par-
blood pressure. ticular, Dr. Mickey does not have experience with research on
For each of the five projects, you are only able to specify the the immune system, so he cannot lead Project Hope. His fam-
medical ailment the research should address since you do not ily also has a history of manic-depression, and you feel that he
know what compounds will exist and be effective without would be too personally involved in Project Stable to serve as
research. an effective project leader. Dr. Mickey therefore cannot lead
You also have five senior scientists to lead the five projects. Project Stable. Dr. Kvaal also does not have experience with
You know that scientists are very temperamental people and will research on the immune system and cannot lead Project Hope.
only work well if they are challenged and motivated by the proj- In addition, Dr. Kvaal cannot lead Project Release because he
ect. To ensure that the senior scientists are assigned to projects does not have experience with research on the cardiovascular
they find motivating, you have established a bidding system for system. Finally, Dr. Rollins cannot lead Project Up because
the projects. You have given each of the five scientists 1,000 bid her family has a history of depression and you feel she would
points. They assign bids to each project, giving a higher number be too personally involved in the project to serve as an effec-
of bid points to projects they most prefer to lead. tive leader. Because Dr. Mickey and Dr. Kvaal cannot lead
The following table provides the bids from the five senior two of the five projects, they each have only 600 bid points.
scientists for the five individual projects. Dr. Rollins has only 800 bid points because she cannot lead

Project Dr. Kvaal Dr. Zuner Dr. Tsai Dr. Mickey Dr. Rollins

Project Up 100 0 100 267 100

Project Stable 400 200 100 153 33
Project Choice 200 800 100 99 33
Project Hope 200 0 100 451 34
Project Release 100 0 600 30 800

You decide to evaluate a variety of scenarios you think are one of the five projects. The following table provides the new
likely. bids of Dr. Mickey, Dr. Kvaal, and Dr. Rollins.

a. Given the bids, you need to assign one senior scientist to each
of the five projects to maximize the preferences of the scien- Project Dr. Mickey Dr. Kvaal Dr. Rollins
tists. What are the assignments?
Project Up 300 86 Can’t lead
b. Dr. Rollins is being courted by Harvard Medical School to
Project Stable Can’t lead 343 50
accept a teaching position. You are fighting desperately to
Project Choice 125 171 50
keep her at Tazer, but the prestige of Harvard may lure her Project Hope Can’t lead Can’t lead 100
away. If this were to happen, the company would give up the Project Release 175 Can’t lead 600
project with the least enthusiasm. Which project would not
be done?
c. If Dr. Rollins leaves, you do not want to sacrifice any project Which scientists should lead which projects to maximize
since researching only four projects decreases the probability preferences?
of finding a breakthrough new drug. You decide that either g. You decide that Project Hope and Project Release are too
Dr. Zuner or Dr. Mickey could lead two projects. Under these complex to be led by only one scientist. Therefore, each of
new conditions with just four senior scientists, which scien- these projects will be assigned two scientists as project lead-
tists will lead which projects to maximize preferences? ers. You decide to hire two more scientists in order to staff
d. After Dr. Zuner was informed that she and Dr. Mickey are all projects: Dr. Arriaga and Dr. Santos. Because of religious
being considered for two projects, she decided to change reasons, neither of them want to lead Project Choice and so
her bids. Dr. Zuner’s new bids for each of the projects they assign 0 bid points to this project. The next table lists all
are shown next. projects, scientists, and their bids.

hiL18920_ch03_064-123.indd 122 12/08/17 06:54 AM

 Final PDF to printer

Case 3-7   Additional Cases   123

Project Dr. Kvaal Dr. Zuner Dr. Tsai Dr. Mickey Dr. Rollins Dr. Arriaga Dr. Santos

Project Up 86 0 100 300 Can’t lead 250 111

Project Stable 343 200 100 Can’t lead 50 250 1
Project Choice 171 800 100 125 50 0 0
Project Hope Can’t lead 0 100 Can’t lead 100 250 333
Project Release Can’t lead 0 600 175 600 250 555

Which scientists should lead which projects to maximize

preferences?
Additional Cases
h. Do you think it is wise to base your decision in part g Additional cases for this chapter are also available at the Univer-
only on an optimal solution for a variant of an assignment sity of Western Ontario Ivey School of Business website, cases.
problem? ivey.uwo.ca/cases, in the segment of the CaseMate area desig-
nated for this book.

hiL18920_ch03_064-123.indd 123 12/08/17 06:54 AM