POLYTECHNIC UNIVERSITY OF THE PHILIPPINES 1

WEEK 5
MODELING AND SOLVING LP PROBLEMS IN A SPREADSHEET

Overview:

Very few real-world LP problems involve only two decision variables. So, the graphical
solution approach is of limited value in solving LP problems. However, the discussion of
two-variable problems provides a basis for understanding the issues involved in all LP
problems and the general strategies for solving them.

Every solvable LP problem has a feasible region, and an optimal solution to the
problem can be found at some extreme point of this region (assuming the problem is not
unbounded). This is true of all LP problems regardless of the number of decision variables.
Although it is fairly easy to graph the feasible region for a two-variable linear programming
problem, it is difficult to visualize or graph the feasible region of an LP problem with three
variables because such a graph is three-dimensional. If there are more than three variables,
it is virtually impossible to visualize or graph the feasible region for an LP problem because
such a graph involves more than three dimensions.

Several mathematical techniques exist to solve LP problems involving almost any

number of variables without visualizing or graphing their feasible regions. These techniques
are now built into spreadsheet packages in a way that makes solving LP problems a fairly
simple task. So, using the appropriate computer software, you can solve almost any LP
problem easily. The main challenge is ensuring that you formulate the LP problem correctly
and communicate this formulation to the computer accurately.

In this chapter, we are going to learn how to solve LP problems using spreadsheets.

Module Objectives:

After successful completion of this module, you should be able to:

1. Recognize the implementation of LP models using spreadsheet;

2. Develop understanding of the concept of spreadsheet platforms and software that can
be used to solve linear programming problems;
3. Enumerate and explain the goals and guidelines in implementing effective
spreadsheet models; and
 POLYTECHNIC UNIVERSITY OF THE PHILIPPINES 2

4. Enumerate the different business problems that can be addressed by linear

programming and how it can help in better informed decisions in these industries.

Course Materials:

1.1. Spreadsheet Solvers

The importance of LP (and optimization in general) is underscored by the fact that all
major spreadsheet packages come with built-in spreadsheet optimization tools called
solvers. Microsoft Excel can be used to illustrate how spreadsheet solvers can solve
optimization problems.
Without using a spreadsheet, we can still solve optimization problems by using
specialized mathematical programming packages like:
LINDO (Lindo Systems, Inc.) - https://www.lindo.com/
CPLEX Optimizer (IBM) - https://www.ibm.com/analytics/cplex-optimizer
Xpress-MP

Researchers and businesses use these packages to solve extremely large problems
that do not fit conveniently in a spreadsheet.

Learn more:

Ragsdale, C. T. (2012). Spreadsheet modeling & decision analysis: a practical

introduction to management science. South Western, Cengage Learning. (pp. 46-47)

1.2. Solving LP Problems in a Spreadsheet

Recalling Howie’s problem in Lesson 2:

Howie owns and operates Blue Ridge Hot Tubs, a company that sells two models of
hot tubs: the Aqua-Spa and the Hydro-Lux. Howie purchases prefabricated fiberglass hot
tub shells and installs a common water pump and the appropriate amount of tubing into
each hot tub.

Every Aqua-Spa requires 9 hours of labor and 12 feet of tubing; every Hydro-Lux
requires 6 hours of labor and 16 feet of tubing. Demand for these products is such that each
Aqua-Spa produced can be sold to generate a profit of $350, and each Hydro-Lux produced
can be sold to generate a profit of $300. The company expects to have 200 pumps, 1,566
hours of labor, and 2,880 feet of tubing available during the next production cycle. The
problem is to determine the optimal number of Aqua-Spas and Hydro-Luxes to produce in
order to maximize profits.
 POLYTECHNIC UNIVERSITY OF THE PHILIPPINES 3

The following LP model was formulated in Lesson 2. In this model, X1 represents the
number of Aqua-Spas to be produced, and X2 represents the number of Hydro-Luxes to be
produced:

MAX: 350X1 + 300X2 } Profit

Subject to: 1X1 + 1X2 < 200 } Pump Constraint
9X1 + 6X2 < 1,566 } Labor Constraint
12x1 + 16X2 < 2,880 } Tubing Constraint
1X1 > 0 } Simple Lower Bound
1X2 > 0 } Simple Lower Bound

To solve this problem in a spreadsheet, we need to implement or build the model in a

spreadsheet.

Learn from:

Ragsdale, C. T. (2012). Spreadsheet modeling & decision analysis: a practical

introduction to management science. South Western, Cengage Learning. (p. 47)

1.3. The Steps in Implementing an LP Model in a Spreadsheet

The following four steps summarize what must be done to implement any LP problem
in a spreadsheet:

1. Organize the data for the model on the spreadsheet.

2. Reserve separate cells in the spreadsheet to represent each decision variable in the
algebraic model.
3. Create a formula in a cell in the spreadsheet that corresponds to the objective
function in the algebraic model.
4. For each constraint, create a formula in a separate cell in the spreadsheet that
corresponds to the left-hand-side (LHS) of the constraint.

Learn more:

Ragsdale, C. T. (2012). Spreadsheet modeling & decision analysis: a practical

introduction to management science. South Western, Cengage Learning. (pp. 47-48)
 POLYTECHNIC UNIVERSITY OF THE PHILIPPINES 4

1.4. A Spreadsheet Model for the Blue Ridge Hot Tubs Problem

Learn from:

Ragsdale, C. T. (2012). Spreadsheet modeling & decision analysis: a practical

introduction to management science. South Western, Cengage Learning. (pp. 49-52)

1.5. How Solver Views the Model

Three components of spreadsheet model for Solver:

1. Objective cell
2. Variable cells
3. Constraint cells

Learn more:

Ragsdale, C. T. (2012). Spreadsheet modeling & decision analysis: a practical

introduction to management science. South Western, Cengage Learning. (pp. 53-54)

1.6. Using Risk Solver Platform

After implementing an LP model in a spreadsheet, we still need to solve the problem

being modeled. To do this, we must first indicate to Solver which cells in the spreadsheet
represent the objective function, the decision variables, and the constraints. To invoke
Solver, choose the Risk Solver Platform tab on the ribbon to display the Risk Solver task
pane.

Risk Solver Platform offers a number of analytical tools (for example, Sensitivity
analysis, Optimization, Simulation, Discriminant Analysis, Decision Trees).

Define the objective cell

Define the variable cells
Define the constraint cells
Define the non-negativity conditions
Review the model
Use other options
Solve the problem
 POLYTECHNIC UNIVERSITY OF THE PHILIPPINES 5

Learn more:

Ragsdale, C. T. (2012). Spreadsheet modeling & decision analysis: a practical

introduction to management science. South Western, Cengage Learning. (pp. 54-63)

1.7. Using Excel’s Built-in Solver

Excel’s built-in Solver is easy to use and is capable of solving most of the optimization
problems. However, it lacks a number of powerful and useful features offered by Risk Solver
Platform.

To use the built-in Solver, you must identify the objective cell (and the desired direction
of optimization), the variables cells, and any constraints. The Solver dialog also allows you
to select a solving method (analogous to selections available on the Engine tab in Risk
Solver’s task pane). You then click the Solve button to solve the problem.

For each of the standard optimization problems, we will identify the objective cell (and
whether it should be maximized or minimized), the variable cells, and the constraints. Using
that information, you can use either Excel’s built-in Solver or Risk Solver Platform to solve
the problems.

1.8. Goals and Guidelines for Spreadsheet Design

To achieve the end result of a logical spreadsheet design, your modeling efforts should
be directed toward the following goals:

Communication

A spreadsheet’s primary business purpose is communicating information to

managers. As such, the primary design objective in most spreadsheet modeling tasks
is to communicate the relevant aspects of the problem at hand in as clear and
intuitively appealing a manner as possible.

Reliability

The output a spreadsheet generates should be correct and consistent. This

has an obvious impact on the degree of confidence a manager places in the results
of the modeling effort.
 POLYTECHNIC UNIVERSITY OF THE PHILIPPINES 6

Auditability

A manager should be able to retrace the steps followed to generate the

different outputs from the model in order to understand the model and verify results
Models that are set up in an intuitively appealing, logical layout tend to be the most
auditable.

Modifiability

The data and assumptions upon which we build spreadsheet models can
change frequently. A well-designed spreadsheet should be easy to change or enhance
in order to meet dynamic user requirements.

Some practical suggestions and guidelines for creating effective spreadsheet models
are:

Organize the data, then build the model around the data.
Do not embed numeric constants in formulas.
Things which are logically related should be arranged in close physical
proximity to one another and in the same columnar or row orientation.
A design that results in formulas that can be copied is probably better than one
that does not.
Column or row totals should be in close proximity to the columns or rows being
totaled.
The English-reading human eye scans left to right, top to bottom.
Use color, shading, borders, and protection to distinguish changeable
parameters other elements of the model.
Use text boxes and cell comments to document various elements of the model.

Learn more:

Ragsdale, C. T. (2012). Spreadsheet modeling & decision analysis: a practical

introduction to management science. South Western, Cengage Learning. (pp. 64-66)

1.9. Make vs Buy Decisions

LP is particularly well-suited to problems where scarce or limited resources must be

allocated or used in an optimal manner. Numerous examples of these types of problems
occur in manufacturing organizations.
 POLYTECHNIC UNIVERSITY OF THE PHILIPPINES 7

For example, LP might be used to determine how the various components of a job
should be assigned to multipurpose machines in order to minimize the time it takes to
complete the job. As another example, a company might receive an order for several items
that it cannot fill entirely with its own production capacity. In such a case, the company must
determine which items to produce and which items to subcontract (or buy) from an outside
supplier.

See example of Make vs. Buy Decision:

Ragsdale, C. T. (2012). Spreadsheet modeling & decision analysis: a practical

introduction to management science. South Western, Cengage Learning. (pp. 67-72)

1.10. An Investment Problem

There are numerous problems in the area of finance for which various optimization
techniques can be applied. These problems often involve attempting to maximize the return
on an investment while meeting certain cash flow requirements and risk constraints.
Alternatively, we may want to minimize the risk on an investment while maintaining a certain
level of return.

See example of an investment problem:

Ragsdale, C. T. (2012). Spreadsheet modeling & decision analysis: a practical

introduction to management science. South Western, Cengage Learning. (pp. 72-76)

1.11. A Transportation Problem

Many transportation and logistics problems businesses face fall into a category of
problems known as network flow problems.

See example of a transportation problem:

Ragsdale, C. T. (2012). Spreadsheet modeling & decision analysis: a practical

introduction to management science. South Western, Cengage Learning. (pp. 76-82)

1.12. A Blending Problem

Many business problems involve determining an optimal mix of ingredients. For

example, major oil companies must determine the least costly mix of different crude oils and
other chemicals to blend together to produce a certain grade of gasoline. Lawn care
 POLYTECHNIC UNIVERSITY OF THE PHILIPPINES 8

companies must determine the least costly mix of chemicals and other products to blend
together to produce different types of fertilizer. The following is another example of a
common blending problem faced in the U.S. agricultural industry, which annually produces
goods valued at approximately $200 billion.

See example of a blending problem:

Ragsdale, C. T. (2012). Spreadsheet modeling & decision analysis: a practical

introduction to management science. South Western, Cengage Learning. (pp. 83-89)

1.13. A Production and Inventory Planning Problem

One of the most fundamental problems facing manufacturing companies is that of

planning their production and inventory levels. This process considers demand forecasts
and resource constraints for the next several time periods and determines production and
inventory levels for each of these time periods to meet the anticipated demand in the most
economical way. As the following example illustrates, the multiperiod nature of these
problems can be handled very conveniently in a spreadsheet to greatly simplify the
production planning process.

See example of a production and inventory problem:

Ragsdale, C. T. (2012). Spreadsheet modeling & decision analysis: a practical

introduction to management science. South Western, Cengage Learning. (pp. 89-95)

1.14. A Multiperiod Cash Flow Problem

Numerous business problems involve decisions that have a ripple effect on future
decisions. Many financial decisions involve multiple time periods because the amount of
money invested or spent at one point in time directly affects the amount of money available
in subsequent time periods. In these types of multiperiod problems, it can be difficult to
account for the consequences of a current decision on future time periods without an LP
model. The formulation of such a model is illustrated next in an example from the world of
finance.

See example of a multi-period cash flow problem:

Ragsdale, C. T. (2012). Spreadsheet modeling & decision analysis: a practical

introduction to management science. South Western, Cengage Learning. (pp. 95-106)
 POLYTECHNIC UNIVERSITY OF THE PHILIPPINES 9

1.15. Data Envelopment Analysis

Managers are often interested in determining how efficiently various units within a
company operate. Similarly, investment analysts may be interested in comparing the
efficiency of several competing companies within an industry. Data Envelopment Analysis
(DEA) is an LP-based methodology for performing this type of analysis. DEA determines
how efficiently an operating unit (or company) converts inputs to outputs when compared
with other units.

See example of a DEA:

Ragsdale, C. T. (2012). Spreadsheet modeling & decision analysis: a practical

introduction to management science. South Western, Cengage Learning. (pp. 106-
114)

Source:

Ragsdale, C. T. (2012). Spreadsheet modeling & decision analysis: a practical introduction

to management science. South Western, Cengage Learning.

Activities/Assessment:

1. In creating the spreadsheet models for the problems, cells in the spreadsheets had to
be reserved to represent each of the decision variables in the algebraic models. We
reserved these cells in the spreadsheets by entering values of zero in them. Why didn’t
we place some other value or formula in these cells? Would doing so have made any
difference?

2. Compare and contrast MS Excel’s built-in solver platform and Risk Solver platform.
Which of these two do you think is more sophisticated to use and why?

3. Four goals should be considered when trying to design an effective spreadsheet

model: communication, reliability, auditability, and maintainability. We also noted that
a spreadsheet design that results in formulas that can be copied is usually more
effective than other designs. Briefly describe how using formulas that can be copied
supports the four spreadsheet modeling goals.

4. List and explain the guidelines in creating effective spreadsheet models.

5. Identify and explain briefly the different business problems and decisions that can be
addressed with linear programming.