Evolver7 en

User’s Guide
Evolver
The Genetic Algorithm Solver for
Microsoft Excel
Version 7
July, 2015
Palisade Corporation
798 Cascadilla St.
Ithaca, NY USA 14850
(607) 277-8000
(607) 277-8001 (fax)
http://www.palisade.com (website)
sales@palisade.com (e-mail)
Copyright Notice
Copyright  2008-2015, Palisade Corporation
Trademark Acknowledgments
IBM is a registered trademark of International Business Machines, Inc.
Palisade, Evolver, TopRank, BestFit and RISKview are registered trademarks of Palisade
Corporation.
RISK is a trademark of Parker Brothers, Division of Tonka Corporation, and is used under
license.
Microsoft, Excel and Windows are registered trademarks of Microsoft Corporation.
Table of Contents
Chapter 1: Welcome to Evolver 1

What Is Evolver? ...................................................................................1
What Is Optimization? ........................................................................1
Why Use Evolver? ................................................................................2
How does Evolver work? ....................................................................4
Efficient Frontier Analysis .................................................................5
Chapter 2: Getting Started 7
Introduction...................................................................................... 7
About This Version..............................................................................7
Evolver Professional and Industrial Versions ................................7
Working with your Operating Environment ..................................7
If You Need Help .................................................................................8
Evolver System Requirements .........................................................10
Installation Instructions ................................................................ 11
General Installation Instructions ....................................................11
The DecisionTools Suite ...................................................................11
Software Activation ...........................................................................12
Instructional Materials .................................................................. 13
Example Spreadsheets .......................................................................13
Quick Start Tutorials .........................................................................13
Guided Tour Videos ..........................................................................13
Examples and Videos for the XDK .................................................13
Chapter 3: Evolver Tutorial 15

Loading Evolver..................................................................................16
The Evolver Model Dialog ...............................................................17
Selecting the Target Cell ...................................................................18
Adding Adjustable Cell Ranges ......................................................18
Constraints...........................................................................................20
Other Evolver Options ......................................................................23
Running the Optimization ...............................................................25
i
Chapter 4: Evolver Command Reference 31
Evolver Icons................................................................................. 31
Evolver Commands ...................................................................... 33

Model Definition Command ........................................................... 33
Settings Command ............................................................................ 48
Start Optimization Command......................................................... 59
Evolver Watcher ................................................................................. 60
Application Settings Command ..................................................... 69
Constraint Solver Command ........................................................... 70
Appendix A: Optimization 73
Introduction ................................................................................... 73
Optimization Methods ..................................................................... 73
Types of Models ................................................................................ 75
Genetic Algorithms ....................................................................... 77
Introduction ........................................................................................ 77
History ................................................................................................. 77
A Biological Example ........................................................................ 79
A Digital Example ............................................................................. 80
How Evolver's Genetic Algorithm Optimization is
Implemented .................................................................................. 83
OptQuest ........................................................................................ 87
Introduction ........................................................................................ 87
Linear Constraints ............................................................................. 87
Nonlinear Constraints ...................................................................... 87
Improving Speed ........................................................................... 89
Appendix B: Automating Evolver 91
Appendix C: Troubleshooting / Q&A 93
Appendix D: Additional Resources 95

Books ................................................................................................... 95
Articles................................................................................................. 96
Magazines & Newsletters ................................................................ 97
ii
Appendix E: Glossary 99
Index 103
Table of Contents iii

iv
Chapter 1: Welcome to Evolver
What Is Evolver?
The Evolver add-in enables you to find optimal solutions to virtually
any type of deterministic problem, that is, any problem that doesn’t
include explicit uncertainty. Simply put, Evolver finds the best inputs
for optimizing a specified output. You can use Evolver to find the
right mix, ordering, or grouping of variables that produces the
highest profit, the lowest cost, the shortest project completion time,
the lowest risk, and so on. You use Evolver in Excel by first setting up
a typical spreadsheet model in Excel and then evoking Evolver to find
an optimal solution.
Excel provides all of the formulas, functions, graphs, and macro
capabilities that most users need to create realistic models of their
problems. Evolver provides the interface to describe what you want
to optimize, and it provides the engines to perform the optimization.
Together, they can find optimal solutions to virtually any problem
that can be modeled.
Evolver is offered in two editions, Professional and Industrial, to
allow you to select the optimizer with the capacity you need. The
Professional edition allows up to 250 adjustable variables per model,
while the Industrial edition allows unlimited variables.
What Is Optimization?
Optimization is the process of trying to find the “best” solution to a
problem. For example, a company might have three manufacturing
plants, each manufacturing different quantities of different goods.
Given the cost for each plant to produce each good, the costs for each
plant to ship to each store, and the limitations of each plant, the goal
is to find the optimal way to adequately meet the demand of local
retail stores while minimizing the transportation costs.
This is the type of problem that optimization is designed to answer,
but it is certainly not the only type. Optimization models of various
types appear in the fields of operations management, finance,
marking, economics, engineering, and others. They all share a
1
common goal: to select values of inputs to maximize (or minimize) an
objective value, possibly subject to constraints.
Traditionally, most optimization models have been deterministic,
meaning that they don’t include any explicit uncertainty. For
example, if a demand forecast is part of the model, it is entered as a
fixed number, with no uncertainty. Evolver is designed to optimize
only deterministic models. If your model include explicit uncertainty,
in the form of probability distributions for some of the variables, you
can use Palisade’s companion product RISKOptimizer, which is
integrated into the @RISK add-in.
Evolver and RISKOptimizer are quite similar in many ways, but they
differ in one fundamental way. When Evolver needs to check how
“good” a potential solution is, all it has to do is evaluate spreadsheet
formulas, such as the formula for profit, once. In stark contrast,
RISKOptimizer has to run a simulation to discover the range of
possible profits for that particular solution. So RISKOptimizer is able
to solve a larger class of problems, those with uncertainty, but it tends
to be much slower than Evolver.
Why Use Evolver?

No More When you are dealing with large numbers of interacting variables,
Guessing and you are trying to find the best mix, the right order, or the
optimum grouping of those variables, you might be tempted to just
take an “educated guess.” A surprising number of people assume
that any kind of modeling and analysis beyond guessing will require
complicated programming, or confusing statistical or mathematical
algorithms. However, an optimal solution might save millions of
dollars, thousands of gallons of scarce fuel, months of wasted time,
and so on. Now that powerful desktop computers are increasingly
affordable, and software like Excel and Evolver are readily available,
there is little reason to guess at solutions, or waste valuable time
trying out many scenarios by hand.
More Accurate, Evolver allows you to use the entire range of Excel functions to build
More more realistic models of any system. When you use Evolver, you do
Meaningful not have to “compromise” the accuracy of your model because the
algorithm you are using cannot handle real world complexities. Less
powerful solvers (statistical and linear programming tools) force users
to make assumptions about the way the variables in their problem
interact, thereby forcing users to build over-simplified, unrealistic
models of their problem. By the time users have simplified a system
enough that these solvers can be used, the resulting solution is often
the solution to the wrong problem and hence not useul. Any
problems involving large numbers of variables, nonlinear functions,
2
lookup tables, if-then statements, or database queries cannot be
solved by these methods, no matter how simply you try to design
your model.
More Flexible There are many optimization algorithms that work well at solving
small, simple linear and nonlinear types of problems. For larger or
more complex problems, you might be able to write specific,
customized algorithms to achieve good results, but this can require a
lot of research and development. Even then, the resulting program
would probably require modification each time your model changed.
In addition to being able to solve problems with continuous variables,
Evolver can aso solve most combinatorial problems. These are
problems where the variables must be shuffled (permuted) or
combined with each other. For example, choosing the batting order
for a baseball team is a combinatorial problem; it is a question of
swapping players’ positions in the lineup. Evolver can find the
optimal ordering of tasks to be performed as part of a project,
evaluating only solutions that meet specified precedence constraints
(that is, constraints that require certain tasks to be performed before
others). Complex scheduling problems are also combinatorial.
More Powerful Evolver finds better solutions. Most software derives optimum
solutions mathematically and systematically. Too often these
methods are limited to taking an existing solution and searching for
the closest answer that is better. This “local” solution may be far from
the optimal solution. Evolver intelligently samples the entire realm of
possibilities, resulting in a much better “global” solution.
Easier to Use In spite of its power and flexibility advantages, Evolver remains easy
to use because an understanding of the optimization techniques it
uses is not necessary. Evolver doesn’t require you to understand the
“nuts and bolts” of your problem; it just requires a spreadsheet model
that can evaluate how good different solutions are. You just select the
spreadsheet cells that contain the decision variables and tell Evolver
what you want to optimize. Evolver intelligently hides the difficult
technology, automating the “what-if” process of analyzing a problem.
Although there have been many commercial programs developed for
mathematical programming and model-building, spreadsheets are by
far the most popular, with literally millions of users. With their
intuitive row and column format, spreadsheets are easier to set up
and maintain than other dedicated packages. They are also more
compatible with other programs such as word processors and
databases, and they offer more built-in formulas, formatting options,
graphing, and macro capabilities than any of the stand-alone
packages. Because Evolver is an add-in to Microsoft Excel, you have
access to the entire range of functions and development tools to easily
build more realistic models of your problems.
Cost Effective Many companies have hired trained consultants to provide
customized optimization systems. Such systems often perform quite
well, but they can require many months and a large investment to
develop and implement. These systems are also difficult to learn, and
therefore require costly training and constant maintenance. If your
system must be altered, you might need to develop a whole new
algorithm to find optimal solutions. For a considerably smaller
investment, Evolver supplies the most powerful optimization
algorithms available and allows for quick and accurate solutions to a
wide variety of problems. Because it works in an intuitive and
familiar environment, there is virtually no costly training and
maintenance.
You might even want to add Evolver’s optimization power to your
own custom programs. In just a few days, you could use Visual Basic
to develop your own scheduling, distribution, manufacturing, or
financial management system. See the Evolver Developer Kit for
details on developing an Evolver-based application.
How does Evolver work?

Evolver uses OptQuest and Genetic Algorithms to search for
optimum solutions to a problem. In addition, if it determines that a
problem is linear, it uses linear programming algorithms.
OptQuest OptQuest uses metaheuristic, mathematical optimization, and neural
network components to guide the search for optimal solutions to
decision and planning problems of all types. OptQuest’s methods
integrate state-of-the-art metaheuristic procedures, including Tabu
Search, Neural Networks and Scatter Search into a single composite
method.
Genetic Genetic Algorithms (GAs) provide an alternative solution method.
Algorithms GAs mimic Darwinian principles of natural selection by creating an
environment where hundreds of possible solutions to a problem can
compete with one another, and only the “fittest” survive. Just as in
biological evolution, each solution can pass along its good “genes”
through “offspring” solutions so that the entire population of
solutions will continue to evolve better solutions.
4
Linear Evolver uses special-purpose algorithm to solve linear models much
Programming more quickly than if they were solved with general-purpose
algorithms. (This requires that Optimization Mode be left as
Automatic in the Engine tab of the Optimization Settings dialog).
Efficient Frontier Analysis

Efficient frontier analysis is used when there are two competing goals.
You choose one of them as the target for an optimization and
constrain the other to be no “worse” than a specified limit. Then you
perform a sequence of optimizations, each time changing the limit
value for the constraint.
Although efficient frontier analysis is generally used for financial
portfolio optimization, it can be applied to a variety of problems
where there are two competing goals. As one example, you might be
trying to determine a waste management system that minimizes cost
and minimizes the pollution level. Because these goals are pulling in
opposite directions, you might minimize cost while putting various
upper limits on the allowable pollution level. Alternatively, you could
minimize the pollution level by putting various upper limits on the
allowable cost.
Starting with version 7, Evolver can perform efficient portfolio
analysis for any model with two competing goals. You choose one of
the goals as the target to optimize, you add an “efficient frontier” type
of constraint on the other goal, where you list the bounds to try, and
you specify that you are performing an efficient frontier analysis.
Then Evolver solves a sequence of optimizations, one for each bound
you specify, and it presents the results in graphical and tabular form.

6
Chapter 2: Getting Started
Introduction
This chapter describes the contents of the Evolver package and shows
how to install Evolver and attach it to your copy of Microsoft Excel
2007 or higher.
About This Version

This version of Evolver can be installed as a 32-bit program for
Microsoft Excel 2007 or higher or as a 64-bit program for Excel 2010 or
higher.
Evolver Professional and Industrial Versions

Evolver is available in two versions: Professional and Industrial. The
difference is that the Professional version has a limit of 250 adjustable
cells, whereas the Industrial version has no such limit. You should use
Evolver Industrial if you plan to optimize large models with
hundreds or even thousands of adjustable cells.
Working with your Operating Environment

This User’s Guide assumes that you have a general knowledge of the
Windows operating system and Excel. In particular:
• You are familiar with your computer and using the mouse.
• You are familiar with terms such as icons, click, double-click,
ribbon, and window.
• You understand basic concepts such as directory (folder)
structures and file naming.

If You Need Help
Technical support is provided free of charge for all registered users of
Evolver with a current maintenance plan, or it is available on a per
incident charge. To ensure that you are a registered user of Evolver,
please register online at
http://www.palisade.com/support/register.asp
If you contact us by telephone, please have your serial number and
this manual ready. We can offer better technical support if you are in
front of your computer and ready to work.
Before Calling Before contacting technical support, please review the following
checklist:
• Have you referred to the online help?
• Have you checked this manual and reviewed the online
Quick Start tutorial from the Evolver Welcome screen?
• Have you read the README.WRI file? It contains current
information on Evolver that might not be included in the
manual.
• Can you duplicate the problem consistently? Can you
duplicate the problem on a different computer or with a
different model?
• Have you looked at our Web site? It can be found at
http://www.palisade.com. Our Web site also contains the
latest FAQ (a searchable database of tech support questions
and answers) and Evolver patches in our Technical Support
section. We recommend visiting our Web site regularly for all
the latest information on Evolver and other Palisade software.
8 Introduction
Contacting Palisade Corporation welcomes your questions, comments or
Palisade suggestions regarding Evolver. Contact our technical support staff
using any of the following methods:
• Email us at support@palisade.com.
• Telephone us at (607) 277-8000 any weekday from 9:00 AM to
5:00 PM, EST. Follow the prompt to reach technical support.
• Fax us at (607) 277-8001.
• Mail us a letter at:
Technical Support
Palisade Corporation
798 Cascadilla St.
Ithaca, NY 14850 USA
If you want to contact Palisade Europe:
• Email us at support@palisade-europe.com.
• Telephone us at +44 1895 425050 (UK).
• Fax us at +44 1895 425051 (UK).
• Mail us a letter at:
Palisade Europe
31 The Green
West Drayton
Middlesex
UB7 7PN
United Kingdom
If you want to contact Palisade Asia-Pacific:
• Email us at support@palisade.com.au.
• Telephone us at + 61 2 9252 5922 (AU).
• Fax us at + 61 2 9252 2820 (AU).
• Mail us a letter to:
Palisade Asia-Pacific Pty Limited
Suite 404, Level 4
20 Loftus Street
Sydney NSW 2000
Australia

Regardless of how you contact us, please include the product name,
version, and serial number. The exact version can be found by
selecting the Help About command on the Evolver menu in Excel.
Student Telephone support is not available with the student version of
Versions Evolver. If you need help, we recommend the following alternatives:
• Consult with your professor or teaching assistant.
• Go to http://www.palisade.com for answers to frequently
asked questions.
• Contact our technical support department via e-mail or fax.
Evolver System Requirements

System requirements for Evolver include:
• Microsoft Windows XP or higher.
• Microsoft Excel Version 2007 or higher.
10 Introduction
Installation Instructions
General Installation Instructions
The Setup program copies the Evolver system files into a folder you
specify on your hard disk. To run the Setup program in Windows XP
or higher:
Double-click the Evolver Setup.exe (or the DTSuite.exe) from your
download or installation CD and follow the Setup instructions on the
screen.
If you encounter problems while installing Evolver, verify that there
is adequate space on the drive to which you’re trying to install. After
you’ve freed up adequate space, try rerunning the installation.
Removing If you wish to remove Evolver (or the DecisionTools Suite) from your
Evolver from computer, use the Control Panel’s Add/Remove Programs utility and
Your Computer select the entry for Evolver or the DecisionTools Suite.
The DecisionTools Suite

Evolver can be used with the DecisionTools Suite, a set of add-ins for
risk and decision analysis available from Palisade Corporation. The
default installation procedure of Evolver puts Evolver in a subfolder
of a main “Program Files\Palisade” (or “Program Files
(x86)\Palisade) folder. This is similar to how Excel is often installed
into a subfolder of a “Microsoft Office” folder.
One subfolder of this Palisade folder will be the Evolver folder (by
default called Evolver7). This folder contains the Evolver add-in
program file (Evolver.xla) plus example models and other files
necessary for Evolver to run. Another Palisade subfolder is the
System folder. It contains files needed by every program in the
DecisionTools Suite, including common help files and program
libraries.

Software Activation
Activation is a one-time license verification process that is required
for your Palisade software to run as a fully licensed product. An
activation ID is on your printed/emailed invoice and is a dash
separated sequence like "DNA-6438907-651282-CDM". If you enter
your Activation ID during installation, your software is activated at
the end of the installation process and no further user action is
required. If you need to activate your software after installation,
select the License Manager command from the Evolver Help menu.
The License Manager can be used to activate, deactivate, and move
software licenses. It is also used to manage licenses for network
installations. Follow the prompts and dialogs in the License Manager
to perform the desired licensing operation.
12 Installation Instructions
Instructional Materials
The Evolver package includes a number of instructional materials to
help you learn Evolver features and how to model with Evolver.
Example Spreadsheets
Evolver includes about 25 example spreadsheets. Some of these
examples illustrate specific features of Evolver, whereas many others
illustrate potential uses of Evolver in various fields. These examples
not only help you learn how to use Evolver, but they illustrate how
extensively Evolver can be applied.
You can find these example spreadsheets from the Evolver Help
menu. When you click its Example Spreadsheets command, an
“example file list” file opens in Excel. This file contains links to all of
the example files.
Quick Start Tutorials

From the Evolver Welcome screen, which you see when you launch
Evolver or you can access at any time from the Evolver Help menu,
you can click the Quick Start link to see a series of short videos that
lead you through the basic features of Evolver. These Quick Start
videos are intended for beginners, but they are sufficient to get you
started creating your own Evolver models.
Guided Tour Videos

A series of Guided Tour videos is also available from the Evolver
Welcome Screen. These are more in-depth videos, and they lead you
through practically all of the Evolver features, from simple to more
complex.
Examples and Videos for the XDK

Once you are comfortable using Evolver, you might want to learn
how to use the macro language of Excel, Visual Basic for Applications
(VBA), to automate Evolver procedures. Evolver provides an Excel
Developer Kit (XDK) for doing this. You can learn about the XDK
from the Developer Kit (XDK) command on the Evolver Help menu.
There you will see links to example spreadsheet models of Evolver
automation, as well as additional instructional videos on using the
XDK.

14 Instructional Materials
Chapter 3: Evolver Tutorial
In this chapter, you will take a tour through an entire Evolver
optimization one step at a time. If you do not have Evolver installed
on your hard drive, please refer to the previous chapter and install
Evolver before you begin this tutorial.
You will start by opening an exisiting spreadsheet model, and then
you will define the problem to Evolver using the Evolver dialogs.
Finally, you will watch Evolver’s progress as it searches for solutions,
and you can explore some of the many options in the Evolver
Watcher. You can find additional information about any specific
topic in later chapters of this manual.
The problem-solving process begins with a model that accurately
represents a problem. For any given set of decision values, called
adjustable cell values in Evolver terminology, the model evaluates an
“objective function,” the function you want to optimize. As Evolver
searches for solutions, this objective function provides feedback,
telling Evolver how good or bad each solution is. Evolver continues
to search for better and better solutions until it can’t keep making
improvements. At this point, it declares the current best solution the
optimal solution.
15
Loading Evolver
To load Evolver, either click the Evolver icon on your Windows
desktop, or select Palisade DecisionTools and then Evolver from the
Windows Start menu. Each of these methods starts both Microsoft
Excel and Evolver, or it opens Evolve if Excel is already running.
The Evolver When Evolver is loaded, a new Evolver ribbon is visible in Excel. This
Ribbon toolbar contains icons you use to specify Evolver settings and start,
pause, and stop optimizations.
Opening an To review the features of Evolver, this section will lead you through
Example Model one of the example models that are installed when you install Evolver.
To do this:
1) Select Example Spreadsheets from the Evolver Help menu.
This opens a file that lists all Evolver example files.
2) Click the hyperlink to Bakery – Tutorial Walkthrough.xlsx to
open this example file.
The Model sheet contains a simple profit maximization problem for a
bakery business. Your bakery produces six bread products. You
track revenues, costs, and profits from production. You want to
determine the number of cases for each type of bread that maximizes
total profit while satisfying production limit guidelines. The
guidelines you face include 1) meeting the production quota for low-
calorie bread, 2) maintaining an acceptable ratio of high-fiber to low-
calorie, 3) maintaining an acceptable ratio of 5-grain to low-calorie,
and 4) keeping production time within limits for person hours used.
The Evolver Model Dialog

To set the Evolver options for this model, click the Model Definition
icon on the Evolver ribbon. This displays the Evolver Model dialog:
The Evolver Model dialog is designed so you can describe you

optimization problem in a simple, straightforward way. In this
tutorial example, you want to determine the number of cases to
produce for the different bread products to maximize overall total
profit.

Selecting the Target Cell
The Total Profit cell in this example model is the target cell. This is
the cell whose value you are trying to minimize or maximize. To
specify the target cell:
1) Set the Optimization Goal option to Maximum.
2) Enter the target cell, I11, in the Cell field.
Cell references can be entered in Evolver dialog fields in two ways.
You can type the reference directly into the field, or you can click the
Reference Entry icon and then point to the cell.
Adding Adjustable Cell Ranges

Now you must specify the adjustable cells. They contain values that
Evolver can adjust for optimizing. These cells are added and edited
one range at a time through the Adjustable Cells Ranges section of
the Model dialog.
1) Click Add to enter the first range of adjustable cells.
2) Select cells C4:G4.
3) Enter 0 as the Minimum and 100000 as the Maximum. (For
each of these, you can specify a cell reference or simply type
in a number.)
4) In the Values cell, click the down arrow and select Integer
from the dialog.
You need to restrict the possible values for an adjustable cell range to
Entering the
a specific min-max range (step 3). In Evolver this is known as a range
Min-Max Range
for Adjustable constraint. For the bakery example, the range from 0 to 100,000 is
Cells reasonable, although you experiment with the limits, particularly the
upper limit. You can also restrict the possible values to be integers if
you like (step 4).
Now, enter a second cell range to be adjusted:
1) Click Add to specify an additional adjustable cell.
2) Select cell B4, the production level of low-calorie bread.
3) Enter 20,000 as the Minimum and 100,000 as the Maximum.
4) In the Values cell, click the down arrow and select Discrete
from the dialog. Leave the default Step Size value of 2.
By selecting a step size of 2, you are require the number of cases to be
an even integer in the specified range. This is usually done to speed
up the optimization process. In this case, it is perfectly acceptable, for
all practical purposes, to limit the adjustable cell values to such step
values.
If there were additional variables in this problem, you would continue
to add ranges of adjustable cells. In Evolver, you can create an
unlimited number of ranges of adjustable cells.
Later, you might want to edit the adjustable cell settings. To do this,
you select one of the ranges and click Edit. You can also delete a range
by selecting it and clicking Delete.

Selecting a When defining adjustable cells, you can specify a solving method to be
Solving Method used. Different types of adjustable cells are handled by different
solving methods. Solving methods are set for a Group of adjustable
cells and are changed by clicking the Group button and selecting Edit
to display the Adjustable Cell Group Settings dialog. You will use
the default Recipe solving method most often, where each cell’s value
can be changed independently of the others. Because this is selected
as the default method, you don't have to change it for this example.
Constraints
Evolver allows you to enter constraints, in addition to the range
constraints on the adjustable cells. These are conditions that must be
satisfied for a solution to be valid. In this example model, there are
three additional constraints that must be satisfied for a possible set of
production levels to be valid. They are:
1) An upper limit of 50,000 on the total production hours.
2) An acceptable ratio of high-fiber to low-calorie bread (high-
fiber cases produced >= 1.5 * low-calorie cases produced).
3) An acceptable ratio of 5-grain to low-calorie bread (5-grain
cases produced >= 1.5 * low-calorie cases produced).
Each time Evolver generates a trial solution to your model, it checks
whether the constraints are satisfied. If they aren’t, the solution is
invalid and is discarded.
Constraints are displayed in the bottom Constraints section of the
Evolver Model dialog. Two types of constraints can be specified in
Evolver:
• Hard. These are conditions that must be satisfied for a
solution to be valid.
• Soft. These are conditions you would like to be satisfied as
much as possible, but which you might be willing to
compromise for a big improvement in the target cell value.
To add a constraint, click the Add button in the Constraints section of
Adding a
Constraint the main Evolver dialog. This displays the Constraint Settings dialog,
where you enter the constraints for your model.
To enter the constraints for the bakery model, you need to specify
three hard constraints. First, enter the “simple” format constraint on
total production hours:
1) In the Range to Constrain box, enter cell I8.
2) Select the <= operator to the right of Range to Constrain.
3) Enter 50,000 in the Constraining Value(s) box.
4) Click OK to enter this constraint.
5) Optionally, in the Model Definition dialog, enter text such as
“Total Hours” in the Description column of the Constraints
grid.

Now, enter the “formula” format hard constraints:
1) Click Add to display the Constraint Settings dialog again.
2) Click the Entry Style button, and select Formula.
4) In the Constraint Formula box, enter C4>= 1.5*B4.
5) Click OK, and then add descriptive text for this constraint.
6) Proceed similarly to enter the D4>= 1.5*B4 constraint.
The Model dialog with the completed constraints section should look
as follows.
Other Evolver Options
Evolver provides a number of options to control how what happens
during an optimization. The default options often work well, but you
can change any of the options to suit your needs.
Stopping It is difficult to know how long to let Evolver run, especially since
Conditions different types of models require longer runtimes than others.
Optimization algorithms, including those in Evolver, tend to make
quick progress toward an optimal solution and then “spin their
wheels” as they get closer to the optimal solution. In any case, the
stopping conditions allow Evolver to automatically stop when either:
1) a certain number of scenarios or “trials” have been examined, b) a
certain amount of time has elapsed, c) no significant improvement has
been found in the last n trials, or d) an entered Excel formula
evaluates to TRUE. To view and edit the Runtime options:
1) Click the Settings icon on the Evolver ribbon.
2) Select the Runtime tab. For this example model, you can
accept the default setting shown here.
In the Optimization Settings dialog you can select any combination of

these optimization stopping conditions, or none at all. If you select
more than one stopping condition, Evolver will stop as soon as any
one of the selected conditions are met. If you do not select any
stopping conditions, Evolver might run until you stop it manually by
pressing a Stop button, or it might stop when all possible solutions
have been tried.

Here are descriptions of the possible options:
• Trials. This option sets the number of trials for Evolver to
run. In each trial, Evolver evaluates one complete set of
variables or one possible solution to the model.
• Time. Evolver will stop after the specified amount of time
has elapsed.
• Progress. This stopping condition is the most popular
because it keeps track of the improvement and allows Evolver
to run until the rate of improvement is small. For example,
Evolver could stop if 20000 trials have passed and the target
cell value has not improved by more than 0.01%.
• Formula is True. Evolver will stop if the entered Excel
formula evaluates to TRUE in a model recalculation.
While Evolver runs, a set of options are available on the View Tab to
determine what you will see on the screen.
The two options for Show Excel Recalculations are:

• Every New Best Trial. This option redraws the screen each
time Evolver generates a new best answer (that is, better than
any other solution so far), allowing you to see the current
optimal solution at any time during the optimization.
• Every Trial. This option redraws the screen after each trial,
allowing you to see Evolver adjusting the variables and
calculating the target. This is a useful option while you are
learning Evolver, and also each time you use Evolver on a
new model, to verify that your model is calculating correctly.
Running the Optimization
Now, all that remains is to optimize this model. To do this:
1) Click OK to exit the Optimization Settings dialog.
2) Click the Start Optimization icon.
During the run, the Progress window displays: 1) the best solution
found so far, 2) the original value for the target cell when the Evolver
optimization began, 3) the number of trials that have been executed
and number of those trials which were valid (all constraints were
satisfied), and 4) the time that has elapsed in the optimization.
At any time during the run, you can click the Excel Updating Options
icon (the one on the left) to select or deselect live updating of the
screen each trial.

Evolver can also display a running log of each trial solution. This is
The Evolver
Watcher displayed in the Evolver Watcher while Evolver is running. To view
a running log of the trials:
1) Click the Watcher icon (magnifying glass) in the Progress
window to display the Evolver Watcher.
2) Click the Log tab.
In this report the results of each trial solution is shown. The column
for Result shows by trial the value of the target cell that you are trying
to maximize or minimize, in this case, the Total Profit in I11. The
columns for B4 through G4 identify the values in the adjustable cells.
If you sense that the target is not improving, you can click the Stop
Stopping the
Optimization icon to stop the optimization. When the Evolver process stops,
Evolver displays the Stopping Options tab, which offers the following
choices:
These same options will automatically appear when any of the

stopping conditions that were set in the Evolver Optimization
Settings dialog are met.

Evolver can create an optimization summary report that contains
Summary
Report information such as date and time of the run, the optimization
settings used, the value calculated for the target cell and the original
and best value for each of the adjustable cells.
This report is useful for comparing the results of successive

optimizations.
To place the new, optimized mix of production levels for the bakery
Placing the
Results in Your to each of the six types of bread in your worksheet:
Model 1) Make sure the Restore Original Adjustable Cell Values
option is not selected when Stopping Options are displayed.
2) Click the OK button.
You will be returned to the Model sheet, with all of the new variable
values that created the best solution.
You can see that Evolver found a solution with total profit 4,686,792.
However, your result might be slightly higher or lower than this.
This is due to random choices built into Evolver’s algorithms and the
possibility that you might use different stopping conditions. There is
often no guarantee that Evolver has found the optimal solution, but its
solution will usually be extremely close to optimal, certainly close
enough for all practical purposes.
When you save a workbook after running Evolver on it (even if you
“restore” the original values in the adjustable cells after running
Evolver), all of the Evolver settings in the Evolver dialogs will be
saved along with the workbook. The next time that workbook is
opened, all of the most recent Evolver settings load automatically.

Chapter 4: Evolver Command
Reference
Evolver Icons
The following icons are shown on the Evolver ribbon in Excel.
Icon Function Performed
Defines the optimization model: the target cell, the
adjustable cells, and the constraints
Lets you view or make changes to the settings for

the current optimization model
Starts the optimization process
Provides several types of reports of optimization

results
Provides Application Settings and several other

useful utilities
Provides Evolver Help files
31
32
Evolver Commands
Model Definition Command
Defines the goal, adjustable cells, and constraints for a model
The Model Definition command displays the Evolver Model dialog.
This dialog is used to specify an optimization problem to Evolver.
This dialog starts empty with each new Excel workbook, but it saves
its information with each workbook. Therefore, when the workbook
is opened again, the Model dialog will remember its previous settings.

Options in the Model dialog include:
• Optimization Goal. This specifies the kind of optimization
Evolver will run. If Minimum is selected, Evolver search for
adjustable cell values that minimize the target. Similarly, if
Maximum is selected, Evolver will try to maximize the target.
• In contrast, if Target Value is selected (a less likely choice, but
still possible), Evolver will search for adjustable cell values
that produce a value for the target cell as close as possible to
the value you specify. It will stop when the target value is
sufficiently close to the specified value.
• Cell. The target cell to be minimized, maximized, or set to a
specified value. A value for this target cell will be generated
for each trial solution that Evolver generates. The target cell
should contain a formula that depends, either directly or
through a series of calculations, on the adjustable cells. This
formula can include standard Excel functions or user-defined
VBA macro functions.
• Analysis Type. You should choose the Standard type unless
you are calculating an efficient frontier. Then you should
choose the Efficient Frontier type.
Adjustable Cell The Adjustable Cell Ranges table displays each range of adjustable
Ranges cells in a row. One or more adjustable cell ranges can be included in
an Adjustable Cell Group. All cell ranges in an Adjustable Cell
Group share a common solving method; if the Genetic Algorithm is
used, they also share the crossover rate, mutation rate, and operators.
(All of this is explained in detail below.)
34 Evolver Commands
Because the adjustable cells contain the decision variables of the
problem, you must define at least one group of adjustable cells to use
Evolver. Most problems will be described with only one group of
adjustable cells, but more complex problems can require different
blocks of variables to be solved with different solving methods
simultaneously. This unique architecture allows for highly complex
problems to be built up from multiple groups of adjustable cells.
The following options are available for entering Adjustable Cell
Ranges:
• Check boxes. Each adjustable cell range has a check box to its
left. You can activate or deactivate any (or all) of the current
adjustable cell ranges by checking or unchecking its box.
• Add. You can add new adjustable cells by clicking the Add
button. You are immediately asked for the adjustable cell
range, and then a new row is added to the table with the
following fields.
• Range. The range you already specified is in the Range field.
You can change this if necessary by typing or pointing to a
new range.
• Minimum and Maximum. The Minimum and Maximum
entries set the range of acceptable values for each adjustable
cell in the range. These values are required. Evolver will not
allow any values outside the specified ranges (unless you
specify soft constraints). In cases where natural limits don’t
exist (or aren’t known), you should specify limits wide
enough to allow all reasonable solutions, but the tighter you
can keep these limits, the quicker the solution process will be,
because tighter limits restrict the scope of the search.

• Values. The Values entry (which is relevant only for the Recipe
and Budget solving methods) specifies the type of adjustable cell
values Evolver will search through. In each of these settings, only
values between the specified minimum and the specified
maximum are included in the search.
There are three possible settings:

- Any. All real numbers are included. This is the default.
- Integer. Only integer values are included. This is usually
chosen when non-integer values make no sense in your
model.
- Discrete. With this setting, you choose a step size. Then
only multiples of this step size are included. This setting
is useful when “rounded” values of the adjustable cells,
such as multiples of $10, are fine for all practical
purposes.
The default is to include all real numbers in the specified
range. The use of Discrete (or Integer) can dramatically the
number of possible solutions and will often result in much
faster optimizations.
Adjustable Cell Evolver uses groups of adjustable cell ranges to provide more
Groups flexibility in the solution process. Each group can contain multiple
cell ranges. This allows you to build a hierarchy of groups of related
cell ranges. Within each group, each cell range can have its own
minimum-to-maximum range.
All cell ranges in an Adjustable Cell Group share a common solving
method. This is specified in the Adjustable Cell Group Settings
dialog. This dialog is accessed by clicking the Group button. You can
create a new group to which you can add adjustable cell ranges, or
you can edit the settings for an existing group.
36 Evolver Commands
By default, each adjustable cell range you add (by clicking the Add
button) is added to a single group using the Recipe solving method.
However, you have other options.
• To add a new group that uses another solving method, click
the Group button and select New. You will be asked for an
adjustable cell range, and then you will see the Adjustable
Cell Group Settings dialog, where you can choose the solving
method for the new group.
• Once you have two more groups, the adjustable cell table will
look something as follows. You can select either group bar
(here, the Budget group is selected). Then when you click
Add to add another adjustable cell range, it will be added to
the selected group. Alternatively, you can click the Group
button and select Edit to edit the selected group (change its
solving method or description).
Again, remember that each group of adjustable cell ranges shares a

common solving method. Evolver can use any of six solving
methods, depending on the type of optimization model. Three of the
solving methods (Recipe, Order, and Grouping) use entirely different
algorithms. The other three are descendants of these three, adding
additional constraints.
The following section describes each solving method. To get a better
understanding of how each solving method is used, you are also
encouraged to explore the example files included with the software.

Recipe Method The Recipe solving method is the simplest and most frequently used
solving method. You use Recipe whenever the values in the
adjustable cells can be varied independently of one another. You can
think of each variable as the amount of an ingredient in a cake. When
you use the Recipe solving method, you are telling Evolver to
generate values for those variables to find the best mix. The only
constraint you place on a set of Recipe variables is its min-max range.
The Order solving method is the second most popular type, after
Order Method Recipe. An order is a permutation of a list of items, where you are
trying to find the best way to arrange the items. Unlike the Recipe
(and Budget) solving methods, which ask Evolver to generate values
for the chosen variables, this solving method asks Evolver to use the
existing values in your model. An order could represent the order in
which to perform a set of tasks. For example, you might wish to find
the order in which to accomplish five tasks, numbered 1, 2, 3, 4, and 5.
The Order solving method would scramble those values, so one
scenario might be 3, 5, 2, 4, 1. Because Evolver is just trying variable
values from your initial sheet, there is no min-max range entered for
adjustable cells when the Order solving method is used.
The Grouping solving method should be used whenever your
problem involves multiple items to be grouped together in sets. By
default, the number of different groups that Evolver creates will be
Grouping equal to the number of unique values present in the adjustable cells at
Method the start of an optimization. (This default behavior can be altered by
using the Group IDs field, discussed below).
For example, suppose a range of 50 cells contains only the values 1, 2,
3, and 4. (These values don’t have to be consecutive integers, but they
usually are.) When you select the 50 cells and adjust the values using
the Grouping solving method, Evolver will assign each of the 50 cells
to one of the four groups, 1, 2, 3, or 4. All of the groups are
represented by at least one of the adjustable cells. It is like tossing
each of the 50 cells in one of four “bins,” and making sure there is at
least one cell in each bin. Like the Order solving method, Evolver is
arranging existing values, so there is no min-max range to define.
NOTE: When using the “Grouping” solving method, you should not
leave any cells blank, unless you want 0 to be considered one of the
groups.
38 Evolver Commands
When using the Grouping solving method, there are two additional
settings in the Adjustable Cell Group Settings dialog:
• Group IDs (Optional). This setting allows you to specify a
range containing numeric group IDs. Normally Evolver
reads group IDs from the adjustable range. For example, if
the adjustable range is A1:D1, and it contains numbers 1, 1, 3,
2, then Evolver will use 1, 2, and 3 as group IDs. However,
there may be more groups than there are adjustable cells; for
example, you might want to assign items represented by cells
A1:D1 to groups numbered 1 to 5. In this case, the Group IDs
setting allows you to specify a range containing five cells with
numbers 1 to 5 be used as group IDs during optimization.
• All Groups Must Be Used. If this option is checked, every
solution will have members from every group. For example,
if the adjustable cells are A1:D1, and group IDs are 1, 2, and 3,
then Evolver will not try a solution with 1 assigned to all four
cells (with 2 and 3 missing). On the other hand, this solution
can be tried if this check box is not checked.
Budget Method The Budget solving method is similar to Recipe except that all of the
variables’ values must sum to a fixed value. That value is the sum of
the variables’ values at the time an optimization is started.
For example, you might want to find the best way to distribute an
annual budget among a number of departments. The Budget solving
method finds the sum of the initial values for the departments, and it
allows only solutions with that sum.
Instead of using the Budget solving method, you could use the Recipe
solving method and add a constraint on the sum of the values in the
adjustable cells. This approach works well if the OptQuest engine is
used. However, if you are using the Genetic Algorithm, the Budget
method will be more efficient.

Project Method The Project solving method is similar to the Order solving method
except that certain items (usually tasks) must precede others. The
Project solving method is especially useful in project management,
where an ordering of the tasks is require but there are precedence
relationships. When using the Project solving method, it is best to
have the task indexes (1 through the number of tasks) in a single
column. These become the adjustable cells
After you have specified the location of the adjustable cells, you
should specify the location of the preceding tasks cells in the
Preceding Tasks section of the dialog. This is a table of cells that
describes which tasks must be preceded by which other tasks. This is
illustrated in the example below (with rows 26-47 not shown), where
the goal is to find the best ordering of town to visit. For example,
town 5 has precedents 32 and 8, so town 5 can’t be visited until towns
32 and 8 have both been visitied.
The precedence tasks range should be specified as being n rows by m

columns, where n is the total number of tasks (adjustable cells), and m
is the largest number of precedents that any one task has.
40 Evolver Commands
Schedule The Schedule solving method is used to schedule tasks to times. Each
Method task is assumed to take the same amount of time, much as classes at a
school are all of the same length. In this case, the Adjustable Cell
Group Settings dialog lets you directly specify the number of time
blocks to be used.
The example below illustrates this. Each yellow cell specifies a time
slot from 1 to 6 for the class. These are the adjustable cells. The
Constraint Cells consist of the bordered range. This range uses
“codes” for the spelled-out constraints to their left.

Eight kinds of constraints are recognized:
1) (with) The tasks in columns 1 and 3 must occur in the same
time block.
2) (not with) The tasks in columns 1 and 3 must not occur in the
same time block.
3) (before) The task in column 1 must occur before the task in
column 3.
4) (at) The task in column 1 must occur in the time block in
column 3.
5) (not after) The task in column 1 must occur at the same time
or before the task in column 3.
6) (not before) The task in column 1 must occur at the same time
or after the task in column 3.
7) (not at) The task in column 1 must not occur in the time block
in column 3.
8) (after) The task in column 1 must occur after the task in
column 3.
You can enter either a numeric code (1 through 8) or the description
(after, not at, etc.) for a constraint. (All language versions of Evolver
will recognize the English description entered for a constraint as well
as the translated form).
Note: When you select the schedule solving method, integers starting
from 1 are always used regardless of the original values in the
adjustable cells.
42 Evolver Commands
Constraints Evolver allows you to enter constraints, or conditions that must be
satisfied for a solution to be valid. Constraints you have entered are
shown in the Constraints table in the Model Definition dialog. As
with adjustable cell ranges, there is a check box to the left of each
defined constraint. This allows you to activate or deactivate any
constraint in the list.
Adding You click the Add button to add a new constraint. This displays the
Constraints Constraint Settings dialog, where you specify the constraint.
The main entries for a constraint are:

• Range to Constrain. The cell(s) in Excel for which values will
be constrained during the optimization.
• Constraining Values. The values you want to limit the cell to.
The operator entry in the middle controls whether the Range to
Constrain will be >, >=, <, <= or = the Constraining Values.
Constraint Entry Three styles – One-Sided, Two-Sided, and Formula – can be used for
Style entering constraints. The Entry Style button in the Constraints dialog
lets you select the style you want to use.
The One-Sided and Two-Sided styles allow constraints to be entered
using <, <=, >, >= or = relations. A typical Two-Sided constraint
would be 0<=Value of A1<=10, where A1 is entered in the Range to
Constrain, 0 is entered as one constraining value, and 10 is entered as
the second.

A Formula constraint, on the other hand, allows you to enter any
valid Excel formula as a constraint. For example, you could enter the
formula =IF(A1>100, B1>0, B1<0). This constraint requires the value
in cell B1 to be positive or negative depending on the value in cell A1.
Alternatively, the formula can be entered in a cell; if that cell is C1, for
example, then you can enter =C1 in the Formula field of the
Constraint Settings dialog.
In general, entering constraints in the One-Sided or Two-Sided style
helps Evolver find the optimal solution faster. The formula just
discussed could be entered in cell D1 as =IF(A1>100, B1, -B1). Then a
simple One-Sided constraint can be added, requiring that D1>0.
Additional Clicking the More button expands the Constraint Settings dialog to
Options for provide additional options for setting up constraints.
Entering
Constraints
44 Evolver Commands
Other constraint options include:
Constraint Type Two types of constraints can be specified in Evolver:
• Hard, or conditions that must be satisfied for a solution to be
valid. Trial solutions that do not satisfy hard constraints are
discarded.
• Soft, or conditions that you would like to satisy, but can be
violated at a certain “penalty” cost. This penalty cost is
subtracted from or added to the target value to discourage
invalid solutions.
Precision Precision refers to violations of constraints that are so small that
Evolver disregards them and will treat a solution as valid despite
these small violations. This small imprecision in the handling of
constraints relates to the fact that computers can only handle
mathematical operations with finite precision.
Leaving Precision as Automatic is recommended. In this case the
Optimization Summary will report the actual precision number,
which is the maximum violation of the constraint that is disregarded.
Alternatively, a specific number can be entered in the Precision field.
Adding To perform an Efficient Frontier Analysis, you need to specify a
Constraint for special type of constraint with multiple alternative constraining
Efficient values. To do this, you first must select Efficient Frontier as the
Frontier Analysis Type in the Model dialog.
Analysis
Then you select Use for Efficient Frontier in the Constraint Settings
dialog to display the options for defining the list of constraining
values.
The List of Constraining Values option specifies how you will enter
the list of values. For example, suppose you want the following list of
alternative constraining values: 45000, 47000, 49000, 51000, 53000,
55000. You can define the list by specifying that it contains 6 values

between 45000 and 55000 (the Between Minimum and Maximum
option). Alternatively, you can type the 6 values in the dialog (Table
option), or type them in an Excel range (In Excel Range option).
To identify the efficient frontier, Evolver performs multiple
optimizations to find the optimal value for each possible value of the
constraint. In the example here, Evolver would perform an
optimization at each of the constraining values and then graph the
optimal solution across the possible constraining values.
More on Soft Soft Constraints are conditions you would like to satisfy as much as
Constraints possible, but you are willing to violate as a compromise (to improve
the target value). When a soft constraint is violated, a penalty is
attached to the target. The amount of penaly is determined by a
penalty function that you specify when you specify the soft constraint.
More information about penalty functions is as follows:

• Entering a Penalty Function. Evolver has a default penalty
function that is displayed when you first enter a soft
constraint. However, you can replace this with any valid
Excel formula. Any penalty function you enter should
include the keyword deviation. This “placeholder” represents
the absolute amount by which the constraint value is beyond
its limit. For each trial solution, Evolver checks whether the
soft constraint has been satisfied; if not, it places the amount
of deviation in the penalty formula and then calculates the
amount of penalty to apply to the target cell value.
The penalty amount is either added to (for minimization) or
subtracted from (for maximization) the target cell value to
penalize it.
46 Evolver Commands
• Viewing the Effects of an Entered Penalty Function.
Evolver includes an Excel worksheet Penalty Functions and
Soft Constraints.xlsx that can be used to evaluate the effects of
different penalty functions on specific soft constraints and
target cell results.
• Viewing the Penalties Applied. When a penalty is applied

to the target cell due to an unsatisfied soft constraint, you can
see the amount of penalty applied in the Evolver Watcher. In
addition, penalty values are shown in Optimization Log
worksheets, created optionally after optimization.
NOTE: If you place a solution in your worksheet at the end of an
optimization, the calculated target cell result shown in the
spreadsheet will not include any penalties applied due to unsatisfied
soft constraints. You can check the Optimization Log worksheet to
see the penalized target cell value and the amount of penalty imposed
due to each unsatisfied soft constraint.
• Implementing Soft Constraints in Worksheet Formulas.
Penalty functions can be implemented directly in the
formulas in your spreadsheet model so that they become part
of the formula in the target cell. If you implement soft
constraints in this manner, you should not be enter them in
the main Evolver dialog.

Settings Command
Defines a variety of settings for the current optimization
The Settings command opens the Optimization Settings dialog. This
dialog contains five tabs for viewing or changing a variety of settings
for the current optimization.
Optimization The Runtime tab displays Evolver settings that determine the runtime
Settings Dialog of the optimization. These stopping conditions specify how and
– Runtime Tab when Evolver will stop during an optimization. Once you select the
Start Optimization command, Evolver will continue to run, searching
for better solutions and running simulations until the selected
stopping criteria are met. You can turn on any number of these
conditions, or none at all; if you do not select any stopping conditions,
Evolver will run until all possible solutions have been tried, or until
you stop it manually by pressing the Stop button. When multiple
conditions are checked, Evolver will stop as soon as one of the chosen
conditions is met. You can also override these selections and stop
Evolver at any time manually by using the Stop button in the Evolver
Watcher or Progress windows.
48 Evolver Commands
Optimization Optimization Runtime options on the Runtime tab include:
Runtime • Trials. This option, when set, stops Evolver when the given
Options
number of trial solutions have been generated.
The Trials setting is particularly useful when comparing
Evolver’s efficiency when trying different modeling methods.
By changing the way you model a problem, or by choosing a
different solving method, you can increase Evolver’s
efficiency. Having a model run a specified number of trials
will indicate how efficiently Evolver is converging to a
solution, regardless of any differences in the number of
variables chosen, the speed of the computer hardware being
used, or the screen re-drawing time. The Evolver
optimization summary worksheet is also useful in comparing
results across runs. For more information on Optimization
Summary worksheets, see the Evolver Watcher – Stopping
Options section in this chapter.
• Time. This option, when set, stops Evolver from generating
scenarios after the given number of hours, minutes, or
seconds has elapsed. This entry can be any positive number
(600, 5.2, etc.).
• Progress. This option, when set, stops Evolver from
generating scenarios when the improvement in the target cell
value is less than the specified amount (change criterion).
You can specify, as an integer, the number of trials over
which to check the improvement. A percentage value such as
0.01% can be entered as the maximum change value in the
Maximum Change field.
For example, suppose you are trying to maximize the value of
the target cell, and the parameters used in the Progress
stopping condition are 20000 trials, and 0.01% change. Also
suppose that after 10000 trials, the best target cell value found
so far is 2000. If the Progress option is the only stopping
condition selected, when Evolver gets to trial #30000, it will
continue only if it is able to find a target cell value of at least
2000.2 during those last 20000 trials. In other words, if the
target cell values have not improved at least 0.01% over the
last 20000 trials, Evolver assumes there is little more
improvement to be found, and it stops the search. The more
complex the optimization problem, the larger the number of
trials you should specify in this stopping condition.

This is the most popular stopping condition, because it
provides an effective way to stop Evolver after the
improvement rate is slowing down and better solutions are
not being found. If you are viewing the graphs of the best
results on the Progress tab of the Evolver Watcher, you will
see the graphs flatten out for a while before this condition is
met and Evolver stops. The Progress option is really just an
automatic way to do what you could do yourself with the
graph: let it run until the improvement levels off.
• Formula is True. This stopping condition causes the
optimization to stop whenever the entered (or referenced)
Excel formula evaluates to TRUE during the optimization.
• Stop on Error. This stopping condition causes the
optimization to stop whenever an Error value is calculated for
the target cell.
NOTE: You can also select no stopping conditions, and Evolver will
run until you press the Stop button on the Evolver Progress or
Watcher window, or until it determines that all the possible
solutions have been tried.
50 Evolver Commands
The Efficient Frontier Runtime tab on the Optimization Settings
Optimization
Settings Dialog dialog displays Evolver settings that determine the runtime of an
– Efficient Efficient Frontier analysis. If you do not select any stopping
Frontier conditions, Evolver will run until all the possible solutions have been
Runtime Tab tried, or until you stop it manually by pressing the Stop button.
When multiple conditions are checked, Evolver will stop as soon as
one of the chosen conditions is met. You can also override these
selections and stop Evolver manually at any time by clicking the Stop
button in the Evolver Watcher or Progress windows.
The options on the Efficient Frontier Runtime tab include:

• Trials per Constraining Value. If this field is set to say 10000,
say, and you defined the Efficient Frontier analysis by
specifying 6 alternative values in a constraint, the Efficient
Frontier analysis will stop after 60000 trial solutions (6 x 10000
simulations).
• Time. This option, when set, stops Evolver from analyzing
scenarios after the given number of hours, minutes, or
seconds has elapsed. This entry can be any positive number
(600, 5.2, etc.).

• Progress. This option, when set, stops Evolver from analyzing
scenarios when the improvement in the target cell value is
less than the specified amount (change criterion). You can
specify, as an integer, the number of trials over which to
check the improvement. A percentage value such as 0.01%
can be entered as the maximum change value in the
Maximum Change field. The improvement in the target cell
value needs to be understood in the context of the fact that in
the Efficient Frontier analysis, you are running multiple
parallel optimizations. The analysis will stop when the
progress is insufficient in each of the parallel optimizations.
Optimization The Engine tab on the Optimization Settings dialog selects the
Settings Dialog optimization engine and settings that will be used during an
– Engine Tab optimization. Evolver uses two optimization engines, OptQuest and
Genetic Algorithm, to search for optimum solutions to a problem.
The Initial Seed field refers to the random number generator of the
optimization algorithm (Genetic Algorithm or OptQuest). By setting
this field to a fixed integer value, you can ensure that the same
sequence of solutions will be generated in two optimizations, given
the same initial values in the adjustable cells. By leaving this field set
to Automatic, you let Evolver select the initial seed randomly.
52 Evolver Commands
Optimization In most cases, Evolver can automatically detect which engine will
Mode give the best and fastest solutions to your model. The Automatic
setting specifies this. However, there might be times when you wish
to specify an engine to use. You might, for example, have a model
that requires the Project or Schedule solving methods that are
available only with the Genetic Algorithm engine.
Optimize Using Two optimization engines, Genetic Algorithm and OptQuest, are
available to search for optimum solutions to a problem.
The Genetic Algorithm engine mimics Darwinian principles of
natural selection by creating an environment where hundreds of
possible solutions to a problem can compete with one another, and
only the fittest survive. Just as in biological evolution, each solution
can pass along its good “genes” through “offspring” solutions so that
the entire population of solutions will continue to evolve better
solutions.
The OptQuest engine uses metaheuristic, mathematical optimization,
and neural network components to guide the search for best solutions
to decision and planning problems of all types. OptQuest’s methods
Search, Neural Networks, Scatter Search, and Linear/Integer
Programming, into a single composite method.
Genetic Genetic Algorithm Settings on the Engine tab include:
Algorithm
Settings Population Size. The population size tells Evolver how many
organisms (or complete sets of variables) to store in memory at any
given time. Although there is still much debate and research
regarding the optimal population size to use on different problems,
we generally recommend using 30-100 organisms in your population,
depending on the size of your problem (bigger population for larger
problems). The common view is that a larger population takes longer
to settle on a solution but is more likely to find a global solution
because of its more diverse gene pool.
Crossover and Mutation. One of the most difficult problems with
searching for optimal solutions, when your problem has seemingly
endless possibilities, is in determining where to focus your energy. In
other words, how much computational time should be devoted to
looking in new areas of the “solution space” and how much time
should be devoted to fine-tuning the solutions in the population that
have already proven to be fairly good?
A big part of the genetic algorithm success has been attributed to its
ability to achieve this balance. The structure of the GA allows good
solutions to “breed,” but it also keeps “less fit” organisms in memory
to maintain diversity, reasoning that a latent “gene” will prove
important to the ultimate solution.
Crossover and Mutation are two parameters that affect the scope of
the search, and Evolver allows you to change these parameters before,
and also during, the evolutionary process. This way, you can help the
GA by deciding where it should focus its energy. For most purposes,
you should not need to adjust the default crossover and mutation
settings (0.5 and 0.1 respectively). However, if you want to fine-tune
the algorithm to your problem, conduct comparative studies, or just
to experiment, you can try adjusting these two parameters.
• Crossover. The crossover rate can be set to between 0.01 and
1.0, and reflects the likelihood that future scenarios or
“organisms” will contain a mix of information from the
previous generation of parent organisms. In other words, a
rate of 0.5 means that an offspring organism will contain
roughly 50% of its variable values from one parent and the
remaining values from the other parent. A rate of 0.9 means
that roughly 90% of an offspring organism’s values will come
from the first parent and 10% will come from the second
parent. A Crossover rate of 1 means that no crossover will
occur, so only clones of the parents will be evaluated.
• The default rate used by Evolver is 0.5. Once Evolver has
started solving a problem, you can change the crossover rate
by using the Evolver Watcher (see the Evolver Watcher
section in this chapter).
• Mutation Rate. The mutation rate can be set to between 0.0
and 1.0, and reflects the likelihood that future scenarios will
contain some random values. A higher mutation rate means
that more mutations, or random “gene” values, will be
introduced into the population. Because mutation occurs
after crossover, setting the mutation rate to 1 (100% random
values) will effectively prevent the crossover from having any
effect, and Evolver will generate totally random scenarios.
If all the data of the optimal solution is somewhere in the
population, then the crossover operator alone is enough to
eventually piece together the solution. Mutation has proven to be
a powerful force in the biological world for many of the same
reasons that it is needed in a genetic algorithm; it is vital to
maintaining a diverse population of individual organisms,
thereby preventing the population from becoming too rigid, and
unable to adapt to a dynamic environment. As in a genetic
algorithm, it is often the genetic mutations in animals that
eventually lead to the development of critical new functions.
54 Evolver Commands
For most purposes, the default mutation setting does not need
adjustment. However, you can change it to fine-tune Evolver’s
performance on complex problems. You might want to boost the
mutation rate if Evolver’s population is fairly homogenous and no
new solutions have been found in the last several hundred trials.
Typical setting changes are from 0.06 to 0.2. Once Evolver has started
solving a problem, you can change the mutation rate dynamically by
using the Evolver Watcher (see the Evolver Watcher section later in
this chapter).
By selecting Auto from for the Mutation rate, auto-mutation rate
adjustment is selected. Auto-mutation rate adjustment allows
Evolver to increase the mutation rate automatically when an organism
"ages" significantly, that is, when it has not changed over an extended
number of trials. For many models, especially where the “best”
mutation rate is not known, the Auto setting is preferred.
Operators Evolver includes selectable genetic operators when used with the
Recipe solving method. Clicking the Operators button in Engine tab
allows you to select a specific genetic operator (such as heuristic
crossover or boundary mutation) to be used when generating possible
values for the adjustable cells. In addition, you can have Evolver
automatically test all available operators and identify the one that
performs best for your problem.
Genetic algorithms use genetic operators to create new members of

the population from current members. Two of the genetic operators
Evolver employs are mutation and crossover. The mutation operator
determines whether random changes in “genes” (variables) will occur
and how they occur. The crossover operator determines how pairs of
members in a population swap genes to produce “offspring” that may
be better solutions than either of their “parents.”
Evolver includes the following specialized genetic operators:
• Linear Operators. These are designed to solve problems
where the optimal solution lies on the boundary of the search
space defined by the constraints. This mutation and
crossover operator pair is well suited for solving linear
optimization problems.
• Boundary Mutation. This is designed to quickly optimize
variables that affect the result in a monotonic fashion and can
be set to the extremes of their range without violating
constraints.
• Cauchy Mutation. This is designed to produce small changes
in variables most of the time, but can occasionally generate
large changes.
• Non-uniform Mutation. This produces smaller and smaller
mutations as more trials are calculated. This allows Evolver
to fine-tune solutions.
• Arithmetic Crossover. This creates new offspring by
arithmetically combining the two parents (as opposed to
swapping genes).
• Heuristic Crossover. This uses values produced by the
parents to determine how the offspring is produced. It
searches in the most promising direction and provides local
fine-tuning.
Depending on the type of optimization problem, different
combinations of mutation and crossover operators can produce better
results than others. In the Operators tab of the Adjustable Cell Group
Settings dialog, when using the Recipe solving method, you can select
any number of operators. When multiple selections are made,
Evolver will test valid combinations of the selected operators to
identify the best performing ones for your model. After a run, the
Optimization Summary worksheet ranks each of the selected
operators by their performance during the run. For subsequent runs
of the same model, selecting just the top performing operators can
lead to faster, better performing optimizations.
56 Evolver Commands
The View tab in the Optimization Settings dialog displays Evolver
Optimization
Settings Dialog settings that determine what will be shown during an optimization.
– View Tab Options on the View tab include:
• Minimize Excel at Start. This option selects to minimize
Excel when an optimization starts.
• Show Excel Recalculations. This specifies to update Excel
either with Every New Best Trial, or at the end of Every
Trial.
Optimization VBA macros can be run at different times during an optimization and
Settings Dialog during each trial solution. This allows you to develop custom
– Macros Tab calculations that will be invoked during an optimization.
Macros can be executed at the following times during an

optimization:
• At the Start of the Optimization. The macro runs after the
Run icon is clicked, prior to the first trial solution being
generated.
• Before Recalculation of Each Trial. The macro runs before
recalculation of each trial that is executed.
• After Recalculation of Each Trial. The macro runs after
recalculation of each trial that is executed.

• After Storing Output. The macro runs after each trial that is
executed and after the target cell’s value is stored in the
optimization engine. Consequently, any changes made to
cell values will have no effect on the information passed to
the optimization engine on that trial. This macro can be used
to update you on optimization progress (or to log
optimization progress).
• At the End of the Optimization. The macro runs when the
optimization is completed.
This feature allows calculations that can be performed only through
the use of a macro to be made during an optimization. Examples of
such macro-performed calculations are iterative looping calculations
and calculations that require new data from external sources.
The Macro Name defines the macro to be run.
58 Evolver Commands
Start Optimization Command
Starts an optimization
The Start command starts an optimization process for the active
model. As soon as Evolver is running, you will see the following
Evolver Progress window.
The Progress window displays:

• Trial. The total number of trials that have been executed, and
can also indicate how many of these trials were valid (all the
hard constraints were satisfied).
• Runtime. The elapsed time in the run.
• Original. The original value in the target cell.
• Best. The best target cell value found so far.
The toolbar at the bottom of the Progress window includes the
following options (from left to right):
• Display Excel Updating Options. This turns Excel Updating
Options on or off. With this option on, the screen will update
every iteration.
• Display Evolver Watcher. This displays the Evolver Watcher
window.
• Run. This resumes the Evolver process after it has been
paused.
• Pause. This allows you to pause the Evolver process. While
it is paused, you can open and explore the Evolver Watcher
and change parameters, look at the whole population, view a
status report, or copy a graph.
• Stop. This stops the optimization process.

Evolver Watcher
The magnifying glass icon on the Evolver Progress window displays
the Evolver Watcher. Evolver Watcher is responsible for regulating
and reporting on all Evolver activity.
From Evolver Watcher, you can change parameters and analyze the
progress of the optimization. You can also see real-time information
about the problem and information on Evolver’s progress in the status
bar across the bottom of Evolver Watcher.
Evolver Watcher The Progress Tab graphically shows how results are changing, by
– Progress Tab trial, for the target cell. Progress graphs show the trial count on the X-
axis and target cell value on the Y-axis. You can right-click either
graph to display the Graph Options dialog, where you can customize
the graph.
60 Evolver Commands
Graph Options The Graph Options dialog displays settings that control the titles,
Dialog legends, scaling and fonts used on the displayed graph.

Evolver Watcher When an Efficient Frontier analysis is underway, the Progress tab
– Progress Tab graphically shows how results are changing by optimization (for each
(Efficient possible efficient frontier constraint value), along with the current
Frontier) Efficient Frontier graph itself.
To identify the efficient frontier, Evolver performs multiple

optimizations to find the optimal value for each value of the
constraint. Instead of performing these optimizations separately in a
sequential manner, Evolver employs a more efficient algorithm in
which the multiple optimizations are performed as part of a single
analytical process. Each line in the Progress for Constraining Values
graph shows the progress for the optimal value at each possible
constraint value.
The Efficient Frontier graph shows the current optimal value across
all possible constraint values.
62 Evolver Commands
The Summary tab displays a summary table of adjustable cell values
Evolver Watcher
– Summary Tab tested during the optimization. If the Genetic Algorithm engine is in
use, tools for adjusting the crossover and mutation rate are also
displayed.
Genetic Algorithm Parameters allow you to change the Crossover

and Mutation rates of the genetic algorithm as the optimization is in
progress. Any changes made here will override the original setting of
these parameters and will take place immediately.
We almost always recommend using the default crossover of 0.5. You
might try turning it up as high as about 0.4 if you want to find the
best solution and are willing to wait longer. You can find more
information about these settings earlier in this chapter.

The Log tab displays a summary table of each trial run during the
Evolver Watcher
– Log Tab optimization. The log includes the results for the target cell, each
adjustable cell, and all constraints.
The Show options select to show a log of All Trials or only those
Trials where there was a Progress Step (that is, where the target cell
value improved). The log includes:
1) Elapsed Time. The start time of the optimization.
2) Result. The target cell value being optimized (including
penalties for soft constraints).
3) Input columns. The values in the adjustable cells.
4) Constraint columns. Indicating whether the constraints are
satisfied.
64 Evolver Commands
If the Genetic Algorithm engine is being used, the Population tab is
Evolver Watcher
– Population shown. This displays a grid that lists all the variables of each
Tab organism (each possible solution) in the current population. These
organisms are ranked in order from worst to best. Because this table
lists all organisms in the population, the Population Size setting in the
Evolver Settings dialog determines the number of organisms listed
here. (The default is 50.) In addition, the first column shows the value
of the target statistic for each organism.
Evolver Watcher If the Genetic Algorithm engine is being, the Diversity tab is shown.
– Diversity Tab The plot on the Diversity tab assigns colors to adjustable cell values,
based on how much the value of a given cell differs across the
population of organisms (solutions) that are stored in memory at a
given point. (Using the genetic optimization terminology, this is an
indication of the diversity that exists in the gene pool.) Each vertical
bar in the plot corresponds to one adjustable cell. Horizontal stripes
within each bar represent the values of that adjustable cell in different
organisms (solutions). The colors of the stripes are assigned by
dividing the range between the minimum and maximum value for a
given adjustable cell into 16 equal-length intervals; each of the
intervals is represented by a different color. For example, in the
picture the fact that the next-to-last vertical bar is single-color means
that the correseponding adjustable cell has the same value in each
solution in memory.

Evolver Watcher When you click the Stop button, the Evolver Watcher dialog Stopping
– Stopping Options tab is displayed. This dialog also appears as soon as one of
Options Tab the specified stopping conditions has been met. This tab includes the
options available for updating your worksheet with the best
calculated values for adjustable cells, restoring original values, and
generating an optimization summary report.
66 Evolver Commands
The Reports to Generate options include the following reports. (The
same options are also available from the Reports menu on the Evover
ribbon.)
• Optimization Summary. This summary report contains
information such as date and time of the run, the optimization
settings used, the value calculated for the target cell, and the
value for each of the adjustable cells.

• Log of All Trials. This report logs the results of all trials
performed. The values in magenta indicate that a constraint was
not met.
• Log of Progress Steps. This report logs the results of all trials that
improved the target cell value.
68 Evolver Commands
Application Settings Command
Displays the Application Settings dialog where program
defaults can be set
The Application Settings command (from the Utilities menu) opens
the Application Settings dialog. This lets you view or change a wide
variety of Evolver settings that will be used each time Evolver runs.
These include Stopping Defaults, Default Crossover and Mutation
Rates, and others.

Constraint Solver Command
Runs the Constraint Solver
The Constraint Solver command (from the Utilities menu) enhances
Evolver’s ability to handle model constraints. When Evolver runs an
optimization, it is assumed that the initial adjustable cell values
satisfy all the hard constraints, that is, the initial solution is valid. If
this is not the case, the algorithm can run many simulations before
finding a first valid solution. The problem is that if a model contains
multiple constraints, it can be difficult to find an initial valid solution.
If an Evolver model contains multiple hard constraints, and the
optimization process fails with all solutions invalid, you will be
notified so that the Constraint Solver can be run. The Constraint
Solver runs an optimization in a special mode, in which the goal is to
find a solution that satisfies all the hard constraints. The optimization
progress is shown in the same way as in regular optimizations. The
Progress Window shows the number of constraints that are met in the
original and best solutions.
Note: The Constraint Solver is useful when optimizing with the

Genetic Algorithm, whereas OptQuest optimizations do not
generally require the use of the Constraint Solver. Therefore, the
description in this section is relevant only for the Genetic Algorithm.
70 Evolver Commands
A button in the Progress window allows you to switch to the Evolver
Watcher. In the Constraint Solver mode the details of optimization
progress are available as in regular mode optimizations, in Progress,
Summary, Log, Population and Diversity tabs. However, in
Constraint Solver mode, the Watcher contains an additional
Constraint Solver tab. This tab shows the status of each hard
constraint (Met or Not Met) for the Best, Original, and Last solution.
A Constraint Solver optimization stops automatically when a solution

meeting all the hard constraints is found; it can also be stopped by
clicking a button in the progress window or in the Evolver Watcher.
After a Constraint Solver run, you can choose to keep the Best,
Original, or Last solution (from the Evolver Watcher Stopping
Options tab) just like in regular-mode optimizations.
Note there is no need to set up the Constraint Solver before a run. It
uses the settings specified in the model. The only difference is the
optimization goal, to find a solution that satisfies all the hard
constraints.
71
72 Evolver Commands
Appendix A: Optimization
Introduction
Palisade offers two optimization add-ins, Evolver and
RISKOptimizer. Evolver is a separate add-in in the Palisade
DecisionTools Suite. It is intended for optimization of deterministic
model, that is, models with no uncertainty. RISKOptimizer, which is
totally integrated into @RISK, combines optimization and simulation
to find optimal solutions to models with uncertainty. This appendix
provides background information on the optimization methods used
by Evolver. Because there is some overlap between Evolver and
RISKOptimizer methods, the discussion also includes some mention
of RISKOptimizer.
Optimization Methods
Traditional Excel-based optimization has focused on deterministic
models. Such models are comprised of:
• A “target” cell that you want to minimize or maximize.
• A set of input or “adjustable cells” whose values you control.
• A set of constraints that must be satisfied, often specified with
expressions such as TotalCost<=100 or A11>=0.
During an optimization, the adjustable cells are varied across
allowable ranges you specify. For each possible set of adjustable cell
values, the model is recalculated and a new value for the target cell is
generated. When the optimization is complete, an optimal solution (a
combination of adjustable cell values) has been found. This solution
is the combination of adjustable cell values that yields the optimal
value for the target cell while satisfying all of the constraints.
Some optimization models are much more difficult to solve than
others. For difficult problems, such as a model for finding the
shortest route between 200 cities, it is not feasible to examine every
possible solution. Such an approach would require weeks or months
of calculations on the fastest computers.
To solve such problems, it is necessary to search through a subset of
all possible solutions. This is accomplished with an algorithm. An
algorithm is a step-by-step description of how to solve a problem.
One type of algorithm is a hill-climbing algorithm:
1) Start at any initial point.
2) Move away from this point in some direction to get to a new
point.
3) If the target value at the new point is better, stay and repeat
step 2. If it is lower, go back to the previous point and try
again.
Numerous hill-climbing variations have been developed and
implemented in optimization software, but the devil is in the details,
especially in step 2. It isn’t at all clear how to choose the direction, or
how far to move in that direction, to get the next solution.
Unfortunately, a specific algorithm that works well on one problem or
type of problem does not always work well on other problems.
Hill-climbing algorithms have one other critical drawback. They tend
to find only local optimal solutions, not global optimal solutions.
Imagine a terrain where there are many hilltops but only one that is
highest. This highest hilltop represents the desired solution, the global
maximum, but because hilltop algorithms tend to search only locally
when making the move in step 2, they can easily find a lower hilltop,
a local maximum, and then stop, never reaching the global maximum.
Although hill-climbing algorithms can be modified so that they don’t
necessarily get stuck at local hilltops, they are naturally most
successful on problems guaranteed to have a single local (and global)
hilltop. There are many such problems, but there are also many that
do not have this attractive property.
74 Introduction
Types of Models
Several different types of models are typically optimized.
Linear Models In linear models, the target function and all constraint functions are
linear functions of the inputs. If a model is represented algebraically,
it is easy to identify whether the model is linear: all expressions will
be sums of products of constants and variables, where “variables”
refer to the adjustable cells. With complex spreadsheet models, it is
not always that easy to spot linearity or the lack of it. However, an
advantage of Evolver (and RISKOptimizer) is that you don’t need to
verify that a problem is, or is not, linear. Evolver will figure it out.
Linear models have been fairly easy to solve since the advent of
computers and the invention by George Dantzig of the simplex
method. A simple linear model can be solved most quickly and
accurately with a linear programming algorithm. Evolver can solve
linear models (with no uncertainty) efficiently. Its Industrial version
can handle an unlimited number of variables and constraints
Nonlinear There are very few real-world models that are truly linear. Many real
Models problems can be approximated well by linear models, as has been
documented in hundreds of successful applications of linear
programming. However, many problems are inherently nonlinear.
This can be due to economies of scale, complex chemical interactions
in oil blending, nonlinearities in financial portfolio measures of risk,
and many others. Such models cannot be represented realistically by
linear models. The nonlinearity is inherent and must be dealt with.
Unfortunately, these models are much more difficult to solve than
linear models. To deal with nonlinear models, Evolver and
RISKOptimizer don’t uses hill-climbing algorithms. Rather, they use
global optimization methods: Genetic Algorithms and the OptQuest
engine. This lets them jump around in the solution space, examining
many combinations of input values without getting stuck at a local
optimal solution. These methods retain in computer memory the
information on previously attempted solutions, and then they use that
information to better guess which scenarios are likely to be successful.
Non-smooth Many problems require the use of lookup tables and databases. For
Models example, in choosing the quantities of different materials to buy, you
might need to look up the prices charged for different quantities.
Tables and databases make models discontinuous (non-smooth). This
makes it difficult for hill-climbing to find optimal solutions. To
optimize these types of non-smooth models, Evolver and
RISKOptimizer have offered the Genetic Algorithms since their initial
releases. In version 6, the OptQuest optimization engine was added,
representing state-of-the-art methods in non-smooth optimization.
Fortunately, the Palisade products do not require you to identify a
model as non-smooth. The optimization algorithms in the Palisade
software work well on both smooth and non-smooth models, and the
software can choose the most appropriate algorithm automatically.
Combinatorial There is a large class of models that are very different from the
Models models discussed so far. Models where the outputs involve changing
the order of existing input variables, or grouping subsets of the
inputs, are called combinatorial models. These problems are usually
very difficult to solve because they often require exponential time;
that is, the amount of time needed to solve a problem with 4 variables
might be 4 x 3 x 2 x 1, and doubling the number of variables to 8 raises
the solving time to 8 x 7 x 6 x 5 x 4 x 3 x 2 x 1, or a factor of 1,680. The
number of variables doubles, but the number of possible solutions
that must be checked increases 1,680 times. For example, choosing
the starting lineup for a baseball team is a combinatorial problem. For
9 players, you can choose one out of the 9 as the first batter. You can
then choose one out of the remaining 8 as the second batter, one of the
remaining 7 will be the third, and so on. There are thus
9x8x7x6x5x4x3x2x1 (9 factorial) ways to choose a lineup of 9 players.
This is about 362,880 different orderings. Now if you double the
number of players, there are 18 factorial possible lineups, or
6,402,373,705,000,000 possible lineups!
The Evolver and RISKOptimizer algorithms, Genetic Algorithm and
OptQuest, are both capable of intelligently searching through the
possible permutations. This is much more practical than searching
through all possibilities, and it is much more efficient than examining
purely random permutations.
76 Introduction
Genetic Algorithms
Introduction
Evolver uses genetic algorithms as one of its optimization methods.
This section provides background information on genetic algorithms
to give insights on how they are used for optimizing models.
History
The first genetic algorithms were developed in the early 1970s by John
Holland at the University of Michigan. Holland was impressed by
the ease in which biological systems could perform tasks that eluded
even the most powerful super-computers. For example, animals can
flawlessly recognize objects, understand and translate sounds, and
generally navigate through a dynamic environment almost
instantaneously.
For decades, scientists have promised to replicate these capabilities in
machines, but this has proved to be very difficult. Most scientists
agree that any complex biological system that exhibits these qualities
has evolved to get that way.
Evolution has produced systems with amazing capabilities through
Evolution
Theory relatively simple, self-replicating building blocks that follow a few
simple rules:
1) Evolution takes place at the level of the chromosome. The
organism doesn’t evolve, but only serves as the vessel in which the
genes are carried and passed along. It is the chromosomes that are
dynamically changing with each re-arrangement of genes.
2) Nature tends to make more copies of chromosomes which
produce a more “fit” organism. If an organism survives long
enough, and is healthy, its genes are more likely to be passed along to
a new generation of organisms through reproduction. This principle
is often referred to as “survival of the fittest.” Remember that “fittest”
is a relative term; an organism only needs to be fit in comparison to
others in the current population to be “successful.”
3) Diversity must be maintained in the population. Seemingly
random mutations occur frequently in nature, and these ensure
variation in the organisms. These genetic mutations often result in a
useful or even vital feature for a species’ survival. With a wider
spectrum of possible combinations, a population is also less
susceptible to a common weakness that could destroy them all (such
as a virus) or other problems associated with inbreeding.
Once evolution is broken down into these fundamental building
blocks, it becomes easier to apply these techniques to the
computational world.
Holland began applying these properties of evolution to simple
strings of numbers that represented chromosomes. He first encoded
his problem into binary strings (rows of 1s and 0s) to represent the
chromosomes, and then he had the computer generate many of these
“bit” strings to form a whole population of them. A fitness function
was programmed that could evaluate and rank each bit string, and
those strings that were deemed most “fit” would exchange data with
others through a “crossover” routine to create “offspring” bit strings.
Holland even subjected his digital chromosomes to a “mutation”
operator, which injected randomness into the resulting “offspring”
chromosomes to retain diversity in the population. This fitness
function replaced the role of death in the biological world to
determine which strings were good enough to continue breeding and
which would no longer be kept in memory.
The program kept a given number of these “chromosomes” in

memory, and this entire “population” of strings continued to evolve
until they maximized the fitness function. The result was then de-
coded back to its original values to reveal the solution.
Holland’s original genetic algorithm was quite simple, yet remarkably
robust in finding optimal solutions to a wide variety of problems.
Many custom programs today solve very large and complex real-
world problems using only slightly modified versions of this original
genetic algorithm.
78 Genetic Algorithms
As interest swelled in academic circles, as serious computational
Modern power began moving its way into mainstream desktop machines,
Adaptations of standards like Microsoft Windows and Excel made design and
Genetic
maintenance of complex models easier. The use of real numbers
Algorithms
rather than bit string representations eliminated the difficult task of
encoding and decoding chromosomes.
The popularity of the genetic algorithm is now growing
exponentially, with seminars, books, magazine articles, and
knowledgeable consultants popping up everywhere. The
International Conference of Genetic Algorithms is already focusing on
practical applications, a sign of maturity that eludes other “artificial
intelligence” technologies. Many Fortune 500 companies, from
brokerage firms to power plants, phone companies, restaurant chains,
automobile manufacturers and television networks, employ genetic
algorithms regularly to solve real-world problems.
A Biological Example
Consider a simple example of evolution in the biological world (on a
small scale). Here, “evolution” means any change in the distribution
or frequency of genes in a population. Of course, the interesting thing
about evolution is that it tends to lead to populations that are
constantly adapting to their environments.
Specifically, consider a population of mice. These mice exhibit two
sizes, small and large, and they exhibit two colors, light or dark. Our
population consists of the following eight mice:
One day, cats move into the neighborhood and start eating mice. It
turns out that darker mice and smaller mice are harder for the cats to
find. Therefore, different mice have different odds of avoiding the
cats long enough to reproduce. This affects the nature of the next
generation of mice. Assuming the old mice die soon after
reproducing, the next generation of mice looks like this:
Notice that large mice, light mice, and especially large, light mice, are
having trouble surviving long enough to reproduce. This continues in
the next generation:
Now the population consists mostly of small, dark mice, because

these mice are better suited for survival in this environment than
other kinds of mice. Similarly, as the cats begin to go hungry with
less mice to eat, perhaps those cats who prefer a lunch of grass will be
better adapted, and pass along their grass-loving gene to a new
generation of cats. This is the central concept of “survival of the
fittest.” More precisely, it could be phrased “survival until
reproduction.” In evolutionary terms, being the healthiest bachelor in
the population is worthless because you must reproduce for your
genes to influence future generations.
A Digital Example
Now consider a problem with two variables, X and Y, that produce a
result Z. If you calculated and plotted the resulting Z for every
possible X and Y values, you would see a solution “landscape”
emerge. If you are trying to find the maximum “Z”, the peaks of the
function are “good” solutions, and the valleys are “bad” ones.
When a genetic algorithm is used to maximize the function, it starts
by creating several possible solutions or scenarios at random (the
black dots), rather than just one starting point. It then calculates the
function’s output for each scenario and plots each scenario as one dot.
Next it ranks all of the scenarios by height, from best to worst. It
keeps the scenarios from the top half and discards the others.
Each of the three remaining scenarios duplicates itself, bringing the
number of scenarios back up to six. Now comes the interesting part:
Each of the six scenarios is composed of two adjustable values, X and
Y. The scenarios pair off with each other at random. Each scenario
exchanges the first of its two adjustable values with the corresponding
value from its partner. For example:
Before After
Scenario 1 3.4, 5.0 2.6, 5.0
Scenario 2 2.6, 3.2 3.4, 3.2
This operation is called crossover. When the six scenarios randomly

mate and perform crossover, there is a new set of scenarios:
In the above example, the original three scenarios, a, b, and c, paired

up with the duplicates, A, B, C, to form the pairs aB, bC, bA. These
pairs then switched values for the first adjustable cell, which is
equivalent in the diagram to exchanging the x and y coordinates
between pairs of dots. The population of scenarios has just lived
through a generation, with its cycle of “death” and “birth.”
Notice that some of the new scenarios result in lower height than any
in the original generation. However, one scenario has moved high up
on the tallest hill, indicating progress. If the population evolves for
another generation, it might lead to the following:
You can see how the average performance of the population of
scenarios increases over the last generation. In this example, there is
not much room left for improvement. This is because there are only
two genes per organism, only six organisms, and no way for new
genes to be created. This means there is a limited gene pool. The gene
pool is the collection of all the genes of all organisms in the
population.
Genetic algorithms can be made much more powerful by replicating
more of the inherent strength of evolution in the biological world;
increasing the number of genes per organism, increasing the number
of organisms in a population, and allowing for occasional random
mutations. In addition, the algorithms can choose the scenarios that
will live and reproduce with a random element that has a slight bias
towards those that perform better, instead of simply choosing the best
performers to breed.
How Evolver's Genetic Algorithm Optimization is
Implemented
This section describes more specifically how Evolver’s Genetic
Algorithm optimization is implemented.
Note: You do not need to know this material to use Evolver.
The majority of Evolver’s genetic algorithm technology such as the
Recipe and Order solving methods are based on academic work in the
genetic algorithm field over the last decade or two. However, most of
the descendant solving methods included with Evolver, and the
multiple groups of adjustable cells, backtracking, strategy, and
probability features are unique to Evolver.
Evolver uses a steady-state approach. This means that only one
organism is replaced at a time, rather than an entire “generation”
being replaced. This steady-state technique has been shown to work
as well or better than the generational replacement method. To
determine the equivalent number of “generations” Evolver has run,
you can divide the number of individual trials it has explored by the
size of the population.
Selection When a new organism is created, two parents are chosen from the
current population. Organisms that have high fitness scores are more
likely to be chosen as parents.
In Evolver, parents are chosen with a rank-based mechanism. Instead
of some genetic algorithm systems, where a parent's chance of being
selected for reproduction is directly proportional to its fitness, a
ranking approach offers a smoother selection probability curve. This
prevents good organisms from completely dominating the evolution
from an early point.
Crossover Because each solving method adjusts the variables in different ways,
Evolver employs a different crossover routine optimized for each type
of problem.
The basic Recipe solving method performs crossover using a uniform
crossover routine. This means that instead of chopping the list of
variables in a given scenario at some point and dealing with each of
the two blocks (called “single-point” or “double-point” crossover),
two groups are formed by randomly selecting items to be in one
group or another. Traditional x-point crossovers can bias the search
with the irrelevant position of the variables, whereas the uniform
crossover method is considered better at preserving schema, and can
generate any schema from the two parents.
The order solving method performs crossover using a similar
algorithm to the order crossover operator described in L. Davis’
Handbook of Genetic Algorithms. This selects items randomly from
one parent, finds their place in the other parent, and copies the
remaining items into the second parent in the same order as they
appear in the first parent. This preserves some of the sub-orderings in
the original parents while creating some new sub-orderings.
Mutation Like crossover, mutation methods are customized for each of the
different solving methods. The basic Recipe solving method
performs mutation by looking at each variable individually. A
random number between 0 and 1 is generated for each of the variables
in the organism, and if a variable gets a number that is less than or
equal to the mutation rate (for example, 0.06), that variable is
mutated. The amount and nature of the mutation is automatically
determined by a proprietary algorithm. Mutating a variable involves
replacing it with a randomly generated value (within its valid min-
max range).
To preserve all the original values, the Order solving method
performs mutation by swapping the positions of some variables in the
organism. The number of swaps performed is increased or decreased
proportionately to the increase and decrease of the mutation rate
setting (from 0 to 1).
Replacement Because Evolver uses a rank-ordered rather than generational
replacement method, the worst-performing organisms are always
replaced with the new organism that is created by selection,
crossover, and mutation, regardless of its fitness “score”.
Hard constraints are implemented with Palisade’s proprietary
Constraints
“backtracking” technology. If a new offspring violates some
externally imposed constraints, Evolver backtracks towards one of the
parents of the child, changing the child until it falls within the valid
solution space.
86
OptQuest
Introduction
The OptQuest Engine incorporates metaheuristics to guide its search
algorithm toward better solutions. This approach remembers which
solutions worked well and recombines them into new, better
solutions. In addition, this technique does not get trapped in local
solutions or get thrown off course by noisy (uncertain) model data.
The OptQuest Engine combines Tabu search, scatter search, integer
programming, and neural networks into a single, composite search
algorithm that provide maximum efficiency in identifying new
scenarios.
Linear Constraints
OptQuest generates solutions that almost always meet all the
specified linear constraints, and time is saved by not evaluating
invalid solutions. (OptQuest can occasionally generate a solution that
does not meet a linear constraint, due to the fact that computers
cannot handle calculations with infinite precision.)
The Bakery.xlsx and several other example files that accompany
Evolver illustrate OptQuest linear constraint handling. All of the
constraints are linear, and all the solutions generated by OptQuest
will be valid. More specifically, the “SumProduct” formula in the
constrained cell expresses a linear function of the adjustable cells.
Other constrained cells are also linearly dependent on the adjustable
cells.
Nonlinear Constraints
OptQuest handles nonlinear constraints efficiently, including
situations in which the original values of adjustable cells are invalid
(they do not satisfy the specified constraints). The Genetic Algorithm
generally requires the original adjustable cell values to satisfy the
constraints. If the original solution is invalid, the Constraint Solver
tool finds a valid solution, providing a starting point for an
optimization that uses the Genetic Algorithm. OptQuest does not
require the use of the Constraint Solver. If the original solution is
invalid, OptQuest can start the optimization by generating a sequence
of invalid solutions. However, during this stage, it collects
information about how much each solution falls short of meeting the
constraints, with the goal of finding valid solutions.
88 OptQuest
Improving Speed
When you use Evolver to solve a problem, you are using both the
Evolver library of compiled routines to control the process and Excel’s
spreadsheet evaluation functionality to examine different scenarios.
A large percentage of the time used by Evolver is actually used by
Excel as it recalculates your spreadsheet model. There are a number
of things that you can do to speed up Evolver optimization and
Excel’s recalculation process.
• The speed of Evolver is directly related to the speed of your
computer processor. If you double the speed of your processor,
Evolver will be able to evaluate twice as many trials in the same
amount of time.
• Try to avoid redrawing in your window. Drawing graphics and
numbers on the screen takes time, sometimes more than half the
time spent optimizing. If you have charts or graphs on the sheet,
they will slow down the recalculation time significantly. You can
tell Excel not to spend time drawing while Evolver is solving a
problem by turning off the “Show Excel Recalculations” option in
the View tab of the Optimization Settings dialog, or by selecting
“Minimize Excel at Start” in the same tab. You can see how much
faster your problem is working by watching the status bar.
• You can set tighter ranges for the adjustable cells. This creates a
smaller area in which Evolver must search for solutions.
However, make sure that your ranges allow enough freedom for
Evolver to explore all realistic solutions.
90 Improving Speed
Appendix B: Automating
Evolver
Evolver comes with a complete macro language for building custom
applications that use Evolver's capabilities. Evolver's custom
functions can be developed in Visual Basic for Applications (VBA) for
setting up and running optimizations and displaying the results from
optimizations. For more information on this programming interface,
see the Developer Kit (XDK) items on the Evolver Help menu.
91
92 Improving Speed
Appendix C: Troubleshooting /
Q&A
This section answers some commonly asked questions regarding
Evolver and keeps you up to date on common questions, problems,
and suggestions. After reading this section, you can call Palisade
customer support at the numbers listed in the beginning of this
manual.
Q: Why am I having trouble getting a valid solution from Evolver?

A: Make sure that the Evolver dialog is set up correctly. Most of the
problems are associated with the setting of the variables. Each
group of adjustable cells should be exclusive, in that no single cell
or range of cells is being treated with more than one solving
method.
Q: Can Evolver deal with concepts or categories instead of just

numbers?
A: Evolver can indirectly deal with any kind of data. You can use a
lookup table in Excel to translate between integers and strings of
text. Evolver (like all computer programs) ultimately can only
deal with numbers, but your interface can use those numbers to
represent and display any strings.
Q: Even though I’m filling in the dialogs the same way, and letting
Evolver run the same amount of time, why does Evolver
sometimes find different solutions?
A: As is the case with natural selection in the biological world, the
Evolver genetic algorithm will not always follow the same path
when searching for solutions (unless you use a fixed random
number generator seed). Ironically, it is this “unpredictability”
that allows Evolver to solve more types of problems, and often
find better solutions, than traditional techniques. Evolver’s
genetic algorithm engine is not just executing a series of pre-
programmed commands, or plugging values through a
mathematical formula. It is efficiently experimenting with many
random hypothetical scenarios simultaneously, and then refining
93
the search through many “survival-of-the-fittest” operators that
also contain random elements. A similar reason can be given in
the case of OptQuest, the other optimization algorithm in Evolver.
OptQuest experiments with different possible solutions, keeping
track of the results of the experiments performed so far. The
randomness involved in the process keeps it from returning a
locally optimal solution if there is a better globally optimal
solution.
It is possible to specify settings that will make two Evolver
optimizations return identical results. To do this, type any integer
in the Initial Seed field in the View tab in the Optimization
Settings dialog. In addition, the “Trials” or “Progress” stopping
condition needs to be selected in the Runtime tab of that dialog.
Q: Why is the best solution found not changing?
A: You might have specified the wrong target cell in the Evolver
Model dialog, a cell that is not affected by the adjustable cell
values. A proper target cell has a formula that depends, directly
or indirectly, on the adjustable cells.
Q: Some of the cells in my spreadsheet model contain “####”

symbols.
A: If the cell is too small to display all of its contents, it will display
several #### signs. Increase the size of the cell.
Q: Evolver is working OK, but is there any simple way to get better
results?
A: Consider loosening the constraints in the problem, including
variable ranges. You can also change some of your hard
constraints to soft constraints via penalty functions. Finally,
remember that the longer you let Evolver explore the possibilities,
the more likely it is to find the optimal solution.
The more scenarios Evolver can run through, the better. Speed
optimization by turning off the “Show Excel Recalculations”
option in the View tab of the Optimization Settings dialog.
94 Improving Speed
Appendix D: Additional
Resources
The following list represents a select sampling of genetic algorithm
and artificial-life-related materials. A star (*) indicates a Palisade
favorite.
Books
• Bolles, R.C., & Beecher, M.D. (Eds.). Evolution and Learning. Lawrence
Erlbaum, 1988.
• Beer, R.D. Intelligence as Adaptive Behavior: An Experiment in
Computational Neuroethology. Academic Press, 1990.
• Davis, Lawrence. Genetic Algorithms and Simulated Annealing. Morgan
Kaufman, 1987.
* Davis, Lawrence. Handbook of Genetic Algorithms. Van Nostrand
Reinhold, 1991.
• Darwin, Charles. On The Origin of Species. Penguin Classics, 1985
(originally 1859).
* Dawkins, Richard. The Selfish Gene. Oxford University Press, 1976.
• Eldredge, N. Macroevolutionary Dynamics: Species, Niches, and Adaptive
Peaks. McGraw-Hill, 1989.
• Fogel, L., Owens, J., and Walsh, J. Artificial Intelligence through Simulated
Evolution. John Wiley and Sons, 1966.
• Goldberg, David. Genetic Algorithms in Search, Optimization, and
Machine Learning. Addison-Wesley, 1989.
• Holland, J.H. Adaptation in Natural and Artificial Systems. University of
Michigan Press, 1975.
• Koza, John. Genetic Programming. MIT Press, 1992.
* Langton, C.L. Artificial Life. MIT Press, 1989.
• Levy, Steven. Artificial Life. Pantheon, 1992.
• Meyer, J.-A., & S.W. Wilson (Eds.). Proceedings of the First International
Conference on Simulation of Adaptive Behavior: From Animals to
Animats. MIT Press/Bradford Books, 1991.

* Proceedings of the Sixth International Conference (ICGA) on Genetic
Algorithms. Morgan Kaufman Publishing, 1995.
• Proceedings of the Workshop on Artificial Life. Addison-Wesley
Publishing, 1990.
• Rawlins, Gregory. Foundations of Genetic Algorithms. Morgan Kaufman
Publishing, 1991.
• Richards, R.J. Darwin and the Emergence of Evolutionary Theories of Mind
and Behavior. U. Chicago Press, 1987.
• Williams, G.C. Adaptation and Natural Selection. Princeton U. Press,
1966.
Articles
* Antonoff, Michael, “Software by Natural Selection,” Popular Science, p. 70-
74. (October, 1991)
• Arifovic, Jasmina, “Genetic Algorithm Learning and the Cobweb Model,”
Journal of Economic Dynamics & Control v. 18 p. 3 (January, 1994)
* Begley, S. “Software au Naturel,” Newsweek p. 70 (May 8, 1995)
• Celko, Joe, “Genetic Algorithms and Database Indexing,” Dr. Dobb’s
Journal p. 30 (April, 1993)
• Ditlea, Steve, “Imitation of Life,” Upside Magazine p. 48 (November, 1994)
• Gordon, Michael, “User-based Document Clustering by Redescribing
Subject Descriptions with a Genetic Algorithm,” Journal of the American
Society for Information Science v. 42 p. 311 (June, 1991)
• Hedberg, Sara, “Emerging Genetic Algorithms,” AI Expert, p. 25-29
(September, 1994)
• Hinton, G.E., & Nowlan, S.J, “How Learning Can Guide Evolution,”
Complex Systems 1: p. 495-502 (1987).
* Kennedy, Scott, “Genetic Algorithms: Digital Darwinism,” Hitchhicker’s
Guide to Artificial Intelligence. Miller Freeman Publishers (June, 1995)
• Kennedy, Scott, “Five Ways to a Better GA,” AI Expert, p. 35-38 (December,
1993)
• Lane, A., “The GA Edge in Analyzing Data,” AI Expert p. 11 (June, 1995)
• Lee, Y.C. (Ed.), “ Evolution, Learning, and Cognition,” World Scientific
(1988)
• Levitin, G and Rubinovitz, J., “Genetic Algorithm for Linear and Cyclic
Assignment Problem,” Computers & Operations Research v. 20 p. 575
(August, 1993)
• Marler, P., & H.S. Terrace. (Eds.). The Biology of Learning. Springer-
Verlag (1984)
96 Improving Speed
• Mendelsohn, L., “Evolver Review,” Technical Analysis of Stocks and
Commodities. p. 33 (December, 1994)
• Maynard Smith, J., “When Learning Guides Evolution,” Nature 329: p. 761-
762 (1987)
• Murray, Dan, “Tuning Neural Networks with Genetic Algorithms,” AI
Expert p. 27 (June, 1994)
• Wayner, Peter, “Genetic Algorithms: Programming Takes a Valuable Tip
from Nature,” Byte Magazine v. 16 p. 361 (January, 1991)
Magazines & Newsletters

• Advanced Technology for Developers (monthly newsletter). Jane
Klimasauskas, Ed., High-Tech Communications, 103 Buckskin Court,
Sewickley, PA 15143 (412) 741-7699
• AI Expert (monthly magazine). Larry O’Brien, Ed., 600 Harrison St., San
Francisco, CA 94107 (415) 905-2234. *Although AI Expert ceased
publishing in the spring of 1995, its back issues contain many useful
articles. Miller-Freeman, San Francisco.
• Applied Intelligent Systems (bimonthly newsletter). New Science
Associates, Inc. 167 Old Post Rd., Southport, CT 06490 (203) 259-1661
• Intelligence (monthly newsletter). Edward Rosenfeld, Ed., PO Box 20008,
New York, NY 10025-1510 (212) 222-1123
• PC AI Magazine (monthly magazine). Joseph Schmuller, Ed., 3310 West
Bell Rd., Suite 119, Phoenix, AZ 85023 (602) 971-1869
• Release 1.0 (monthly newsletter). Esther Dyson, Ed., 375 Park Avenue,
New York, NY 10152 (212) 758-3434
• Sixth Generation Systems (monthly newsletter). Derek Stubbs, Ed., PO Box
155, Vicksburg, MI, 49097 (616) 649-3592

98 Improving Speed
Appendix E: Glossary
For additional information on any term, refer to the Index section of
this manual.
Algorithm A mathematically based step-by-step method for solving a certain

kind of problem. All computer programs are built by combining
algorithms.
Adjustable Cell A spreadsheet cell whose value can be adjusted by Evolver to try to
optimize the value of the target cell. An adjustable cell should always
contain a number, not a formula.
Constraints Constraints are conditions that should be satisfied for a solution to be
valid. Hard constraints must be satisfied. Soft constraints can be
violated to some extent, but this penalizes the target to be optimized.
Crossover In Evolver, the term crossover is used to express the computational
equivalent to crossing over genetically, where an exchange between
variables yields new combinations of scenarios.
Efficient This term is used in optimization models with two objectives. It
Frontier represents the set of all solutions that are not “dominated” by other
solutions. Specifically, a solution is on the efficient frontier is the only
way to improve one objective will make the other objective worse.
Objective This is the term commonly used to indicate the target cell value to be
Function optimized. It is a function of the adjustable cell values.
Genetic This is a type of optimization algorithm inspired by, and somewhat

Algorithm similar to, the process of evolution in the biological world, where the
fittest survive to reproduce.
Generation In the field of genetic algorithms, each completely new population of
“offspring” solutions is a new “generation.” Some genetic algorithm
routines mate all members of a population at once, creating a whole
new “generation” of offspring organisms that replaces the previous
population. Evolver evaluates and replaces one organism at a time
(rank-ordered) and thus does not use the term “generation” in its
documentation.
99
Global Optimal This is the best solution, in terms of the target cell value, among all
Solution possible solutions that satisfy the constraints.
Group of This is a set of one or more adjustable cells ranges that are treated as a
Adjustable cells group. Each group will use the same solving method. For example,
there could be two groups, one using the Recipe solving method and
the other using the Order solving method.
Hard This is a constraint that must be satisfied. Any trial solution that
Constraints doesn’t satisfy all of the hard constraints is invalid and ignored.
Hill-Climbing An optimization procedure that starts from one trial solution and
Algorithm repeatedly moves to the next trial solutions in small steps in the
direction that will most improve the target cell value. Hill-climbing
algorithms are simple, but they have two drawbacks. First, much
work can be needed to find the direction of most improvement.
Second, the algorithms usually seach only locally, so they can get
stuck at a local optimum and never reach the global optimum.
Local Optimal This refers to a solution that is better, in terms of the target cell value,
Solution than all nearby solutions. However, there can be a better solution, the
global optimal solution, somewhere else in the solution space.
Mutation In the biological world, gene mutation is the source of variation
needed for effective natural selection. Likewise, a genetic algorithm
uses mutation techniques to maintain diversity in a population of
possible solutions.
OptQuest This is a set of optimization algorithms. OptQuest’s methods
Search, Neural Networks, and Scatter Search into a single composite
method.
Penalty This is a function that assigns a penalty to the violation of a soft
Function constraint, where larger violations lead to larger penalties. The
penalty is added to (for minimization) or subtracted from (for
maximization) the target cell value, so that trade-offs can be made.
Population This is the set of trial solutions that Evolver's Genetic Algorithm keeps
in memory, so that new (and better) solutions can be generated from
these.
Solution This term is typically used to indicate any set of values for the
adjustable cells. It is often preceded by the word “trial” to indicate
that it isn’t necessarily the optimal solution.
100 Improving Speed

Soft Constraints These are constraints that can be violated, but only at the expense of
incurring a “penalty.”
Solving Method This specifies the approach Evolver will use to optimize a particular
type of problem. Evolver includes six of these methods: Recipe,
Order, Budget, Grouping, Project, and Schedule.
Target Cell This is the cell in your model to be optimized. It is also referred to as
the objective function.
Trials Each trial in Evolver’s optimization process corresponds to a new trial
solution.
Appendix E: Glossary 101

102 Improving Speed
Index
A Guided Tour Videos, 13
Icons, 31
Adjustable Cell Group. See Evolver Improving Speed, 89
Adjustable Cell Range. See Evolver Installation, 11
Application Settings Command. See Instructional Materials, 13
Evolver Commands License Manager, 12
Loading, 16
Optimization Goal, 34
B Professional and Industrial
Bakery Example Tutorial, 15 Versions, 7
Budget Method. See Solving Methods Progress Window, 59
Quick Start Tutorials, 13
Reports, 66
C Ribbon, 16
Constraint. See Evolver Software Activation, 12
Constraint Settings Dialog. See Stopping Conditions, 23, 49
Evolver Dialogs Student Version, 10
Constraint Solver Command. See Summary Report, 28
Evolver Commands System Requirements, 10
Target Cell, 18, 34
Troubleshooting, 93
D Uninstalling, 11
Watcher, 26, 60
DecisionTools Suite. See Palisade
Diversity Tab, 65
Corporation
Log Tab, 64
Developer Kit (XDK). See Evolver
Population Tab, 65
Deviation Keyword. See Evolver
Progress Tab, 60
Stopping Options Tab, 66
E Summary Tab, 63
What Is It?, 1
Efficient Frontier Analysis, 5, 45
Why Use It?, 2
Evolver
Evolver Commands
Adjustable Cell Group, 20, 34, 36
Application Settings, 69
Adjustable Cell Range, 18, 34
Constraint Solver, 70
Analysis Type, 34, 45
Model Definition, 33
Automating with VBA, 91
Settings, 48
Constraint, 20, 43
Start Optimization, 59
Deviation Keyword, 46
Evolver Dialogs
Hard, 20, 45
Constraint Settings, 43
Penalty Function, 46
Model Definition, 17, 33
Precision, 45
Optimization Settings, 48
Soft, 20, 45, 46
Efficient Frontier Runtime Tab,
Developer Kit (XDK), 13, 91
51
Example Spreadsheets, 13
Engine Tab, 52
Index 103
Macros Tab, 57 Non-smooth Models, 75
Runtime Tab, 48 Objective Function, 15
View Tab, 57 Running, 24
Stopping, 27
G Summary Report, 28
What Is It?, 1
GA. See Genetic Algorithm Optimization Settings Dialog. See
Genetic Algorithm, 4, 53, 77 Evolver Dialogs
Crossover, 53, 83 OptQuest, 4, 53, 87
Evolver Implementation, 83 Linear Constraints, 87
Mutation, 53, 84 Nonlinear Constraints, 87
Operators, 55 Order Method. See Solving Methods
Parameters, 63
Population Size, 53
References, 95
P
Replacement Method, 84 Palisade Corporation
Settings, 53 Contacting, 9
Global Optimal Solution. See DecisionTools Suite, 11
Optimization Getting Help, 8
Grouping Method. See Solving Penalty Function. See Evolver
Methods Project Method. See Solving Methods
H R
Hard Constraint. See Evolver Recipe Method. See Solving Methods
L S
Linear Programming, 5 Schedule Method. See Solving
Simplex Method, 75 Methods
Local Optimal Solution. See Settings Command. See Evolver
Optimization Commands
Simplex Method, 75
Soft Constraint. See Evolver
M Solving Methods, 20, 36
Model Definition Command. See Budget, 39
Evolver Commands Grouping, 38
Model Definition Dialog. See Evolver Order, 38
Dialogs Project, 40
Recipe, 38
Schedule, 41
O Start Optimization Command. See
Objective Function. See Optimization Evolver Commands
Optimization Stopping Conditions. See Evolver
Combinatorial Models, 76
Global Optimal Solution, 74 T
Linear Models, 75
Local Optimal Solution, 74 Target Cell. See Evolver
Methods, 73
Nonlinear Models, 75
104 Improving Speed

105

Evolver7 en

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Evolver7 en

Загружено:

Авторское право:

Доступные форматы

User’s Guide

Chapter 1: Welcome to Evolver 1

Chapter 2: Getting Started 7

Chapter 3: Evolver Tutorial 15

Evolver Commands ...................................................................... 33

Appendix B: Automating Evolver 91

Appendix C: Troubleshooting / Q&A 93

Appendix D: Additional Resources 95

Table of Contents iii

Why Use Evolver?

How does Evolver work?

Efficient Frontier Analysis

Chapter 1: Welcome to Evolver 5

About This Version

Evolver Professional and Industrial Versions

Working with your Operating Environment

Chapter 2: Getting Started 7

Chapter 2: Getting Started 9

Evolver System Requirements

The DecisionTools Suite

Chapter 2: Getting Started 11

Quick Start Tutorials

Guided Tour Videos

Examples and Videos for the XDK

Chapter 2: Getting Started 13

The Evolver Model Dialog

The Evolver Model dialog is designed so you can describe you

Chapter 3: Evolver Tutorial 17

Adding Adjustable Cell Ranges

Chapter 3: Evolver Tutorial 19

Chapter 3: Evolver Tutorial 21

In the Optimization Settings dialog you can select any combination of

Chapter 3: Evolver Tutorial 23

The two options for Show Excel Recalculations are:

Chapter 3: Evolver Tutorial 25

These same options will automatically appear when any of the

Chapter 3: Evolver Tutorial 27

This report is useful for comparing the results of successive

Chapter 3: Evolver Tutorial 29

Lets you view or make changes to the settings for

Starts the optimization process

Provides several types of reports of optimization

Provides Application Settings and several other

Provides Evolver Help files

Chapter 4: Evolver Command Reference 33

Chapter 4: Evolver Command Reference 35

There are three possible settings:

Again, remember that each group of adjustable cell ranges shares a

Chapter 4: Evolver Command Reference 37

Chapter 4: Evolver Command Reference 39

The precedence tasks range should be specified as being n rows by m

Chapter 4: Evolver Command Reference 41

The main entries for a constraint are:

Chapter 4: Evolver Command Reference 43

Chapter 4: Evolver Command Reference 45

More information about penalty functions is as follows:

• Viewing the Penalties Applied. When a penalty is applied

Chapter 4: Evolver Command Reference 47

Chapter 4: Evolver Command Reference 49

The options on the Efficient Frontier Runtime tab include:

Chapter 4: Evolver Command Reference 51

Genetic algorithms use genetic operators to create new members of

Macros can be executed at the following times during an