User Guide

556 E. Technology Way

Orem, Utah 84097
801.223.4600
ii

Disclaimer Copyright Information

The information in this guide is provided by Pro- © 2010 ProModel Corporation
Model Corporation to document SimRunner. The
All rights reserved
content of this manual is subject to change with-
out notice and does not represent a commitment Printed in the United States of America
on the part of ProModel Corporation. The soft-
ware described in this guide is supplied under a
license agreement and may be copied only under
the terms of the license agreement. No part of this
SimRunner, ProModel, MedModel, and Service-
guide may be reproduced, transmitted, or distrib-
Model are registered trademarks of ProModel
uted by any means, electronic or mechanical, for
Corporation.
purposes other than the owner’s personal use
without express written permission from Pro- Text indicator graphics Copyright New Vision
Model Corporation. Technologies Inc.
Microsoft is a registered trademark and Windows
is a trademark of Microsoft Corporation.
IBM is a registered trademark of International
Business Machines Corporation.
Netware and Novell are a registered trademarks
of Novell, Inc.

9/22/10
 SimRunner iii
User Guide

Table of Contents

Table of Contents ......................................................................... iii

Introduction ....................................................................................1

About the User Guide ........................................................... 1

Symbols and Notation .................................................... 2
Product Support .............................................................. 2
Modeling Services .......................................................... 3
Reporting Suggestions ................................................... 4

Chapter 1:
SimRunner Concepts ...........................................................................................5

What is SimRunner? ............................................................. 5

Benefits .......................................................................... 5
Where Do I Begin? ............................................................... 6
Concepts and Theories ................................................... 7
General Procedure ........................................................ 10
Pitfalls .......................................................................... 11

Chapter 2:
Getting Started ...................................................................................................13

The Interface ....................................................................... 13

Set Up a Project .................................................................. 14
iv

Select Model/Project .................................................... 14

Define Objectives ......................................................... 14
Define Inputs ................................................................ 17
Set Options ................................................................... 19
Analyze Model ................................................................... 22
Optimize Model .................................................................. 25
Introduction .................................................................. 25
Optimization Concepts ................................................. 26
SimRunner Optimization Techniques .......................... 27
Menus ................................................................................. 29
File menu ...................................................................... 29
Options menu ............................................................... 29
Help menu .................................................................... 29

Chapter 3:
Building Projects ................................................................................................31

Getting Started .................................................................... 31

Set Up Project ..................................................................... 32
Analyze Model ................................................................... 36
Optimize Model .................................................................. 38

Appendix ......................................................................................41

Glossary ........................................................................................43

Bibliography .................................................................................47

Index .............................................................................................49
 SimRunner 1
User Guide

Introduction

0.1 About the User Guide Appendix

The SimRunner User Guide is a reference that Contains a list of suggested readings.
will guide you through the process of building
simulation optimization projects. This guide Glossary
contains detailed information on the use of Contains definitions and descriptions for many of
features and capabilities found within SimRunner the terms and concepts common to simulation
and serves to complement the product training. and modeling.

Chapter 1 Bibliography
SimRunner Concepts
Referenced materials.
As an introduction to simulation optimization,
this chapter begins by reviewing some of the
theory behind optimization. The remainder of the
discussion revolves around the basic approach to
building, running, and examining results from
models.

Chapter 2
Getting Started
Introduces the product interface and allows you
to become more familiar with the SimRunner
environment. Contains an in-depth discussion
and explanation of each step you must follow to
build, run, and examine an optimization project.

Chapter 3
Building Projects
Steps you through the process of creating a
project.
2
About the User Guide

0.1.1 Symbols and Notation 0.1.2 Product Support

To better help you navigate this text, please Technical support is available to all licensed
review the following symbols and conventions. SimRunner users with current maintenance and
support agreements. Support representatives are
Keyboard glad to answer specific questions you may have
and offer direction in solving specific modeling
The names of keys are displayed in capital letters. challenges you encounter. See also Modeling
For example, ESC refers to the Escape key and Services on page 3.
CTRL refers to the Control key.
Keys are frequently specified in combinations or
in a sequence of keystrokes. For example,
CTRL + L means to hold down the CTRL key
while pressing L. When key commands are set
off by commas (e.g., ALT + N, R), press and Technical support
release each of these keys (or key combinations)
in the order listed. The term “arrow keys” refers Technical support is available via: internet, e-
collectively to the ×, Ø, Õ, and Ö cursor keys. mail, telephone, and fax. When you contact
technical support, please be prepared to provide
Text your user profile and a description of any
problems you experience.
Specific text you are asked to type is shown in
bold type. For example, if you are directed to User Profile
type cd simrun, you would type the lowercase • Your name and company
letters “cd” followed by a space and the letters
• The license # found on your security key
“simrun.”
• Hardware make and configuration
Placeholders for things such as file names and • Network information (if applicable)
directories are shown in italics. For example, if • Version number of Windows and SimRun-
you are directed to type filename.res, enter the ner
name of the file you wish to use (e.g.,
model_1.res). Problem Description
• Brief description of the problem you are
experiencing
• What you were doing when the problem
occurred
• The exact wording of any messages that
appeared on your screen
 SimRunner 3
User Guide

Internet 0.1.3 Modeling Services

www.promodel.com/solutionscafe
If you find yourself in need of extra help or
Take a look at our Knowledgebase of Frequently specific expertise to complete your simulation
Asked Questions (FAQs), Tips and Techniques, project, let us help you. PROMODEL Modeling
and other valuable software information. Services will meet all of your needs with fast,
accurate results at competitive rates.
E-mail [email protected] Whether simple or complex, partial or complete,
When you contact technical support via e-mail, PROMODEL Modeling Services can create any
send your user profile and a description of the model you require. With our vast experience in
problem you encountered. producing simulation models for many diverse
applications, we are in a unique position to
Telephone (888) PRO-MODEL evaluate your system and isolate specific
improvements. We work closely with you during
Speak to a technical support engineer and resolve the development process to ensure that the model
the problem over the phone. Our support lines are we create is complete and precise.
open Monday through Friday from 6:00 AM to
6:00 PM MST. With each simulation, PROMODEL Corporation
provides a complete, comprehensive analysis of
After Hours Support (801) 362-8324 your system. We document conclusions and
results derived from the project and present you
In our ongoing effort to serve you better,
with statistics suitable for company presentations
technical support is also available after hours.
and briefings—you may even distribute a limited
Please have ready your user profile and a
version of the model which allows repeated
description of the problem you encountered.
executions and minor revisions.
Fax (801) 226-6046
Send a fax of the listing created when you select
Print Text from the File menu along with your
user profile and a description of the problem you
encountered. For more information on modeling
services, please contact:

Phone (801) 223-4600

Fax (801) 226-6046
4
About the User Guide

0.1.4 Reporting Suggestions

It is our goal to make SimRunner the ultimate
output analysis, decision support, and
optimization tool. To do this, we need your input.
Please feel free to submit comments and ideas on
how we may improve the SimRunner software
and documentation.

Send us your comments

SimRunner Product Team

ProModel Corporation
556 E. Technology Way
Orem, UT 84097
Phone (801) 223-4600
Fax (801) 226-6046
E-mail [email protected]
 SimRunner 5
User Guide

Chapter 1:
SimRunner Concepts

1.1 What is SimRunner? 1.1.1 Benefits

SimRunner is a decision support tool used to help Simulation allows you to find solutions to
optimize simulated processes. SimRunner takes complex problems without incurring the
your existing ProModel, MedModel, or tremendous costs associated with a trial and error
ServiceModel simulation models, evaluates them approach. Optimization helps ensure that the
for you, then performs tests to find better ways to solutions you implement are at or near their
achieve the results you desire. optimal values.

Typically, most people use simulation tools to Unlike manual simulation experimentation which
predict and improve system performance or to answers only specific what-if questions,
establish the relationships between various SimRunner automatically seeks a course of
system elements. By modeling the actual facility action that will help optimize your entire
or process, you can conduct what-if analyses to system’s performance—essentially answering
determine the best way to improve system how to meet your objectives.
performance—this is optimization. Although To help you present your findings, SimRunner
SimRunner cannot guarantee it will identify the allows you to output various types of reports:
optimal solution for all processes every time, it
will find better solutions than you would likely • Data reports (for spreadsheets)
get on your own. • Analysis reports (for text and word pro-
cessing reports)
With each optimization project, SimRunner runs
• Graphical reports (for charts and graphs)
sophisticated optimization algorithms on your
model to help you optimize multiple factors You can print each of these reports or export them
simultaneously. Each project requires a validated to other applications through common clipboard
model, a standard by which to measure system functions.
performance (an objective function), and a group
of factors that SimRunner may change to
improve system performance.
6 Chapter 1: SimRunner Concepts
Where Do I Begin?

1.2 Where Do I Begin? and we examine the model’s output statistics to

see if we achieved the desired outcome. This is
not a bad approach if you have one decision
Start with a validated model variable you are trying to optimize—but what if
Once you complete and validate your simulation you are trying to optimize multiple decision
model, you are ready to begin an optimization variables at once? Interaction becomes very
project. If you are not working with a valid complex and requires more advanced
model, you don’t need to perform an optimization optimization methods like SimRunner.
until the output from the simulation is valid.
Define macros
Identify your simulation type When you build your model, you must define a
It is important to properly identify the simulation macro for any variable you want to optimize.
type—terminating or non-terminating. What does This provides SimRunner with a series of values
it mean to refer to a simulation as terminating or it can change as it seeks to optimize your model.
non-terminating? A terminating system stops In SimRunner, these values are called factors. For
when some key event occurs like the end of the information on how to add and modify macros,
day. When you come back the next day, you start consult your ProModel, MedModel, or
fresh again. A non-terminating system is not ServiceModel User Guide.
necessarily a system that never stops production,
rather it is a system that resumes from the point it Screen factors
left off. For both terminating and non-terminating Part of the simulation process is to evaluate the
simulations, you need to determine the relationships that exist between model elements,
appropriate run length and number of or factors. Often, you will take the time to adjust
replications. For non-terminating simulations it is some part of your model to find that the
also necessary to determine the warm-up period. adjustment has no impact on system
performance. Factor screening is the process of
Determine if model is a true candidate for identifying which model elements (factors) do
optimization not affect the output of your model, and
Not every simulation model is built with the narrowing your search to include only those
express purpose of optimizing some particular factors that affect the model’s output. Be
element. Many simulation models are built to discriminant in your selections.
demonstrate the relationships that exist between
various elements of your system. If optimization Relax and wait
is appropriate, define the objective function—the While some models contain relatively few factors
output statistics used to measure the performance that you can quickly optimize, others contain
of proposed solutions. many. In a previous inventory reduction project, a
high-end computer took approximately 24 hours
Use simulation to identify and examine poten- to compute what it esteemed to be the optimal
tial solutions value for the model. Although it took a long time
Simulation has always been trial and error when to produce this result, the net savings were
it comes to optimization. We have some sort of tremendous.
optimization method that we apply to the model
 SimRunner 7
User Guide

Consider the results 1.2.1 Concepts and Theories

While there is no promise that SimRunner will
SimRunner uses both genetic and evolution
identify the optimal solution to your process, it is
strategies algorithms with its primary algorithm
possible. SimRunner will, however, find better
being evolution strategies. The specific design of
solutions than you would likely get with your
these algorithms for SimRunner is based on the
own trial and error experimentation. The surest
work of Dr. Royce Bowden and other experts.
way to know the optimal solution to any model is
Evolutionary algorithms have been extensively
to run an infinite number of replications of all
evaluated by academics and have proved reliable
possible inputs. Since this is not an option, take
in optimizing simulated systems.
into consideration the number of experiments you
are able to perform and act accordingly.
Evolutionary algorithms
An evolutionary algorithm is a numerical
optimization technique based on simulated
evolution. Solutions generated from the
evolutionary algorithm must adapt to their
environment in order to survive. Since each
potential solution returns a specific result, you
must establish an objective function to measure
the performance of each solution. If the returned
value falls within the acceptable range defined by
the objective function, SimRunner will continue
to examine the value as it searches for the optimal
solution.

Example
To help you better understand how SimRunner’s
optimization process works, consider the
following analogy. If you and a group of
explorers found yourselves on the slopes of a
mountain, in the dark, with nothing but radios
and altimeters, how would you find the summit?

þ À Œ œ
à Õ
Profile view
8 Chapter 1: SimRunner Concepts
Where Do I Begin?

The first step would be to establish the current Again, after everyone moves various distances,
altitude of everyone in your group by recording you stop to take another altimeter reading.
each person’s altimeter reading.

œ
œ Œ
À À
þ Ã
Œ Õ
þ
à Topographical view: Reading 3
Õ
Topographical view: Reading 1
By taking the average of the group’s new
Next, you would direct your group members to altimeter readings, you can see that the overall
wander out in any direction for various distances, reading increased. This confirms that you are
then stop and review their new altimeter moving in the right direction. Just to be sure, you
readings. At the second reading, you find that repeat the process and take another reading.
some of your group has moved higher and some
have moved lower.
œ
œ þ Œ
ÃÀ Õ
À Œ
Õ
þ Ã Topographical view: Reading 4

Topographical view: Reading 2 As you will notice from the example, your group
is beginning to converge upon a single point. The
By comparing the results of each altimeter more you repeat the data collection process, the
reading, you can determine the general direction closer the group will get to one another, and the
in which to proceed. Since you want to climb to more certain you will be that you are at or near
the summit, you proceed in the general direction the summit. Just to be on the safe side, however,
that had the highest reading (in this case, 2). you even send a few members of your group out
to remote parts of the terrain to be sure that there
 SimRunner 9
User Guide

is nothing higher than the point you already value would take a lifetime. More than likely,
identified. you will never get the answer.

ΠWhat you want is an algorithm that can

efficiently explore the response surface (the
þ output from the model) and focus on those areas
œ returning good answers—without evaluating
ÀÕ everything. To learn more about the algorithms
Ã
used in SimRunner, consult the Suggested
Readings on page 41.

Topographical view: Reading 5

Once your average altimeter reading is equal to

the best in your group, you have converged upon
the summit and can be confident that you have
achieved your goal.

ÃŒœ
þÀÕ

Topographical view: Reading 6

Conceptually, this is how SimRunner works—

only the terminology is different. Instead of
altimeters, explorers, and tests, you will use
objective functions, input factors, and
replications. Each time the explorers move,
SimRunner calls this a generation. The concepts
are the same.

Not an exhaustive search

In spite of what some may think, you don’t want
an algorithm that will do an exhaustive search. In
a previous inventory reduction project, modelers
estimated the number of possible solutions to be
around 9.38 x 1037—it was a big project. To
perform an exhaustive search for the optimum
10 Chapter 1: SimRunner Concepts
Where Do I Begin?

1.2.2 General Procedure Step 4: Evaluate suggestions

The following is an overview of the process you The fourth step is to consider and evaluate
will use to perform an optimization of your SimRunner’s suggestions. This is crucial because
system. SimRunner will often identify several candidate
solutions to your problem and you may, for
reasons not addressed in the model, prefer one
Step 1: Create, verify, and validate solution over another. You may also wish to make
The most important preparation you can make for additional model runs (replications) and look at
an optimization project is a validated model. It is confidence intervals to further evaluate
not enough to simply create a model—it will SimRunner’s list of possible solutions.
profit you nothing if the model does not reflect
the real operation. Once you validate the model, Step 5: Apply solution
you are ready to begin.
Once you identify the solution that best fits your
needs, implement the solution.
Step 2: Build a project
With your model prepared for evaluation, create a
new SimRunner project and identify the response
statistic you wish to target. Using these response
statistics, define an objective function by which
to gauge system performance. SimRunner will
use this objective function to measure system
improvement. Next, select the input factors you
will allow SimRunner to use as it determines how
best to achieve system improvement. When you
optimize the model, SimRunner tests each input
factor to seek the combination of factors that will
result in the greatest improvement of model
performance.

Step 3: Run experiments

Once you select the input factors and define the
objective function, you can use SimRunner to
automatically conduct a series of experiments on
your model. SimRunner runs your model for you
and tests a variety of possible combinations of
values. After it completes the tests, SimRunner
lists the test results in order of the most to the
least successful combination of factor values.
 SimRunner 11
User Guide

1.2.3 Pitfalls
If you follow the general procedure stated
previously, your chances of success are very
good. Typically, projects fail because the:
• Model is not valid
• Analysis considers insignificant factors
• Analysis ignores significant factors
• Objective Function is inappropriately for-
mulated
• Test results are not scrutinized
12 Chapter 1: SimRunner Concepts
Where Do I Begin?
 SimRunner 13
User Guide

Chapter 2:
Getting Started

2.1 The Interface process is divided into three phases, each

containing a series of steps necessary to complete
The SimRunner interface provides you with easy the phase. As you move from phase to phase
access to every step necessary to create an (displayed at the top of the dialog), a list of steps
optimization project. The project building for the phase appears in the left pane.

Phase 1, 2, 3

Steps for
current phase
14 Chapter 2: Getting Started
Set Up a Project

2.2 Set Up a Project 2.2.2 Define Objectives

This section provides detailed information about The objective of your project is the final outcome
each step required to build an optimization you want to achieve. SimRunner measures your
project. progress toward this goal using an objective
function. An objective function is composed of
response statistics, a min/max or target range, and
a specific weight you wish to apply to each
response statistic.

2.2.1 Select Model/Project

Select the model you will use for your project.

Response category
The response category is the type of statistic you
wish to use to evaluate your model. Response
Create new project / Select model Enter the name
categories include model elements such as
of the model you wish to optimize. You will enter
locations, entities, resources, and variables.
a name for the project when you save it.
Open existing project / Select project Select an
existing project. If you open an existing project, it
is not necessary to select the model you will
use—the model is stored as part of the
optimization project.
 SimRunner 15
User Guide

Response statistics
Simply put, response statistics are those values
you wish to improve. Once you define your
targeted improvements, you are ready to define Please note
how you want them to perform—the objective for
the response statistic.
Maximized objective functions return positive val-
ues. Minimized objective functions return negative
values.

Target Range Check this option to enter a

specific target range within which you want the
final result.
Statistic’s weight Weights serve as a means of
Objective for response statistic load balancing for statistics that might bias the
objective function. Since most simulation models
The objective for the response statistic refers to
produce a variety of large and small values, it is
the way in which you want to effect change for
often necessary to weight these values to ensure
that item. If you are trying to increase the overall
that the objective function does not
output of a system, you would maximize the
unintentionally favor any particular statistic. For
response statistic. Likewise, you could minimize
example, suppose that a run of your model
the statistic or target a specific range within
returns a throughput of .72 and an average WIP
which you want the result. Finally, enter the
of 4.56. If you maximize throughput and
reward or weight for each response statistic—
minimize WIP by applying the same weight to
larger numbers signify greater rewards in
both (W1=W2), you will bias the objective
situations that require more than one objective.
function in favor of WIP:
Maximize[(W1)*(Throughput)] = .72
Minimize[(W2)*(WIP)] = 4.56
Max Check this option if you want to increase
the final value of this statistic. .72 4.56
Min Check this option if you want to decrease
the final value of this statistic.

In this case, since you want to ensure that both

statistics carry equal weight in the objective
function, you will apply a weight of 6.33
(W1=6.33) to throughput and 1.0 (W2=1.0) to
16 Chapter 2: Getting Started
Set Up a Project

WIP to make them of equal weight in the combination and displays the updated objective
objective function. function for the project.
Maximize[(W1)*(Throughput)] = 4.56
Minimize[(W2)*(WIP)] = 4.56

4.56

Objective function
.72
The objective function is an expression used to
quantitatively evaluate a simulation model’s
performance. By measuring various performance
characteristics and taking into consideration how
In situations where it is necessary to favor one
you weigh them, SimRunner can measure how
statistic over another, balancing the statistics first
well your system operates. However, SimRunner
will make it easier to control the amount of bias
knows only what you tell it via the objective
you apply. For example, if you apply a weight of
function. For instance, if your objective function
12.67 (W1=12.67) to throughput and 1.0
measures only one variable, Total_Throughput,
(W2=1.0) to WIP, the objective function will
SimRunner will attempt to optimize that variable.
consider throughput to be twice as important as If you do not include an objective function term
WIP (adapted from Harrell, Ghosh, and Bowden to tell it that you also want to minimize the total
2000). number of operators used, SimRunner will
assume that you don't care how many operators
you use. Since the objective function can include
multiple terms, be sure to include all of the
response statistics about which you are
concerned.
Please note
SimRunner’s capacity to include many different
Typically, you will need to experiment with your response statistics in an objective function gives
model to identify the weight ratio necessary to it tremendous capability. For example, the
balance statistics. objective function below signifies that you wish
to maximize the total ovens and cooktops
processed while minimizing the total resource
cost. The numeric weighting factors indicate that
maximizing Total Ovens Processed is the most
Response statistics selected for
important, followed by maximizing Total
objective function Cooktops Processed, then minimizing Total
After you define the objective for the response Resource Cost.
statistic, you may click the Add button to include Z =Max:10 * (Total Ovens Processed) +
the statistic as part of the objective function. Max:5 * (Total Cooktops Processed) +
SimRunner combines the statistics into a linear Min:2 * (Total Resource Cost)
 SimRunner 17
User Guide

Perhaps the best way to define your objective 2.2.3 Define Inputs
function is in terms of cost or profit. When
possible, this allows you to use a simple, single In every system, there are controllable and
response statistic like maximize profit without uncontrollable factors that determine the
concern over assigning meaningful weights to outcome of the process. Controllable factors
multiple response statistics. include staffing, equipment, schedules, and
facilities. Uncontrollable factors refer to such
things as arrival rates.
Often, relationships exist between various
controllable factors. How do you identify these
relationships and use them to seek the best value
Example: Maximize up to a target for your objective function?—SimRunner.
SimRunner allows you to target controllable
The following example shows how to achieve a model factors and determine which combination
throughput target level. Suppose you want to target a of values for those factors will elicit the behavior
production rate of 300 to 325 units per day. In the tar- you desire.
get range fields, assign a range of 300 to 325 and enter
a weight of 1. With each test it performs, SimRunner examines
the results to see if the results will produce the
effect you require from the objective function.
For information about how SimRunner performs
Log time these tests and evaluates data, see Concepts and
Because the LOG statement does not have a Theories on page 7.
separate data structure (it occurs during the
processing logic), SimRunner does not have
direct access to log data. However, if you create a
variable to represent the logged time, you can use
the variable as part of the objective function.

Macros listed in model

All macros defined in your model are displayed.
SimRunner will use these macros to improve the
18 Chapter 2: Getting Started
Set Up a Project

value of your objective function. Typically, these few inputs as possible—only those you anticipate
are the controllable factors of your model. will affect the value of the objective function.

Macro properties Define macros

Macro properties describe the basic attributes of Before you can use a model element as an input
each macro used in your project. factor, you must define a macro for that element.
Macros allow SimRunner to control the value of
the factor and feed it back into the model.

Variables
Data type Refers to the numeric type of the data If you want to use a variable from your model as
(integer or real) SimRunner will use. Typically, an input factor, you must define a corresponding
you will use integers to represent the number of macro and set the variable equal to the macro at
resources (e.g., people) and real numbers to some point in your model. To do this, you can:
represent time values or percentages (e.g.,
• Enter the name of the macro in the vari-
machine processing times or product mix).
able’s “initial value” field.
Default value Refers to SimRunner’s initial • Set the variable equal to the macro in the
setting for the macro—the value SimRunner will model’s processing logic.
use to analyze the model before conducting the • Set the variable equal to the macro in the
optimization. See Analyze Model on page 22. model’s initialization logic.
Lower and upper bound The limits (constraints)
within which this value must fall during Resource & location capacities
subsequent tests. If you defined these bounds in In order to use a resource or location capacity as
an RTI for the macro, you may override them an input factor, you will need to define a macro
here if you wish. that represents the capacity or number of
resources and place this macro in the appropriate
Macros selected as input factors field for the resource or location. For instance, if
you have a resource named Operator_1, you can
Displays a list of all macros you selected for use create a macro named Number_Of_Operators to
as input factors. Typically, you will want to use as represent the “number of units” for the operator.
When you run SimRunner, the macro
Number_Of_Operators will appear on the list of
input factors. Testing this factor will change the
number of units of Operator_1 in the model.
 SimRunner 19
User Guide

2.2.4 Set Options

Please note

SimRunner will not recognize any macro contain-

ing non-numeric characters (e.g., “N”, “)”, or
“e+”) as a valid input from the macro’s "text"
field.

Distribution parameters
Since macros must be completely numeric in the
"text" field for SimRunner’s use, macros that
represent distributions will not appear in the list
of available input factors. A macro can be used to
represent a distribution parameter. For example, Optimization options
suppose you have a normal distribution, N(10,2),
with an average of 10 and a standard deviation of
2. You can create a macro, Dist_1_Average, with
a value of 10 and use this macro as the text value
of the distribution: N(Dist_1_Average, 2). This
will allow you to manipulate the distribution
from SimRunner. Optimization profile SimRunner provides three
optimization profiles: Aggressive, Moderate, and
Cautious. The optimization profile is a reflection
of the number of possible solutions SimRunner
will examine. For example, the cautious profile
tells SimRunner to consider the highest number
Please note of possible solutions—to cautiously and more
thoroughly conduct its search for the optimum.
Using ProModel, MedModel, or ServiceModel, As you move from aggressive to cautious, you
you can set up macros in your simulation model will most often get better results because
as run-time interfaces (RTI’s). SimRunner reads SimRunner examines more results—but not
the RTI’s as the default values for the input fac- always. Depending on the difficulty of the
tor’s lower and upper bounds. If you do not problem you are trying to solve, you may get
define RTI’s, SimRunner will use the factor’s cur- solutions that are equally good. If you are pressed
rent value as the lower and upper bounds. for time or use relatively few input factors, a
more aggressive approach—the aggressive
profile—may be appropriate.
20 Chapter 2: Getting Started
Set Up a Project

3. Min No. of generations The fewest iterations Sim-

Runner will use to conduct the analysis—this impacts
the maximum time allocated to the optimization
search. Typically, you will set the minimum number of
Please note generations to 1; however, if you specify a very large
value, you can force SimRunner to continue to seek the
optimal solution even after it satisfies the convergence
The optimization profile affects the number of percentage.
solutions SimRunner will evaluate before it con-
verges. An aggressive profile generally con-
verges quickly, but will have a lower probability
of finding the optimum. Simulation options

Convergence percentage With every experiment,

SimRunner tracks the objective function’s value
for each solution. By recording the best and the
average results produced, SimRunner is able to
monitor the progress of the algorithm. Once the
best and the average are at or near the same Disable animation Uncheck to activate the
value, the results converge and the optimization animation during run time. If you enable the
stops. The convergence percentage controls how animation, it will take longer to conduct the
close the best and the average must be to each analysis but will not affect the model’s results.
other before the optimization stops. A high
percentage value will stop the search early, while Number of replications per experiment The
a very small percentage value will run the number of times the model will run in order to
optimization until the points converge. conduct the experiment—to estimate the value of
the objective function for a solution. (Typically,
you will never conduct an optimization with only
one replication.)
Warm-up time The amount of time the model
must run before it reaches steady-state.
Advanced Options
Run time The total run length of the model.
1. Options: the following options are available if
Confidence level If you specify more than one
selected on the Advanced Options menu. Select
Options | Advanced... to enable Max and Min genera- replication, SimRunner will compute and display
tion selection. a confidence interval based on the mean value of
the objective function and its sample variance.
2. Max No. of generations The most iterations Sim-
The confidence level allows you to specify a
Runner will use to conduct the analysis. This impacts
90%, 95%, or 99% confidence interval. Although
the maximum time allocated to the optimization
search. Very high values allow SimRunner to run until you cannot determine the true mean value of the
it satisfies the convergence percentage. objective function for a given experiment (you
can only estimate it), the confidence interval
 SimRunner 21
User Guide

provides a range that hopefully includes the true

mean. The higher the value you select for the
confidence level, the more confident you can be
that the interval contains the true mean. However,
note that a 99% confidence interval is generally
wider that a 90% confidence interval given a
fixed number of replications.

Please note

Typically you want SimRunner to run the algo-

rithm until it fully converges. Do this by specify-
ing a very low value for convergence percentage,
such as 0.01, and, if running in advance mode, a
very high value for maximum number of genera-
tions, such as 99999, and a value of 1 for mini-
mum number of generations.
22 Chapter 2: Getting Started
Analyze Model

2.3 Analyze Model does not reflect reality. When you run a non-
terminating simulation, you must run it long
The first step in conducting any analysis of your enough for the operation to stabilize—to reach
simulation model’s output is to make sure that steady state. The warm-up time is how long it
your model produces output with the degree of takes the simulation to reach steady state. The
accuracy you desire. If all of the aspects of your end of the warm-up time is the point at which you
model are deterministic (with no random want to begin sampling statistical data—data
variance in any part of the model), you will get a representative of how the system normally
precise output statistic with one replication. In operates.
most models, however, random variance is
present in several places. The result of using
random inputs in the model is that the output is
also random. Because of this, you should think of
a simulation experiment as a sampling technique.
Since you don’t know the exact, true value or Please note
population statistic, the best you can do is collect
a sample large enough to make a good estimate— When you define a warm-up period, the model
you must run multiple replications. runs for a specific period of time at the beginning
In general, with each additional replication you of each replication before results are recorded.
run, you generate a more accurate estimate of the
population statistic your are trying to measure in
the objective function. For example, suppose you
wish to maximize the utilization of a given How do you know how long the warm-up period
resource by changing several factors in the should be? SimRunner uses an approach based
model. For each experiment, you may need to run on Welch’s graphical method. SimRunner
ten replications to get an accurate estimate of the displays a time series graph of the objective
utilization for a specific combination of settings. function’s current value and moving average
If your estimate isn’t accurate enough, you will based on several replications of the simulation.
not be able to tell if improvements to a system are As the objective function begins to stabilize, the
the result of changes or simply an effect of moving average graph will appear to "flatten out"
random variance. SimRunner helps you select the around a specific value or range. This behavior
appropriate number of replications for both signals the end of the warm-up period.
terminating and non-terminating systems. It is quite possible that you may be unable to
For non-terminating systems, SimRunner’s determine the warm-up period from just one look
Objective Function Time Series graph shows the at the graph. If you are unable to make this
value of the objective function as it changed determination, run the model again with a longer
during the simulation based on several run length.
replications. At the very beginning of a
simulation run, the simulation model is “empty.”
There are no entities in the system and often all
the statistics are set to zero. This is an artificial
condition and, for non-terminating simulations,
 SimRunner 23
User Guide

Define experiment Output recording time interval (period)

SimRunner records output statistics during
regular intervals of time. This sampling
technique provides you with an observation of
the output statistic for each time interval. The
output recording time interval refers to the length
of each time interval. The recording time interval
should be large enough to ensure that simulated
events occur, and that the model produces an
output value for each response statistic listed as
part of the objective function. For example, given
a job arrives to a queue every 30 minutes, you do
not want to select a recording time interval of less
than 30 minutes if the average time jobs wait in
the queue is part of the objective function. Be
sure to set the recording time interval to observe
at least one observation during each interval.
In the case of a non-terminating system, the Number of test replications The number of
warm-up period is the amount of time the model replications needed to conduct the analysis. The
must run for the objective function to exhibit number of test replications should be five or more
statistical regularity—to reach a steady-state. depending on how long you have to wait for
This occurs when the distribution of the objective results.
function values is the same from one time period
to the next. SimRunner helps you identify the Percent error in objective function estimate Given
length of the warm-up period. that you can only estimate the true value of the
objective function, you have to decide how
Experimental parameters accurate your estimate needs to be. Entering a 10
indicates that you wish to estimate the required
number of replications needed to obtain an
estimate of the average value of the objective
function to within 10% of its true value.
Confidence level The confidence level (90%,
95%, or 99%) used to approximate the number of
replications needed to estimate the value of the
objective function to within the stated "percent
Simulation run length The total duration of the error" of its true value. The higher the confidence
simulation. If you run a non-terminating level, the larger the number of replications
simulation, the run length should be sufficient to needed will be.
allow the model to warm up and reach steady
state. Start with a large initial value for the run
length.
24 Chapter 2: Getting Started
Analyze Model

Conduct analysis Estimate required number of replications

Warm-up periods Enter the number of time

periods you wish to use as the warm-up period
when estimating the required number of
replications. Periods refer to the frequency with
which you recorded data from the model. For
example, if your output recording time interval is
four hours, you may find that five periods will
suffice (20 hours).
No. of replications This field is automatically
updated as the warm-up periods field is defined.

Start analysis

Run Begin analyzing the model.

Final Report Produces an analysis report which

describes the results of each experiment.
Analysis status Displays the completion status of
the analysis.

Warm-up detection for steady-state estimates

Moving average window Allows you to adjust the
number of periods used to compute the moving
average plot. The plot is automatically redrawn.
 SimRunner 25
User Guide

2.4 Optimize Model

2.4.1 Introduction
Example
Often, the reason for building a simulation model
of a system is to answer questions such as: “What The following diagram shows the relationship between
are the optimal settings for to minimize an optimization algorithm and a simulation model
(or maximize) ?” You can think of the (adapted from Harrell, Ghosh, and Bowden 2000).
simulation model as a black box that imitates the
actual system. When you present inputs to the
black box, the box produces outputs that estimate Input: (x1, x2, x3, ...)
Optimization Simulation
how the actual system will respond. In the Algorithm Model
question above, the first blank represents the
input(s) to the simulation model that you control.
(These inputs are often called decision variables
or factors.) The second blank represents the Output Response: (performance measure =y)
performance measure(s) of interest, computed
from the output response of the simulation model
when you set the factors to specific values (see If you were to evaluate all combinations of the differ-
ent values for the input factors (X1, X2, X3, …) and
example below).
plot the output response generated by the simulation
In the question, “What is the optimal number of model, you would create what is called a response sur-
material handling devices needed to minimize the face of the output for that model. Think of the response
time that workstations wait for material?”, the surface as a mountainous region with many peaks and
input factor is the number of material handling valleys. In the case of a maximization problem, you
devices and the performance measure is the time wish to climb to the highest peak. In the case of a min-
imization problem, you wish to descend to the lowest
that workstations wait (computed from the output
valley.
response of the simulation model). The goal is to
locate the optimal value for each factor that
minimizes or maximizes the performance
measure of interest.
26 Chapter 2: Getting Started
Optimize Model

2.4.2 Optimization Concepts Limit the number of input factors

Don't include input factors that have relatively
To optimize functions or the output from
little impact on model results. Although
simulation models, you must apply some kind of
SimRunner can handle multiple factors, each
method to evaluate different combinations of
factor you add increases the amount of time
input factors until you arrive at what you consider
necessary to optimize the model. So, the fewer
the optimal solution. If you take several different
the number of factors, the faster you will get
input/output combinations (i.e., function
results.
evaluations), you should be able to establish
which variables cause improvement in the output
Set good, tight bounds
response and try a few more function evaluations.
By ascending the “peak” (in the case of a It takes a lot longer to find an answer between 1
maximizing problem), you will work your way to and 1000 than between 1 and 10. The tighter the
the top. bounds you specify for your input factors, the
faster you will get results.
Once you climb to a peak (i.e., reach a local
optimum), how do you determine whether your
Formulate a good objective function
local optimum is the global optimum—the very
best? How do you know that there is no better Since SimRunner uses the objective function to
combination of inputs? The surest way would be evaluate a solution’s performance, it is absolutely
to try every possible combination of input necessary that you include the right output
factors. This is not feasible for most problems. responses in the objective function. This allows
What you can do, however, is try to find and SimRunner to test several combinations of
compare several local optimums. This will factors to see which arrangement will improve
greatly increase your chance of finding the best, the desired aspect of the system’s performance.
or global, solution. Remember also that your answers are only as
precise as the system's randomness allows. If
In many situations, it is impossible to formulate a your model gives results that are accurate only to
single equation that precisely describes a
system’s behavior. For this reason, techniques
±5, SimRunner’s solutions will be equally
such as simulation are used to model and study accurate (±5).
these systems. Although simulation itself does
not optimize, it acts as a function to tell you
(based on your model of the system) what results
you will achieve from a given set of input factors.
SimRunner conducts a variety of tests to seek
ideal operation levels for your model.
When setting up your optimization, you can do
several things to greatly improve SimRunner's
performance:
 SimRunner 27
User Guide

2.4.3 SimRunner Optimization Run Begin optimizing the model.

Techniques Stop Halt the optimization.

SimRunner intelligently and reliably seeks the Performance plot Plots the best objective
optimal solution to your problem based on the function value found throughout the optimization
feedback from your simulation model by process.
applying some of the most advanced search Final report The final report contains several of
techniques available today. The SimRunner the best solutions found. SimRunner sorts these
optimization method is based upon Evolutionary solutions to allow you to evaluate them further
Algorithms (Goldberg 1989, Fogel 1992, and before you make your final decision. These
Schwefel 1981). results can be saved to a file using the Export
Evolutionary Algorithms are a class of direct Optimization Data option found in the File menu.
search techniques based on concepts from the
theory of evolution. The algorithms mimic the Convergence status
underlying evolutionary process in that entities
adapt to their environment in order to survive.
Evolutionary Algorithms manipulate a
population of solutions to a problem in such a
way that poor solutions fade away and good
solutions continually evolve in their search for Phase 1 and Phase 2 SimRunner sometimes uses a
the optimum. Search techniques based on this genetic algorithm before using the evolution
concept have proven to be very robust and have strategies algorithm based on the characteristics
solved a wide variety of difficult problems. They of the problem being solved. Phase 1 displays the
are extremely useful because they provide you convergence status of the genetic algorithm.
with not only a single, optimized solution, but Phase 2 displays the convergence status of the
with many good alternatives. evolution strategies algorithm.
Generation The current generation. Each
Seek optimum generation represents a collection of experiments
performed on the model.
Experiment The current simulation experiment.
28 Chapter 2: Getting Started
Optimize Model

Response plot Edit chart Accesses the controls you will need to
change the appearance of the Surface Response
Plot.

Independent variable 1 & 2 Select the input

factors you wish to use to produce a partial
response surface of the model’s output.
Update chart Updates the chart to reflect the data
you selected.
 SimRunner 29
User Guide

2.5 Menus 2.5.2 Options menu

2.5.1 File menu

Clear Optimization Data Clears Optimization

Data from the data grid.
Clear Analysis Data Clears Analysis Data
from the data grid.
Advanced... Provides an option to control
Minimum and Maximum Generations. This
effects the Setup Project | Set Options screen,
New Starts a new project. providing additional optimization control to the
user.
Open Model Opens an existing simulation model.

Open Project Opens an existing optimization

project.
2.5.3 Help menu
Save Saves the current optimization project to its
original path and filename.
Save As... Allows you to save the project under a
different name or in a different location.
Help Topics Provides access to the help
Export Optimization Data Exports the data found system.
in the optimization grid to a comma-delimited
file of your choice. About SimRunner Provides information
about SimRunner.
Recently opened files Displays a list of all
recently opened files.
Exit Quits SimRunner.
30 Chapter 2: Getting Started
Menus
 SimRunner 31
User Guide

Chapter 3:
Building Projects

3.1 Getting Started

In this chapter, you will create a new project from
a validated model that contains macros and
examine how the model might perform more
optimally.
32 Chapter 3: Building Projects
Set Up Project

3.2 Set Up Project

Step 2: Define objectives

1. Select the response category and statistic for each

Step 1: Select model item you wish to monitor.

1. Enter the name of the model you will use in the

Create new project field or browse to select the
model.

2. Click Add (the down arrow button). The statistic

will appear in the selected response statistics window.
You can remove portions of the function by selecting
the up arrow button.

2. Click open to select a model.

3. Click Next to continue.
 SimRunner 33
User Guide

3. Enter the objective for the response statistic (Max,

Min, Target range) and the weight to apply to the statis-
tic. Select the Update button to apply the changes to
the objective function
Step 3: Define inputs

1. Select a macro you will use in the project from the

macros list.
2. Click the down arrow button to add the macro to the
input factors list.

4. Repeat this process to add any other response statis-

tics. If you wish to remove a statistic, select the statis-
tic to remove, then click the up arrow button. Once you
have selected all the statistics you will use, click Next.
34 Chapter 3: Building Projects
Set Up Project

3. Select the numeric type, then enter the default

value, the lower bound, and the upper bound. Select
Update to reflect the changes in the macro selected.

Step 4: Set options

1. Select the optimization profile and enter the conver-

gence percentage you wish to use.

4. Repeat this process to add any other response statis-

tics. If you need to change the properties for a macro
that has already been added, select the macro to
change, make the macro property change (Integer,
Real, value and bounds), then click Update. If you
wish to remove a statistic, select the statistic to remove,
then click the up arrow button. Once you have selected 2. Select the confidence level.
all the statistics you want use, click Next.
 SimRunner 35
User Guide

3. Enable or disable the animation as necessary and

specify the number of replications per experiment you
wish to run.

4. Enter the warm-up time and run time for the model.

5. Click Apply, then Next to continue.

36 Chapter 3: Building Projects
Analyze Model

3.3 Analyze Model 2. Enter the number of test replications

Step 1: Define experiment parameters

1. Enter the simulation run length and the output

recording time interval (period).

3. Enter the percent of error in the objective function

estimate and select a confidence level.

4. Click Apply, then Next to continue.

 SimRunner 37
User Guide

4. Click the Final report button to view the analysis

report.

Step 2: Conduct analysis

1. Click Run to begin the analysis

2. If you wish to adjust the number of periods used to

compute the moving average (for steady state esti-
mates), adjust the value for the Moving Average Win-
dow. The moving average plot will update to match the
current settings.

3. Enter the warm-up time. SimRunner will suggest

the number of replications you should run for the opti-
mization project.
38 Chapter 3: Building Projects
Optimize Model

3.4 Optimize Model 3. Click Final Report to display a written summary of

the optimization.

Step 1: Seek optimum

1. Click Run to start the optimization.

Step 2: Response plot

1. Enter the independent variables (macro input fac-

tors) you wish to use in the response plot and click
Update chart.
2. Click Performance Plot to view the objective func-
tion performance throughout the optimization.
 SimRunner 39
User Guide

2. If you wish to modify the appearance of the graph,

click Edit Chart. The following dialog will appear.

3. After you modify the appearance of the graph, click

OK.
40 Chapter 3: Building Projects
Optimize Model
 SimRunner 41
User Guide

Appendix

Suggested Readings Optimization

Harrell, C. R.; B. Ghosh; and R. O. Bowden.
ProModel Corporation recommends the follow-
2000. Simulation Using ProModel. New
ing works to help you obtain a better knowledge
York: McGraw-Hill.
of simulation and optimization.
Bowden, R. O. and J. D. Hall. 1998. Simulation
Simulation Optimization Research and Development.
Banks, Jerry, John S. Carson, II., Berry L. Nel- Proceedings of the 1998 Winter Simulation
son, and David M. Nichol. 2000. Discrete Conference 1693-1698.
Event System Simulation. N.p.: Prentice-Hall
Inc. Hall, J. D. and R. O. Bowden. 1997. Simulation
Optimization by Direct Search: A Compara-
Harrell, C. R.; B. Ghosh; and R. O. Bowden. tive Study. Sixth International Industrial
2000. Simulation Using ProModel. New Engineering Research Conference 298-303.
York: McGraw-Hill.
Hall, J. D.; R. O. Bowden; and J. M. Usher. 1996.
Law, A. M. and W. D. Kelton. 2000. Simulation Using Evolution Strategies and Simulation to
Modeling and Analysis. New York: Optimize a Pull Production System. Journal
McGraw-Hill. of Materials Processing Technology 61:47-
52.
Bateman, R. E.; R. G. Bowden; T. J. Gogg; C. R.
Harrell; J. R. A. Mott. 1997. System Goldberg, D. 1989. Genetic Algorithms in
Improvement Using Simulation. Orem, Utah: Search, Optimization, and Machine Learn-
PROMODEL Corporation. ing. Addison Wesley, Massachusetts.
Harrell, Charles R. and Kerim Tumay. 1995. Sim- Schwefel, H. P. 1981. Numerical Optimization of
ulation Made Easy. N.p.:Industrial Engineer- Computer Models. Chichester: John Wiley
ing Press. and Son.
42
 SimRunner 43
User Guide

Glossary

Analyze Model Population best

Analyze model runs tests to help you identify the

initial bias (warm-up period) for non-terminating

Fitness [F(x)]
simulations and appropriate number of
Population
replications. converged
Best – Average
CvgPercentage ≥ ------------------------------------------
Best
Confidence Level Population average

Used for approximating the required number of

replications and for computing confidence Generations
intervals. In the case of confidence intervals, it
specifies the proportion of confidence intervals
that contain the true mean.
Data type
Convergence Percentage The numeric type (integer or real) of the input
factor you will use. Typically, you will use inte-
As SimRunner moves through the optimization gers to represent resources (e.g., people) and real
process, it computes statistical data for the solu- numbers to represent time values or percentages
tions it evaluates. From each generation, Sim- (e.g., machine processing times or product mix).
Runner records the best solution and the average
of all solutions. When the population best and
Generation
average are within the specified convergence per-
centage, SimRunner halts the optimization and
A complete cycle of evaluating a population of
the optimization is said to have "converged."
"parent" solutions and selecting the best parents
to produce a population of "offspring" solutions
for the next generation.

Independent variable
See Input factors.
44

Input factors cessed and resource utilization you could mea-

sure how “well” a certain simulation scenario ran
Those factors for which SimRunner will seek by measuring Z, where:
optimal values. When you define input factors,
you must provide the lower and upper bounds. Z = (Total Processed) + (Resource Utilization)
An example of an input factor is the number of
operations used. Optimization
Testing various what-if scenarios to determine the
Lower bound best way to conduct operations.
The lowest numeric value of the input factor. The
minimum value an input factor may have for Optimization Profile
optimization experimentation purposes.
The optimization profile controls the size of the
population of solutions SimRunner uses to seek
Macros the optimum. Aggressive is the smallest popula-
A macro is a placeholder for an often-used tion, Moderate is a larger population size, and
expression used in an expression or logic field. Cautious is the largest population.
You can type a macro once, then substitute the
macro’s name anywhere in the model to use its Output Factor
value (as often as necessary).
See Objective function.

Moving Average
Output Recording Time Interval
When the model’s output response is erratic, it is
The length of the individual time periods for
useful to "smooth" it with a moving average. A
which SimRunner computes statistics, as used
moving average is constructed by calculating the
within the Analyze Model step.
arithmetic average of the n most recent data
points in the data set. The value of n is called the
moving average window. As the value of n Pre-analysis
increases, the "smoothness" of the moving aver- See Analyze Model.
age plot also does.
Project
Objective function
When you conduct an analysis in SimRunner,
An expression used to quantitatively evaluate a you create a project containing the model and the
simulation model’s performance. By measuring results from any tests you run.
various performance characteristics and weight-
ing them, an objective function is a single mea-
sure of how well a system performs. Run length
The time length for which SimRunner will run
SimRunner allows you to include many different
your model for each experiment.
performance characteristics in one objective
function. For example, if you want an objective
function to include a measure of total entities pro-
 SimRunner 45
User Guide

Simulation
Simulation is the act of creating a model to repre-
sent an actual location (e.g., plant floor, bank
lobby, or emergency room) or abstract, logical
process.

Steady-state
The point at which the output of the model exhib-
its statistical regularity—the distribution of the
output is the same from one time period to the
next.

Upper bound
The highest numeric value of the input factor. In
other words, “no more than this.”

Warm-up Time
The amount of time the model must run for the
objective function to reach a steady state in the
context of non-terminating simulations.

Weighting
This value signifies the importance of maximiz-
ing or minimizing the variable. For example,
while it may be important to maximize the pro-
duction level, it may be more important to main-
tain current personnel levels.
46
 SimRunner 47
User Guide

Bibliography

References
This manual referenced the following works. The
list below will allow you to find these works eas-
ily as you seek to increase your knowledge of
simulation and optimization.

Simulation
Harrell, C. R.; B. Ghosh; and R. O. Bowden.
2000. Simulation Using ProModel.McGraw-
Hill, Massachusetts.

PROMODEL Corporation. 2000. ProModel,

MedModel, ServiceModel User Guides.
Orem, Utah: PROMODEL Corporation.

Optimization
Fogel, D. 1992. Evolving Artificial Intelligence.
Ph.D. Thesis, University of California, CA.

Goldberg, D. 1989. Genetic Algorithms in

Search, Optimization, and Machine Learn-
ing. Addison Wesley, Massachusetts.

Schwefel, H. P. 1981. Numerical Optimization of

Computer Models. Chichester: John Wiley
and Son.
48
 SimRunner 49
User Guide

Index

A F
About SimRunner 29 Factors
Activity capacities 18 input 44
Animation output 44
disable 20 File
menu 29

B
G
Bound
lower 44 General procedure 10
set 26 Getting to know SimRunner 13
upper 45

H
C
Help
Concepts and theories 7 menu 29
Help topics 29

D
I
Data
type 18, 43 Independent variables 43
Define Input factors 44
macros 18 limit 26
Disable animation 20 Interval width 44
Distribution parameters 19 Introduction
what is simrunner? 5
E
K
Exit 29
Export Optimization Data 27, 29 Keyboard 2
50

L P
Limit Pitfalls 11
input factors 26 Pre-analysis 44
Location capacities 18 Print 29
Log time 17 Product support 2
Lower bound 44 Project 44
new 29
open 29
M save 29
ProModel
product team 2
Macros 44
define 18
Maximize
up to a target R
example 17
MedModel Resource capacities 18
product team 2 RTI 19
Menus Run length 44
file menu 29 Run-time interface 19
help menu 29
options menu 29
Modeling S
services 3
Moving Average 44
Save
as 29
project 29
N ServiceModel
product team 2
New project 29 Simulation 45
Number of replications Statistical Advantage 43
test 23 Steady-state 45
Support, technical 2
Symbols and notation
O keyboard 2
text 2
Objective function 16, 44
formulate 26
Open project 29 T
Optimization 5, 44
concepts 26 Target
techniques 27 maximize up to 17
Options Technical support 2
menu 29 Text 2
Output Theories and concepts 7
factor 44 Time
log 17
 SimRunner 51
User Guide

U
Upper bound 45

V
Variables 18
independent 43

W
Weighting 45
Where do I begin? 6
general procedure 10
pitfalls 11
52