4.

11 SIMULATION
Simulation isthe appearance of
reality. In MSS,
experiments (e.g., what-if simulation is a technique
for conducting
analyses) with a computer on a
Typically, real model of a management system.
decision-making situations involve some
deals with randomness. Because DSS
semistructured or unstructured
situations, reality is
easily represented by complex,which may not be
optimization or other models but
can often be handled by simula
tion. Simulation is
one of the most commonly used DSS
and Seilaa et al. (2003). methods. See Application Case
4.8

APPLICATION CASE 4.8

Improving Maintenance Decision Making in the Finnish Air Force Through Simulation

The Finnish Air Force wanted togain

efficiency in probabilities was available). They used several meth
its maintenancesystem in order to keep as
many air- ods
ods to acquire
acquire and secure data, such as asking
craft aspossible safely available at all
times for train- experts in aircraft maintenance fields at different
ing,missions, and other tasks, as needed. A discrete levels for their opinions and designing a model that
event simulation program similar to those used in allowed the confidential data to be input into the
Imanufacturing was developed to accommodate system. Also, the simulations were compared to
Worktorce issues, task times, material handling actual performance data to make sure the simulated
delays,and the likelihood of equipment failure. results were aCcurate.

The developers had to consider aircraft avail The maintenance program was broken into
ability,resource three levels:
requirements for international oper-
alions, and the periodic maintenance program: The

information for normal conditions and conflict 1. The organizational level, in which the fighter
Conditions was input into the simulation program squadron takes care of preflight checks, turn
because the maintenance schedule could be altered around checks (which occur when an aircraft
Irom onesituation to another. returns), and other minor repairs at the main

The developers had to estinate some informa- command airbase in normal conditions
tion due to 2. The intermediate level, in which more compli
confidentiality, especially with regards
to Conflict cated periodic maintenance and failure repairs
narios (no data of battle-damage
172 Part Il • Computerized lecision Support

of the
taken care of at the air command repair commands,
air
are system state: characteristiCS
main-

shop at the main airbase in normal conditions tenance needs, and flight operations; accumulated Aio

3. The depot-level, in which all major periodic hours; and the location of each aircraft. Excel spre

maintenance is taken care of and is located sheets were used for data input and output. Additionall

away from the main airbase parameters of some of the input data were
estimated
from statistical data or based on information from auk
During conflict conditions, the system is decen for ti
mater experts. These inchuded probabilities
tralized from the main airbase, The maintenance levels jeCt
between failures, damage sustained during a sinole
just described may continue to do the exact same
flight mission, the duration of each type of periodi
repairs, or periodic maintenance may be eliminated.
maintenance, failure repair, damage repair, the times
Additionally, depending on need, supplies, and capa between flight and the duration of a
missions, mission
bilities, any of the above levels may take care of any
This simulation model was sO successful that the Finnich
maintenanceand repairs needed at any time during
conflict conditions.
Any, in collaboration with the Finnish Air Force. has

now devised a simulation model for the maintenance

The
simulation model was implemented using
for some of its new transport helicopters.
Arena software based on the SIMAN language and
Source: Based on V. Matila, K. Virtanen, and T. Raivio,
invoked using a graplhical user interface (GUD that was "ImprovingMaintenance Decision Making in the Finnish Air Force

executed using Visual Basic for Applications (VBA).The Though Simulatíon," Interfaces, Vol. 38, No. May/June 2008,

3,
input data incuded simulation parameters and the initial pp. 187-201.

Major Characteristics of Simulation

Simulation is notstrictly a type of model;models generally represent reality, whereas sim

ulation typically imitates it. In a practical sense, there are fewer simplifications of reality

in simulationmodels than in other models. In addition, simulation is a technique

for
conducting experinments. Therefore, it involves testing specific values of the decision or

uncontrollable variables in the model and observing the impact on the output variables
At DuPont,decision makers had initially chosen to purchase more railcars; however, an
alternative involving better scheduling of the existing railcars was developed,tested, and
found to have excess capacity, and it ended up saving money.
Simulation is a descriptive rather than a normative method. There is no automatic
Search for an optimal solution. Instead, a simulation model describes or predicts the charac
teristics of a given system under different conditions. When the values thecharacteristic
of
are computed, the best of several alternatives can be selected. The simulation proces
usually repeats an experiment many times to obtain an estimate (and a variance) of the over
all effect of certain actions. For most situations, a
computer simulation is approprtae,
there are some well known manual simulations (e.g, a city police department simulated is
patrol car scheduling with a carnival gamewhee).
Finally, simulation is normally used only when a problem is toocomplex to be treaieu
using numerical opimization techniques. Complexity in this situation means either that U
rollen cannat be formulated for ptimization (e.g., becausethe assumptions do nol
that the formulation is toolarge, that there are too many or
interactions among the variables,
that the problem is stochasticin nature (i.e., exhibits risk or uncertainty).

Advantages of Simulation
SimulatÍon is used in MSS for the following reasons:

•The theory is faily straighforward.

•A great amount of time compression can be attained, quickly giving a manager
SOme

feel as to the long-term (1- to 10-year)

effects of many policies.
 Simulation is descriptive
Chapter 4 • Modeling and Analysis 173
rather than
what-if questions. normative. This allows the
Managers can use a trial-and-error manager to pose
and can do so faster, at approach to problem
less expense, solving
.A manager can experiment to more accurately, and with less
risk.
determine which
the environment are
An accurate
really
simulation model
forcing the MSS builder
important and with ,decision variables and

requires an intimate
to constantly
different alternatives
knowledge of the
.
which parts

problem,thus
DSS development interact with the
because the developer and [Link] is desirable for

.standing of the
problem and the potential
The model is built
.he simulation
from the
manager both gain a
decisions available.
manager's perspective.
better under

model is built for one

any other [Link], particular problem and
no generalized typically cannot solve
every component understanding is required of
. in the model
Simulation can handle an
corresponds to part of the real
extremely wide variety
system.
the manager;

and staffing, as well of problem types,

as higher-level such as inventory
managerial functions, such as
planning. long-range
•Simulation generally can include the
real complexities of problems;
are not necessary. For simplifications
example, simulation can use
rather than real probability
apprOximatetheoretical distributions. distributions
•Simulation automatically
produces many important
•Simulation isoften the only performance measures.
DSS modeling method that can
unstructured problems. readily handle relatively
•Some relatively easy-to-use simulation packages (e.g., Monte Carlo
available. These include simulation) are
add-in spreadsheet packages (e.g., ®RISK),
gram software, Java-based (and other influence dia
Web development)packages,and
interactive simulation systems the visual
to be discussed shortly.

Disadvantages of Simulation
The primary disadvantages of simulation are as follows:
•An optimal solution cannot be
• Simulation model constructionguaranteed, but relatively good ones are found.
generally
can be a slow and costly process,
modeling systems are although newer
use than ever.
easier to
•Solutions and from a simulation study are
inferences
usually not transferable to
other problems because the
model incorporates unique problem
•Simulation is sometimes so easy to explain to factors.

often overlooked. managers that analytic methods are

•Simulation softwaresometimes requires special skills because of the complexityof
the formal solution method.

The Methodology of Simnulation

Simulation setting up a model of a real
involves
system and conducting
Iments on The methodology consists of the
it. repetitive exper
following steps, as shown in Figure 4.12:
1. Define theproblem. We
examine and classify the real-world problem,
specifying
why a simulation approach is appropriate. The
system's boundaries,environment,
and other such aspectsof problem
clarification are handled here.
Construct the simulation model. This step
involves determination thevariables of

and their relationships, as well as data

gathering, Often the process is described by
using a flowchart., and then a
computer program is written.
3. Test and valdate the model. The
simulation model must properly
represent the
System being studied, Testing and validation
ensure this.
Part Il • Computerized Decision Support

Real-world
Changethe
problem realworld!

Test and Design and Evaluate the

Construct
Define the
validate conduct the experiments' Irmplemen
the
the the experiments results
simulation resuts
problen model model

Do-overFeedback

The Process of Simulation

FIGURE 4.12

model has been proven an

the experiment. When the

valid,
experime
4. Design the simulation is part this
how long to run step.

of
There
is designed. Determining
objectives: accuracy and cost.
i is
also prudent to

It
are twoimportant and conflicting

mean and median cases for random variables), best-case (eg

identify typical
(e.g.,

and worst-case (e.g., high-cost, loW-revenue)

scenaio
low-cost, high-revenue),
decision variables and environment in whid

These help establish the ranges of the

the simulation model.
work and also assist in debugging
to
Conducting the experiment involves issues ranging

5. Conduct the experiment.

from random-number generation to result presentation. to
The results must be interpreted. In addition
standand

6. Evaluate the results.

statistical tools, sensitivity analyses can also be used.
involves
Implement tbe results. The implementation of simulation results

the
7.
of success
implementation. However, the chances
are betr

same issues asany other

than wth
the manager is usually more involved with the simulation process
because
levels of managerial involvementgenerally lead to higher leves
other models. Higher
of implementation success.

Many simulation They are typically developed along u

packages are Web ready.
lines of the DSS architecture shown in Figure 3.4, where
a user connects to the man se
through a Web browser. This server connects to optimization servers and
database se

and the optimization and database servers in turn may connect to data warehouses,hot
populate the models.

Simulation Types
simul

Inthe following sections, we describe the major types of simulation: probabilistic and
simulation,
tion,time-dependent and time-independent simulation, object-oriented
visual simulation.

the independer
PROBABILISTIC SIMULATION In simulation, more cetl
of
probabilistic one or
variables (e.g.,the demand in an inventory They follow
problem) are probabilistic.
distributio

probability distributions, which can be either discrete distributions or corntinuous

vaniables

Discrete distributions events (or

• involve a situation with a limited number of

that can take on only a finite number of values. even

• Continuous distributions are situations with unlimited numbers of possible

that follow density functions, such as the normal distribution.

Conduce
The two types of distributions are shown in Table 4.6. Probabilistic simulation
is

with the aid of a used.

technique called Monte Carlo simulation, which DuPont
 Chapter 4 •Modeling and Analysis 175
4.6 Discrete Versus Continuous Probability
TABLE Distributions

Discrete Probability
paily Demand Continuous Probability

5 .10
Daily demand is normally distributed
with a mean of 7 and a standard
deviation of 1.2.

6 .15

7 .30
.25

.20

TIME-DEPENDENTVERSUS TIME-INDEPENDENT SIMULATION Time-independent refers to a

siuation in which it is not importantto know exactly when the event occurred. For
example, we may know that the demand for a certain product is three units per day, but we
do not care wben during the day the item is demanded. In some situations, time may not
be a factor in the simulation at all, such as in steady-state plant control design. However, in
waiting-line problems applicable to e-commerce, it is important to know the precise time of
arival (to know whether the customerwill have to wait). This is a time-dependent situation.

OBJECT-ORIENTED SIMULATION There have been some advances in the area of develop
ing simulation models by using the object-oriented approach. SIMPROCESS (CACI
Products Company, [Link]) is an object-oriented process modeling tool that lets the
user create a simulation model by using screen-based objects (e.g., metro cars). Unified
Modeling Language (UML) is a modeling tool that was designed for object-oriented and

object-based systemsand applications. Because UML is object oriented,

it Could be used in
practice for modeling complex, real-time systems. UML is particularly well suited for mod
eling. A real-time system is a software system that maintains an ongoing, timely interaction

with its environment; examples include many DSS and information and communication
systems (see Selic, 1999). Also, Java-based simulations are essentially object oriented.

VisUAL SIMULATION The graphical display of computerized results, which may incude
animation, is one of the most successful developments in computer-human interaction

and problem solving. We describe this in the next section.

Simulation Softvware

packages are available for a variety of decision-making

situations.
Hundreds of simulation
newer implementation packages are Java based.
NOSt run as Web-based systems,and
in Java. PC software packages
in
Many online demonstrations of simulations are written
[Link]) and the Excel add-ins Crystal
de Analytica (Lumina Decision Systems,
and@RISK (Palisade Corp., [Link]).
Decisioneering, [Link])
and SIMUL8 (SIMUL8
Web-based systems include WebGPSS (GPSS, [Link]), Arena
is
for simulation
Corp.,
[Link]). A major commercial size software
([Link]).

Simulation Examples
to analyze a call center. Jovanovic
Saltzman and (2001) used a simulation approach
Mehrotra This isimpor-
(2002) determined systems via simulation.
how to scheduletasks in distributed simulation to improve
tant when Dronzek (2001)used
managing grid computer networks. care system,
critical changes in a health
care in a military
hospital. He analyzed proposed
APPLICATION CASE 4.9
SimulationApplications

Simulation vary widely. Some are in scien

applications •IT/sOftwareproducts. CNA Financial Com

tific disciplines, others are in business, and yet others uses simulation tools from iRise, Inc., to deter.
are, essentially, video/computer [Link] following mine how potential software system require.
isa sampler of several simulation applications: ments will work out. The simulations brine

products to life and show business users how

E-learning. Tacit skill training is expensive, they will function in practice. In this way.

and people can experience walkthroughs

if the firm can communicate requirements from
without one-on-one human interaction, the business users to software developers more
savings can be dramatic. For example,the U.S. accurately, which, in turn, shortens therequire
Department of Agriculture (USDA)developed a ments cycle. This is critical when development is

Web-based simulation system to provide more offshored (i.e.,outsourced to a firm in another

than 400 geographically dispersed federal COuntry). Serena developed its ProcessVie
employees training on how to use the Composer software to simulate, visualize, and
PeopleSoft 8.0 human resource management prototype business processes and application
rewon
This has greatly reduced
sysiem. The cost of developing e-learning sim requirements.
ulations, even with custom video, can be less due to flawed requirements.
than $20,000. Humana, Inc., used simulation to •Web applications. In a manner similar
tos
train150 top executives in navigating complex CNA, Wachovia Corp. uses iRise simulation
Web app
change initiatives. Business simulations have in user design laboratory to simulate
a

useb.
cations before their release to
internal
been used for decades at leading colleges of Conducs
business (e.g., the Carnegie Mellon Tepper •Military battles. The U.S. military

School of Business). battle simulations running on high-powered

using

•Production scenarlos. Agilent's California graphics engines and supercomputers for

developed
Semiconductor Test division simulates supply software based largely on that
(video and Computer
síde production scenarios to coordinate sourc the entertainment industry training
a
simulate
ing for new product launches, to keep excess games), These Systems and
uncer
stress
and obsolete inventory (caused by engineering environment that reflects the
For example,
changes)low,and to respond quickly to ner tainty of battle situations. training
of flight
orders. One clear benefit has been a reduction simulators save many hours
intelligence eren
of 30to 40percentin the inventory of most time. By including artificial Can

Components, with some reductions as high as bilities, characters in the Simulations

reason on their own.
60percent.
 Pharmaceutical development. Pharmaceutical
shape of the chip caused it to fly
o Was that the
firms sinmulate how drugs interact with the off the manufacturer's conveyer belt if the belt
human body, down to the DNA level, thus
was running too fast. Based on information
accurately predicting their impacts. Savings found through simulation,the shape of the
are realized in the testing phases if potential
chip was redesigned.
compounds prove ineffective in the com
uter simulation. Pharmaceutical development Sources: Modified and abstracted from T. Hoffman, "Simulations
simulations run on high-powered parallel Revitalize Computerworld, August
E-Learning," 4, 2003, pp. 26-27,
S. Boehle, The Next Generation of E-Learning." Training,January
computers.
• Product design. Even simple productscan
benefit from simulation.
2005, pp. 23-31; R. Michel, "Decision Support Smoothes
Semiconductor Test Division's Outsourced Model," Manfacturing
Earlier in this section,
Business Tecbnology, June 2005, p. 30; H. Havenstein,"ApP
we described how General Motors uses simu Simulation in Demand," Computerworld, June 27, 2005, pp. 1, 7;
lations in new product design and manufactur D. Verton, "Simulating Fallujah," Compuierworld, January 31, 2005,
ing, Even the aerodynamics of Pringles chips p. 28, "Million-Atom Biology Simulation Could Improve Drug

have been analyzed using simulationmethods Development, Computeruorld, November 14, 2005, p. 39; P.
Arena, S. Rhody, and M. Stavrianos, "The Truth About Facts," DM
to determine how air flows around their
Review,anuary 2006, pp. 30-34; and D. Briody, The Flight
of the

unique double-saddle shape. The problem Pringle," CIO Insight, August 2005,p. 22.

Section 4.11 Review Questions qeols ti olqooo td t aolontss to

[Link] the characteristics of simulation.

it aaebris iitolis s n
2. List the advantages and disadvantages of simulation
s3 2o0st
3. List and describe the steps
4. List and describe the types of
the methodology of simulation
in

simulation.
w siliISl nisg
1 sei a o
4.12 VISUAL INTERACTIVE SIMULATION
We next examine methods that show a decision maker a representation of the decision
making situation in action as it runs through scenarios of the various alternatives. These
powerful methods overcome some of the inadequacies of conventional methods and
help build trust in the solution attained because they can be visualized [Link]

Conventional Simulation Inadeguaciesasp s aiil ahus

Simulation is a well-established, useful, descriptive, mathematics-basedmethod forgaining

insight into complex decision-making situations. However, simulation does not usually
allow decísion makers to see how a solution to a complex problemn evolves over
(compressed) time, nor can decision makers interact with the simulation (which would be
purposes and teaching). Simulationgenerally reports statistical results at
useful for training
he end of a set of [Link] makers are thus not an integral part of simula
ion development and experimentation, and their experience and judgment cannot
De used directly. If the simulation results do not match the intuition or judgment of the
decision maker, a confidencegap in the results can ocCur.g V ool vioa
Visual
Interactive Simulation

sual interactive simulation (VIS), also known as visual interactive modeling

M) and visual interactive problem solving, is a simulationmethod that lets decision
makers see what the
model is doing and how
it interacts with the decisions made, as they

The technique has been used with great sucess in operations management
made.
O Ihe user can employ his or her knowledge to deternmine and try different decision
Part II • Computerized Decision Support

with theEnhanced learning, about both

model.
strategies while interacting the
tested, can and does oCcur. Decision problenm
and the impact of the alternatives makers
contribute to model validation. Decision
makers who use VIS generally support also
andinus
their results.
the impact of
animated computer graphic displays to present
VIS uses different
that the user can adjust the man
man-
agerial decisions. It differs from regular graphics in
visual model is a graphic decision
making process and see the results of theintervention.
A

used as
making or problemn solving, not just as a
an integral part of decision Communication

VIS displays the effects of different decisions in graphic form on a computer screen, as
device.

done through the GIS in the P&G supply chain redesign through optimization. Some
respond
people better than others to graphical displays, and this type of
interaction
Can
help managers learn about the decision-making For example, Swisher
situation.

(2001) applied an object-oriented visual simulationto examining the functioning fa of

physi-
cian clinic environment within a physician network to provide high-quality, cost-effertu

health care in a family practice. The simulation system identified the most important innr

factors that significantly performance. These inputs, when properly managed, lei
affected
to lower costs and higher service levels. Many Java-based VIS demonstrations run directy
on the Internet.
VIS can represent static or dynamic systems. Static models display a visual image of
the result of one decision alternative at a time. Dynamic
models display systems that
evolve over time, and the evolution is represented byanimation. The latest
visual simula
tion technology has been coupled with the conceptof virtual reality,
where an artificial
world is created for a number of purposes, from
training to entertainment to viewing data
in an artificial landscape. For example, the
U.S. military uses VIS
systems so that ground
troops can gain familiarity with terrain ora city in
order tovery quickly orient themselves.
Pilots also use VIS to gain
familiarity with targets by simulating
software can also include GIS attack runs. The VS
cOordinates.