University of Nebraska - Lincoln

DigitalCommons@University of Nebraska - Lincoln

Industrial and Management Systems Engineering --
Industrial and Management Systems Engineering
Presentations

1-1-2003

Discrete-Event Simulation: Simulation Practices

and Trends
Paul Savory
University of Nebraska at Lincoln, psavory2@gmail.com

Follow this and additional works at: http://digitalcommons.unl.edu/imsepresentations

Part of the Industrial Engineering Commons, and the Operational Research Commons

Savory, Paul, "Discrete-Event Simulation: Simulation Practices and Trends" (2003). Industrial and Management Systems Engineering --
Presentations. Paper 2.
http://digitalcommons.unl.edu/imsepresentations/2

This Article is brought to you for free and open access by the Industrial and Management Systems Engineering at DigitalCommons@University of
Nebraska - Lincoln. It has been accepted for inclusion in Industrial and Management Systems Engineering -- Presentations by an authorized
administrator of DigitalCommons@University of Nebraska - Lincoln.
Simulation Practices and Trends

Paul A. Savory, Ph.D.

Industrial and Management

Systems Engineering
University of Nebraska - Lincoln

University of Nebraska
 What is a Model?

A model is a simplified representation

of a system (or process or theory)
intended to enhance ones ability to
understand, predict, and possibly
control the behavior of the system

University of Nebraska
Why Develop a Model?

an aid to experimentation
a tool of prediction
an aid to thought
an aid to communication
an aid to training and instruction

University of Nebraska
 How the Model is Used

FORMULATION
REAL
SYSTEM MODEL

DEDUCTION

REAL INTERPRETATION MODEL

CONCLUSIONS CONCLUSIONS

University of Nebraska
 Problem Solving Approaches

Computer OR Expert Artificial

Applications Models Databases Systems Intelligence

CERTAINTY UNCERTAINTY
The spectrum goes from physical models or prototypes which are expensive
and time consuming to build and need exact conditions to model the real system
and proceeds to completely abstract mathematical models

University of Nebraska
 Factors Affecting the
Development of a Model
SYSTEM

Purpose Level of Boundaries

Detail
Performance
Measures Design
Alternative
CONCEPTUAL
MODEL

Assessment

IMPLEMENTATION

University of Nebraska
What is a Computer Simulation?

Process of developing a
mathematical-logical
model of a system and
experimenting with that
model on a computer

University of Nebraska
 Why Use Simulation?
The real system does not exist and it is too costly, time-
consuming, hazardous, or impossible to build

The real system exists but experimentation

is expensive, hazardous, or seriously disruptive

The real system exists but experimentation

may destroy it

Mathematical modeling of the system has no practical or

numeric solutions

University of Nebraska
A Job Shop Example

University of Nebraska
 Analysis of a Job Shop
Factors affecting our analysis
set-up times; unique production sequence of
each part; machine breakdowns; limited
waiting space

What types of performance measures would

you like to obtain?
work-in-progress; production time;
machine/worker utilization

University of Nebraska
 A Heuristic Model
One approach to studying this system is to an analytical
technique which seeks to get some unique and/or optimal
solution

Due to this complexity, possibly we can build a heuristic

model
construct a model of the system
observe the behavior of the performance variables over time
based on the observations, hypothesize how the real system
is behaving
These three tasks describe what simulation performs!

University of Nebraska
 Simulation Disadvantages
Simulation models are expensive and time
consuming to develop

Simulation models give estimates of

performance characteristics and offer few
procedures for optimization

Simulation models produce large quantities of

output. This does not mean the model is
correct!

University of Nebraska
 The Simulation Process
Step 1: Problem Definition
Develop the preliminary model

Step 2: Project Planning

Decide on personnel, management support,
computer hardware, and budget for the study

Step 3: System Definition

Identify the small subset of characteristics or
features of the system to include

University of Nebraska
Organizational Activities and
Simulation
PRODUCTION ACTIVITY
40
1. RESEA RCH
2. PROJ ECT
30 3. M A SS PRODUCTION
4. BA TCH PRODUCTION
5. J OB SHOP
20 6. CONTINUOUS FLOW LINE
7. OTHER
10

0
1 2 3 4 5 6 7
Production Type

One third of practitioners perform simulation research

University of Nebraska
 Organizational Activities and
Simulation
USE OF SIMULATION - PURPOSE
30
1. DESIGN
2. RESEA RCH
3. PLA NNING
4. SCHEDULING
20
5. A SSIGNM ENTSA ND A LLOCA TIONS
6. OTHER

10

0
1 2 3 4 5 6

Simulation is highly used for design, research, and planning

Industry uses simulation most for design and planning

University of Nebraska
 Project Contributors
SIMULATION TEAM MEMBERS
30
1. SY STEM A NA LY ST
2. SIM ULA TION DEV ELOPER
3. CONSULTA NT
4. SHOP FLOOR SUPERV ISOR
20
5. M A NA GER
6. OTHER

10

0
1 2 3 4 5 6

Simulation developers and systems analysts usually

lead the analysis team

University of Nebraska
 The Simulation Process
Step 4: Conceptual Model Formulation
Develop the preliminary model

Step 5: Preliminary Experimental Design

What output do you want?

Step 6: Input Data Preparation

Collect data and fit probability distributions to it

University of Nebraska
 The Simulation Process
Step 7: Validation
Does the model describe
the real system?

Step 8: Model Translation

Formulate the model in an appropriate language

Step 9: Verification
Does the computer simulation model execute as
intended?

University of Nebraska
 Hardware Configuration
COMPUTER HARDWARE
50

1. PERSONA L COM PUTERS

40 2. W ORKSTA TIONS
3. M A INFRA M ES
30 4. M INI-COM PUTERS

20

10

0
1 2 3 4

PCs and workstations are widely used simulation platforms

Mainframes are least preferred by industry

University of Nebraska
 Simulation Software
GENERAL LANGUAGE
30
1. C
2. FORTRA N
3. PA SCA L
20 4. BA SIC
5. LISP
6. OTHER
7. NONE
10

0
1 2 3 4 5 6 7

C and FORTRAN are widely used, PASCAL is not used in industry

University of Nebraska
 Simulation Software
SIMULATION LANGUAGES
30
1. SLA M
2. SIM A N
3. GPSS
20 4. SIM SCRIPT
5. OTHER
6. NONE

10

0
1 2 3 4 5 6

SLAM and SIMAN are equally popular

Industry prefers SIMAN to SLAM, Universities prefer SLAM to SIMAN

University of Nebraska
 A Banking System

Customers arrive Service time is

according to an between 2 and 5
exponential distribution minutes
with a mean of 2 minutes

University of Nebraska
 SIMAN Simulation Model
CREATE: EXPON(2):MARK(ArrivalTime)
QUEUE, WaitTellerQ;
SEIZE: Teller;
DELAY: UNIFORM(2,5);
RELEASE: Teller;
TALLY: TimeInSystem, INT(ArrivalTime);
DISPOSE;

University of Nebraska
 The Simulation Process
Step 10: Pilot Runs
Collect a small quantity
of output

Step 11: Validation

Does the simulation model approximate what
would be expected?

Step 12: Final Experimental Design

Design an experiment to yield the desired output

University of Nebraska
 The Simulation Process
Step 13: Make Production Runs

Step 14: Output Data Analysis

Compute averages, standard
deviations, and confidence
intervals on performance variables

Step 15: Documentation, Presentation, and

Implementation

University of Nebraska
 Project Effort Allocation

TIME FOR EACH PHASE

20 1. M ODEL FORM ULA TION
2. M ODEL BUILDING
3. DA TA A CQUISITION
15 4. M ODEL TRA NSLA TION
5. V ERIFICA TION
10 6. V A LIDA TION
7. STRA TEGIC A ND
TA CTICA L PLA NNING
5 8 . EX PERIM ENTA TION
9 . A NA LYSIS OF RESULTS
0 1 0 . IM PLEM ENTA TION A ND
1 2 3 4 5 6 7 8 9 10 DOCUM ENTA TION
Simulation Phase

University of Nebraska
 Available Computer Assistance

COMPUTER EFFORT FOR EACH PHASE

1. M ODEL FORM ULA TION
100
2. M ODEL BUILDING
80
3. DA TA A CQUISITION
4. M ODEL TRA NSLA TION
60 5. V ERIFICA TION
6. V A LIDA TION
40 7. STRA TEGIC A ND
TA CTICA L PLA NNING
20 8 . EX PERIM ENTA TION
9 . A NA LYSIS OF RESULTS
0 1 0 . IM PLEM ENTA TION A ND
1 2 3 4 5 6 7 8 9 10
Simulation Phase DOCUM ENTA TION

University of Nebraska
What You Should Remember

Simulation only provides an

estimate of the true
performance characteristic, not
the characteristic itself

Simulation DOES NOT give an

exact answer!

University of Nebraska