1

©2005 Tag Gon KimEE612 Lecture 3

1 of 13Multiplicity in Modeling and SimulationMultiplicity in Abstraction Level

What are ignored and what are consideredDifferent abstract level different model

Multiplicity in ResolutionRepresentation of a variable in different rangesMulti-resolution modeling

Multiplicity in AspectView of modelingDifferent aspects of modeling different models

Decision of the multiplicities Objectives of Modeling and Simulation

2

©2005 Tag Gon KimEE612 Lecture 3

2 of 13Objective-driven Modeling Simulation: Concept

Questions Modeling Experiments Answers

question 1(Throughput)

question 2(Utilization)

question n

System

Experiment 1 Answer 1

Experiment 2 Answer 2

Experiment 3,4

Experiment 5

Answer 3,4

Answer 5

Experiment k-1

Experiment k

Answer k-1

Answer k

ModelingObjectives

To answerquestion 1,2

Model 1

To answerquestion 3,4,5

Model 2

< <1System

mModels

kExperiments

To answerquestion n-1,n

Model m

3

©2005 Tag Gon KimEE612 Lecture 3

3 of 13Objective-driven M&S: Example

Computing Power?

TemperatrureCharacteristics?

Solidity?

Failure Rate?

ArchitectureModel

Program Exe timeSolidityModel

ExternalVibration

OperationalCorrectness

Temp CharModel

Chan of Temp OperationalCorrectness

ReliabilityModel

ComponentsReliability

MTTF

4

©2005 Tag Gon KimEE612 Lecture 3

4 of 13Experimental Frame Concepts

Simulation of Combinations(i) Alternative scenarios with

an experimental frame(ii) A scenario with

alternative experimental frames

Model 3Model 2

Model 1

Models under TestData

Generator

Acceptor

Receiver

Experimental Frame

Input

Run Control

Statistics

IC

Circuit under Test Experimental Devices

Oscilloscope

Signal Generator

Voltmeter

5

©2005 Tag Gon KimEE612 Lecture 3

5 of 13Purpose of DES Modeling/Simulation

Logical/Behavior Analysis Performance Evaluation

Average waiting timeThroughputUtilization

Properties (Desired states sequence)Safeness (Bad things will not happen)

Liveness (Good things will eventually happen)Example

Totally orderedUnordered/partial orderedEvent Orderin

Time Untimed DES ⊇ Timed DES

State/event sequence

Untimed (time unspecified) DES Model

RequiredInformation

inModeling

S1 S2 S3e1 e2

e2e1

Timed state/event sequence

Timed DES Model

S1 S2 S3e1:T1 e2:T2

e2e1

Complexity High Low

6

©2005 Tag Gon KimEE612 Lecture 3

6 of 13Math Formalisms for DES Analysis

State sequence Timed state sequence

Untimed DES model Timed DES model

Safeness, Liveness Throughput

Required information

Model type

Example

Behavioral analysis(Correctness)

Performance analysis(Efficiency)

Logic base

Process algebra

Set/Bag theory

Temporal Logic

CSPCCS

FSM Petri net

Automata

GSMPMin-Max algebra

Timed FSM Timed-PN

DEVS formalism

7

©2005 Tag Gon KimEE612 Lecture 3

7 of 13

Counting problem

cases to select x out of n

Combination : nCx

!)!(!xxn

n−

Advantages of Formal MethodsCompleteness TestabilityCommunication meansMathematical manipulation

Analysis and Design of

Complex System

Person-1: (1,2,3) (1,2,4) (1,2,5) ……

Person-2: (10,9,8) (10,9,7) (10,9,6)...

Informal method Formal method

Role of Formal (Mathematical) Modeling

8

©2005 Tag Gon KimEE612 Lecture 3

8 of 13Programming vs Formal Modeling

Object: RectangleObject: Rectangle

aabb

Object: TriangleObject: Triangleaa

xx yy

Interface Interface Triangle exactly Triangle exactly on rectangleon rectangle

Programming

Programming

Formal M

odeling

Formal M

odeling

Prog 1

Prog 2

Prog 3 Error !!

9

©2005 Tag Gon KimEE612 Lecture 3

9 of 13

Advantages of Formal Methods• Completeness • Testability• Communication means• Mathematical manipulation

Counting problem DES problem

Case to be counted Behavior of process

Combination : nCx

X S Y

δintδextλtaDEVS formalism

!)!(!xxn

n−

Simulation Environment(DEVSim++, DEVSim-Java)

Informaldescription

FormalModel

SimulationAnalysis

Group Working for Modeling of

Complex Systems

Formal Modeling Framework: Combination vs DEVS

10

©2005 Tag Gon KimEE612 Lecture 3

10 of 13

Event

Acquisition ofsame information

for same event regardless of Newspapers

5W1H :Framework for information

representation and verification

Pressman ReaderCorrespondent

Transfer

Transfer

Transfer

Verification

Verification

Verification

Different writing stylefor different pressman

Framework for Writing Newspaper Article: 5W1H

11

©2005 Tag Gon KimEE612 Lecture 3

11 of 13

DES

Same information on

system behavior

regardless of implementation

language

DEVS :Framework for modeling

and verification

Modeling expert Simulatordeveloper

Transfer

Transfer

Transfer

Domainexpert

Verification

Verification

Verification

Model implementation using different

programming languages

Framework for Discrete Event Modeling: DEVS(3S4F)

12

©2005 Tag Gon KimEE612 Lecture 3

12 of 13

CompletenessCheck what’s missing in 5W1H /3S4F

VerifiabilityCheck each of 5W1H / 3S4F in a model against the real system

Modifiability/ExtensibilityModify/extend each of 5W1H / 3S4F independently

Efficient Maintenance Maintain each of 5W1H / 3S4F separately

Why Framework for Information Modeling

13

©2005 Tag Gon KimEE612 Lecture 3

13 of 13Method for DES Simulation Modeling

Informal Modeling

Formal Modeling

Method Example

Modeler’s world view of system behavior

Mathematical representationof system behavior

Event-orientedProcess-orientedActivity-orientedObject-oriented

DEVS Formalism

Merits/Limitations

Simulation language basedEasy to modeling

Impossible to manipulate mathematically

Sound frameworkMathematical manipulationGeneral purpose language

Cost for learning