DEVS and SES as a Framework for Modeling and Simulation Tool

DevelopmentBernard P. Zeigler

Arizona Center for Integrative Modeling and Simulation

University of ArizonaPresented at SIMUTools 2008 March 3–7 2008 Marseille, France

M&S Framework

Source System

Simulator

Model

Experimental Frame

SimulationRelation

ModelingRelation

• Based on mathematical formalism using system theoretic principles

• Separation of Model, Simulator and Experimental Frame

• Atomic and Coupled types

• Hierarchical construction

x0 x1X

S

Yy0

e

t0 t1 t2

Discrete Event Time Segments

M&S Framework formulated within UML

classes

construction mechanisms

relationshipsconstraints

create classes to satisfy use cases

M&S Framework classes

interpretationas software code(e.g. Java)

Use Cases

UML

models

simulatorsSourcesystems

Experimental& PragmaticFrames

ontologies

collectdata

storemodel

runmodel

construct

model

storedata

simulatemodel

sensorretrievedata

retrievemodel

validatemodel

memory

control-simulator predictive

-simulator

modelevaluator

modelgenerator

real world

model generated

goal-generator

action planningmotor

objectives

M&S

applypragmatic

frame

applyexperimen

talframe

DEVS – Formal Specification of a System

A discrete event system specification (DEVS) is a structureM=<X,S,Y,int, ext, con, ,ta>

whereX is the set of input values,S is a set of states,Y is the set of output values, int :S->S is the internal transition function, ext,:QX->S is the external transition function,

con,:QX->S is the confluent transition function, ta: S->R +

0,∞

WhereQ={(s,e)|sS, 0 e ta(s) } is the total state set,e is the time elapsed since last transition, :S->Y is the output function and R +

0,∞ is the set of positive reals with 0 and

DEVS Hierarchical Modular Models

DEVSModel

Output port

Input port

DEVSModel

message

coupling

SES

Hierarchical Composition, Coupling and Variants are Represented in System Entity Structure

System Entity Structure

Atomic Models

OrdinaryDifferentialEquationModels

Spiking NeuronModels

Coupled Models

Petri NetModels

Cellular Automata

n-Dim Cell Space

PartialDifferentialEquations

Self Organized CriticalityModels

Processing/Queuing/

Coordinating

ProcessingNetworks

Networks,Collaborations Physical

Space

Some Types of Models Represented in DEVS

can becomponents in a coupled model

MultiAgent

Systems

Discrete Time/

StateChartModels

QuantizedIntegratorModels

Spiking Neuron

Networks

StochasticModels

ReactiveAgent

Models

Fuzzy Logic

Models

Concept of DEVS Standard

DEVS

Simulator

Interface

Single

processor

Distributed

Simulator

Real- Time

Simulator

C++

Non

DEVS

DEVS

Model

Interface

Java

Other

Representation

DEVSSimulationProtocol

Virtual- Time

Simulator

DEVSML

DEVS Simulation Protocol

Coordinator

Atoimc1

Non-DEVS Simulator

Atoimc2

simulators.tellAll("initialize“)

simulators.AskAll(“nextTN”)

simulators.tellAll("computeInputOutput“)

simulators.tellAll("sendMessages")

simulators.tellAll("

Coordinator

DEVSModel1

simulators.tellAll("initialize“)

simulators.AskAll(“nextTN”)

simulators.tellAll("computeInputOutput“)

simulators.tellAll("sendMessages")

simulators.tellAll("ApplyDeltFunc”)

putContentOnSimulator

DEVS Simulator

DEVS Simulator

DEVSModel2

?

Finite Deterministic DEVS : FD-DEVS

FDDEVS = <incomingMessageSet, outgoingMessageSet, StateSet, TimeAdvanceTable, InternalTransitionTable, ExternalTransitionTable,OutputTable>whereincomingMessageSet, outgoingMessageSet, StateSet are finite setsTimeAdvanceTable: StateSet → R0,∞

+ (the positive reals with zero and infinity)InternalTransitionTable: StateSet → StateSetExternalTransitionTable: StateSet × incomingMessageSet → StateSet, OutputTable: StateSet → the set of subsets of outgoingMsgSet

Natural Language For FDDEVS•to start hold in PHASE for time SIGMA •when in PHASE and receive MSG go to PHASE’ <eventually> •hold in PHASE for time SIGMA •after PHASE then output MSG •from PHASE go to PHASE’

Semantics defined by mapping into DEVS

FD-DEVS

Example of Natural Language Spec: 1.to start passivate in waitForStart2.when in waitForStart and receive StartAck go to sendFirst3.hold in sendFirst for time 0.5 then output

ReqForAck and go to sendSecond4.hold in ackReceived for time Infinity5.…

waitForStart

sendFirst

sendSecond

sendThird

AckReceived

StartAck

AckAck

0.510

10

10

∞ ∞

lastChance

ReqForAck

ReqForAck

ReqForAck

Ack

• The “right” abstraction of DEVS – retains important timing properties

• Amenable to analysis • Supports automation• Maps to DEVSJAVA• Supplies a skelton that can be extended to full

DEVS• Simple XML expression

Automated Analysis: Based on State Input Pairs Enables Automated Test Model Generation

listen

AwaitingAdv

OrderOffer ProdFeasibilityFalseProdFeasibilityTrue

waitForConfirm

ConfirmDominantState-Input pair:listen ,orderOffer

DominantState-Input pair:listen ,orderOffer

State-Input pair paths:1= listen ,OrderOffer 2 = listen , OrderOffer, awaitingAdv ,ProdFeasibilityFalse3= listen ,OrderOffer awaitingAdv ,ProdFeasibilityTrue4= listen ,OrderOffer awaitingAdv ,ProdFeasibilityTrue waitForConfrm ,Confirm

Dominant State-Input pair paths:24

Defined State-Input pairs:listen ,orderOfferawaitingAdv,ProdFeasibilityFalseawaitingAdv, ProdFeasibilityTruewaitForConfrm ,Confirm

Defined State-Input pairs:listen ,orderOfferawaitingAdv,ProdFeasibilityFalseawaitingAdv, ProdFeasibilityTruewaitForConfrm ,Confirm

Terminal State-Input Pairs: awaitingAdv , ProdFeasibilityFalse

waitForConfrm , Confirm

Terminal State-Input Pairs: awaitingAdv , ProdFeasibilityFalse

waitForConfrm , Confirm

State-Input Pairs mapped To Terminal State-Input pairs:

listen ,orderOffer -> awaitingAdv , ProdFeasibilityFalseawaitingAdv , ProdFeasibilityTrue -> waitForConfrm, Confirm

Generated Test Models in DEVSJAVA SimView

AckCheck

Send inputs

Observe Outputs

Trigger next in Sequence upon pass

Observe Pass/Fai

l

System Entity Structure (SES) : SESBuilder

System Entity Structure/Model Base Repository: Support Automated DEVS Generation and Reuse

Modelconstruction SES

Model Repository

ExperimentalFrame

pruneand transform

Automated Generation of Test Models

SES/XMLOf DSRtexts

SES

PrunedES

FDDEVS atomic model

FDDEVS test models

Augmented Test Models

Automated manual

DEVS/SOA Infrastructure: Supports Deployment and Execution of DEVS Models on the Web

WEBSERVICECLIENT

Middleware (SOAP, RMI etc)Net-centric infrastructure

DEVS Simulator Services

DEVS Modeling Language (DEVML)

DEVSJAVA

DEVSAgent

( Virtual User)

DEVSAgent

(Observer)

WEBSERVICECLIENT

Run Example

• Service Oriented Architecture (SOA) consists of various W3C standards

• Machine-to-machine interoperable interaction over the network based on WSDL interface descriptions

• Client server framework

• Message encapsulated in SOAP wrapper which is in XML

Deploying Models: DEVSML and DEVS/SOA

Applications

• Natural language capture of high level information technology systems requirements

• Automated generation of FDDEVS kernel DEVSJAVA/C++ models for distributed real-time net-centric IT systems testing

• Development of web service workflows using DEVS/SOA

• Network Traffic data capture, focused extraction, and model generation for exercising IT systems e.g., intrusion detection

Conclusions

• DEVS and SES provide a framework based on Systems Theory for M&S tool development

• DEVS standard supports sharable models and repository reuse

• The framework provides an ontology in which tools can be organized and interoperate – missing slots can be identified; overlapping tools can be compared

• The framework supports development of generic tools which in turn support a wide array of domain specific specializations and applications

devsworld.org acims.arizona.edu Rtsync.com

Books and Web Links