DEVS as a Semantic Domain for Programmed Graph Transformation

McGill UniversitySchool of Computer Science

Ph.D. Student in the Modelling, Simulation and Design Lab

MSDL’08

DEVS as a Semantic Domain for Programmed

Graph Transformation

Eugene Syriani

DEVS: A Semantic Domain For Programmed Graph TransformationMSDL’08

REFERENCE• GraBaTS 2008 Tool contest

E. Syriani and H. Vangheluwe: Using MoTif for the AntWorld Simulation Tool Contest

• DEVS Book (2008)E. Syriani and H. Vangheluwe: DEVS: A Semantic Domain For Programmed Graph Transformation

2


OVERVIEW

Background

In the context

Case Study: The extended AntWorld Simulation

The Modular Timed Graph Transformation language (MoTif)

From DEVS to MoTif

The MoTif solution for the AntWorld case-study

Rules & Control Structure

Real-Time Model Transformation & Deployment

Conclusion and Future Work3


IN THE CONTEXT• Model Transformation Graph Transformation

• Controlled Graph Rewriting (rewriting rules)

• Control Primitives :Sequencing, Branching, Looping, Parallelism, Hierarchy

• Time is inherent in our transformation models4


CASE STUDY: ANTWORLD SIMULATION [1]

5[1] http://www.fots.ua.ac.be/events/grabats2008/ (2008/07/21)

EXTENDED


CASE STUDY: ANTWORLD SIMULATION Domain-specific modelling

6


CASE STUDY: ANTWORLD SIMULATION

• Create grid nodes when ant reaches outmost circle

• Differentiate between nodes along main axis

• Ant movement in search mode

• Ant movement in carry mode

• Ant creation, once food is brought back to the hill

• Evaporation of pheromones

Simulation Semantics

7


CASE STUDY: ANTWORLD SIMULATION++

• At any moment in time, a human steps on the map– Kills all ants on the five nodes

– The nodes loose food and pheromones

Extended Simulation Semantics

8


FROM DEVS TO MOTIF

9


DEVS• The Discrete Event System Specification [2]

(DEVS) formalism

• Highly compositional simulation framework

• DEVS:– Blocks

– Ports

– Events

• Semantics: Parallel composition of blocks/models

10[2] Zeigler B.P., Multifacetted Modelling and Discrete Event Simulation/ Academic Press, London (1984)


DEVS META-MODEL


MoTiF [2]• Blocks– Atomic: encapsulate a graph rewriting rule

– Coupled: encapsulate a set of rules (graph grammar)

• Events– Inport: receive the host graph

– Outport(s): send the transformed graph

12[2] Syriani E. and Vangheluwe H.: Programmed Graph Rewriting with DEVS. AGTIVE 2007, LNCS (2008)


MOTIF META-MODEL


MOTIF AS A DEVS MODEL

14



15

• At most one rule will execute



• Synchronize parallel threads

• Success if at least one thread succeeds16



17

The select function:1. If the Selector is in the imminent list, choose the Selector.

2. Among all the rules that still have a match, choose a corresponding ARule from the imminent list at random, no matter what depth it is at inside this CRule.

3. At this point no rule has any unprocessed match left, then choose any of the ARule in the imminent list.

4. Finally, the imminent list contains either custom atomic blocks or Synchronizers.

5. Proceed with a first-in-first-out selection.


MoTiFMODELLING ENVIRONMENT IN AToM3 [3]

18[3] de Lara J., Vangheluwe H., AToM3: A tool for multi-formalism and meta-modelling. FASE, LNCS 2306 (2002), 174-188


MoTiF EXECUTION

COMPILE

GENERATE

IMPORT

SIMULATE

Class Grabfood : def match(): … def execute(): …

class Grabfood(ARule): def __init__(self): ARule.__init__(self, name=‘GF') self.state = ARuleState(Grabfood()) def weightFunction(self): return 1.0

Class Grabfood : def match(): … def execute(): …

class Grabfood: def match(): … def execute(): …

19


MOTIF SOLUTION

20


ANTWORLD RULES

21


SOME ANTWORLD RULES

22

dropFood checkAntOnOutCircle connectNodesInSameCircle

goToSameNodeIn moveTowardsHill grabFood

KillOnNode KillOnNextCircleNodeIn WeakenSameNodeOut


MODELLING OF THE TRANSFORMATION: Round CRule

23


MODELLING OF THE TRANSFORMATION: AntMovements CRule

24


MODELLING OF THE TRANSFORMATION: GenerateCircle CRule

25


MODELLING OF THE TRANSFORMATION: EndOfRound CRule

26


MODELLING OF THE TRANSFORMATION: HumanStep CRule

27


MODELLING OF TRANSFORMATION: SYSTEM

28

• User – Controller – Autonomous loop

• Feed-back to User

• On User interrupt: User Controlled

• Feed-back to User


SYNTHESIS OF APPLICATION [4]• Synthesis of a reactive application

• Discard the UserBehavior block

• Runs on a web browser using AJAX and SVG

• Real-time model transformation (RT-DEVS)

29[4] E. Syriani and H. Vangheluwe. Programmed Graph Rewriting with Time for Simulation-Based Design. ICMT 2008. LNCS, vol.5063, pp. 91-106 (2008)


CONCLUSIONFormalised MoTif in terms of DEVSProvided solution to the AntWorld case-studyExtended solution with user interruptDeployment of an application

? Pure parallel execution, possibly distributed? Re-use and parametrisation of transformation

models? Allowing variable structure

30


31

?

Documents

DEVS as a Semantic Domain for Programmed Graph Transformation