International Semantic Web Doctoral Symposium Research Topic: Representing Discrete-Event Simulation Process-Interaction Models using the Web Ontology Language - OWL

1
International Semantic Web Doctoral
SymposiumResearch TopicRepresenting
Discrete-Event Simulation Process-Interaction
Models using the Web Ontology Language - OWL

• November 7, 2005
• Lee W. Lacy
• PhD Candidate
• University of Central Florida

2
PIMODES Research Overview

• Discrete-event simulation models support
  operations research and other applications
• These models have historically been represented
  in vendor-specific file formats that have made
  sharing/interchange difficult
• Research is being performed to develop an OWL
  ontology that will provide a neutral interchange
  description of these types of models.
• The ontology is being scoped to a particular type
  of discrete event simulation model descriptions
  those that adhere to the process-interaction
  world view
• The interchange of simulation models using the
  ontology will be demonstrated using web-based
  software

3
Presentation Outline

• Purpose Description
• Goal Statement
• Methodology
• Evaluation
• Summary

4
Purpose Description

• Simulation Model Interchange Challenges
• Scoping the Problem
• Subject Domain
• Benefits of Interchange

5
Why Interchange Models?

• Leverage investment in model development through
  reuse
• Higher quality through reuse of validated models
• Speed development lifecycle
• Reduce development costs
• Enable competition of model development
  environments and compliant execution engines
• Potential software manufacturer push-back if not
  presented correctly
• Need to sell the HTML and XML business models
• Shift model development emphasis from programming
  to model quality

6
Scoping the Proposed Research

• Simulation data interchange topic broad
• Various types of simulation data
• Emphasis shifting from code to data
• Simulation models represent one type of data in
  newer data-driven systems
• Model is problem-specific while execution engine
  is problem-independent
• Discrete-event simulations represent one type of
  simulation
• Further scoped by world view

7
Subject Domain
8
Benefits of Model Interchange

• Better
• Reuse of validated models
• Faster
• Quicker to create new models by leveraging
  existing models
• Cheaper
• Lower cost due to reuse instead of creation

9
Goal Statement

• Develop an ontology to support Discrete Event
  simulation model interchange
• Ontology becomes a de facto language
• New ontology/language harmonizes the most
  important aspects of legacy languages
• Legacy models can then be converted to/from the
  new lingua franca enabling interchange

10
Methodology

• Use of OWL Ontologies
• Research Activities
• Research Plan
• Feedback Opportunities
• Anticipated Results

11
Simulation Ontology Representations

• Proposed by Lacy (2004) and Miller Fishwick
  (2004)
• Provides advantages over traditional (e.g., XML)
  approaches
• Requires the development of a meta-model
• Existing modeling languages have implicit
  ontologies
• New explicit ontologies in effect describes a new
  modeling language
• Mappings required from legacy languages to the
  new ontology

12
Web Ontology Language - OWL

• OWL became a World Wide Web Consortium (W3C)
  Standard in February 2004
• OWL will be used to define a Process-Interaction
  Modeling Ontology for Discrete-Event Simulations
  (PIMODES)
• Compliant instance files are represented using
  RDF/XML

13
OWLs Layered Architecture
14
Research Plan
15
A0 View
16
A2 Literature Search (2) Activities

• Formalize domain semantics (2.1)
• Survey existing discrete-event process-interaction
  (2.2)
• Software packages (2.2.2)
• Modeling languages (2.2.3)
• Formalisms/Representation methods (2.2.4)
• Review related simulation information interchange
  research (2.3)
• Describe Semantic Web technology (2.4)

17
Evaluation

• Develop sample model in ProcessModel and Arena
• Convert legacy model representations to DEPIM
• Convert DEPIM representation to legacy formats
• Compare converted models to original models

18
Demonstration Data Flow
19
Summary

• Simulation data is interchanged in a variety of
  ways
• Interchange is best performed with open standards
• OWL can be used to define an ontology for
  Discrete-Event Process-Interaction models
• Use of such an ontology can be demonstrated by
  converting legacy simulation model formats