1 / 16

An Ontology for Agent-based Modeling and Simulation

An Ontology for Agent-based Modeling and Simulation. Scott Christley, Xiaorong Xiang, Greg Madey Dept. of Computer Science and Engineering University of Notre Dame. Supported in part by National Science Foundation, CISE/IIS-Digital Society & Technology, under Grant No. 0222829. Overview.

ocean-ortega
Télécharger la présentation

An Ontology for Agent-based Modeling and Simulation

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. An Ontology for Agent-based Modeling and Simulation Scott Christley, Xiaorong Xiang, Greg Madey Dept. of Computer Science and Engineering University of Notre Dame Supported in part by National Science Foundation, CISE/IIS-Digital Society & Technology, under Grant No. 0222829 Scott Christley, An Ontology for Agent-based Modeling and Simulation

  2. Overview • Motivation • Ontology • Agent-based Modeling and Simulation • Ontological Reasoning • Inference and Automation • Future Work Scott Christley, An Ontology for Agent-based Modeling and Simulation

  3. Motivation • Formalize the process, take a knowledge-based approach to simulation • Underlying assumptions in the model can manifest into artifacts in the simulation results, so formalizing the model makes them more explicit. • Reasoning can provide the modeler with feedback/methodology direction and can automate some of the tasks. • Agent-directed Simulation, computer assisted problem-solving, support tools Scott Christley, An Ontology for Agent-based Modeling and Simulation

  4. Ontology • Establish common vocabulary • Knowledge-based representation of concepts and relationships between those concepts • Allows for automated reasoning • Ontology Web Language (OWL) using Protégé editor. Scott Christley, An Ontology for Agent-based Modeling and Simulation

  5. Agent-based Model • Agents interacting with other agents and its environment within a spatial structure. • Agent is the conceptual unit of interest, defines a boundary between what is internal to the agent versus what is external. • Environment represents global and/or local state information that is external to agents. Abstracts entities in the model that we do not want to explicitly represent as agents. • Spatial structure has implicit notions of locality based upon measures, holds only state specific to those measures. • Multiple agents, environments, and spaces are valid concepts in the model. Scott Christley, An Ontology for Agent-based Modeling and Simulation

  6. Modeling and Simulation Process • Scientific inquiry • ConceptualModel; verbal, abstract model that states the theory or hypotheses for the agent-based model and the goals and objectives of the simulation. • CommunicativeModel; domain-specific ontology that fits within the general agent-based ontology. • ProgrammedModel; software representation of the Communicative Model, in particular an AgentBasedSimulation. • ExperimentalModel; design of experiments using the Programmed Model to produce SimulationData. Scott Christley, An Ontology for Agent-based Modeling and Simulation

  7. Scott Christley, An Ontology for Agent-based Modeling and Simulation

  8. Ontological Reasoning • Reasoning with our knowledge base about the process of modeling and simulation, and ultimately about the domain of interest. • Inference of model assumptions and parameters. • Check model consistency • Automated software programming • Automated design and execution of experiments • Automated validation of simulation results Scott Christley, An Ontology for Agent-based Modeling and Simulation

  9. Inferred Assumptions • Data Assumptions, data collection and analysis • InputModeling, representation of empirical data with an analytic distribution. • Structural Assumptions, model composition and representation • Ontology representation, CommunicativeModel • Software representation, ProgrammedModel • Reasoner can extract these assumptions from the properties and relationships of the concepts in the ontology. • Knowledge of these assumptions implies possible validation tests. Scott Christley, An Ontology for Agent-based Modeling and Simulation

  10. Inferred Parameters • Parameter • Input to the model that is persistent through the whole simulation. • InitialCondition • Values for state variables for just the start of the simulation. • Value assigned as part of ExperimentalModel, may be obtained through ParameterEstimation or attached to a DataSource like a RNG. • Reasoner can infer parameters from the knowledge base; logical query on the properties of agents, environment, and space. Scott Christley, An Ontology for Agent-based Modeling and Simulation

  11. Automated Programming • CommunicativeModel -> ProgrammedModel • Declarative versus Procedural debate • Declarative, “what to do”, high-level specification language, ontology and reasoning • Procedural, “how to do it”, standard programming language • Intermediate approach • High-level structure generated automatically • Detailed behaviors implemented procedurally by modeler. • Ontological representation of programming constructs • Java RePast, ObjC Swarm, classes, instance variables, methods Scott Christley, An Ontology for Agent-based Modeling and Simulation

  12. Model Composition • Composition of multiple, separate CommunicativeModels into a single ProgrammedModel. • Two CommunicativeModels represent same real world phenomena or different real world phenomena. • Merge of domain-specific ontologies requires knowledge of the semantics of interaction. • Semantics in both models • Semantics in one model • Semantics in neither model • Structural equivalence Scott Christley, An Ontology for Agent-based Modeling and Simulation

  13. Automated Experiments • Iterative process of hypothesized model, experimentation, validation and analysis, leading back to changes to the model. • Experimental design as manipulation of the ProgrammedModel • Value assignment for parameters • Enable/disable Actions for the agents, environment, and space • Different implementations • Reasoning about the value of information, produce experiments to test assumptions. Scott Christley, An Ontology for Agent-based Modeling and Simulation

  14. Automated Validation • Compare the implications of a model against the real world phenomena but also applies to all data/structural assumptions. • Subjective tests like FaceValidity • Statistical tests like GoodnessOfFit, TestOfMeans, ConfidenceIntervals, TimeSeriesAnalysis • ModelToModelComparison • Same CommunicativeModel different Programmed Model • Different CommunicativeModel • Reasoner has knowledge about statistical tests and can apply them to the experimental results. Scott Christley, An Ontology for Agent-based Modeling and Simulation

  15. Emergence • Local interaction rules that produce global structure or behavior. • Recognition; you know it when you see it • Conditions for implication and existence Scott Christley, An Ontology for Agent-based Modeling and Simulation

  16. Future Work • Uncertainty and probabilistic reasoning • Learning • Tools • Integration with Agent-based toolkits, visualization, and statistical packages • eScience, Web Services, model composition Scott Christley, An Ontology for Agent-based Modeling and Simulation

More Related