160 likes | 317 Views
A MULTI AGENT SUPPORT SYSTEM FOR MODELING, CONTROL ,DESIGN AND SIMULATION. utrrbudi, utrrfran, utrrtung@savba, kasanicky@neuron.tuke.sk Intitute of Informatics SAS, Dúbravská cesta 9, 841 08 Bratislava. Structure of presentation. Architecture of the system The system’s interfaces
E N D
A MULTI AGENT SUPPORT SYSTEM FOR MODELING, CONTROL ,DESIGN AND SIMULATION utrrbudi, utrrfran, utrrtung@savba, kasanicky@neuron.tuke.sk Intitute of Informatics SAS, Dúbravská cesta 9, 841 08 Bratislava
Structure of presentation • Architecture of the system • The system’s interfaces • Ontology and structured database creation • Example of similarity computation
Architecture of the system • Generic block • User agent • Modeling agent • Control agent • Simulation agent • Monitoring agent
Architecture of the system MODELLING AGENT MODELLING CONTROL AGENT CONTROL STRUCTURED DATABASE USER AGENT SIMULATION AGENT USER SIMULATION MONITORING AGENT
Architecture of the system • Basically agents work according the following scheme: { What user? What system? Search for appropriate methods Identify available tools Select the best tool (CBR method) Create model }
The system’s interfaces • User’s iterface • Introductory questionnaire • Second level questionnaire • Third level questionnaire • Last level questionnaire • Interface between Generic Block and external application • Interface between the system and the structured database
Ontology and structured database creation • Ontology is a very useful framework for organizing knowledge • Protégé (Noy, 2000), Stanford Medical Informatics at Stanford University • OntoEdit (Sure, 2002) AIFB Karlsruhe University
Ontology and structured database creation • An example of ontology in • Protégé
Ontology and structured database creation • Case library • Each case is characterized by a set of attributes. A is a set of attributes If an attribute is not related to the concrete case, a value is set to zero
Ontology and structured database creation • Set of solutions • A Case Base is defined as a space of case descriptions and case solutions
Ontology and structured database creation • Let a current case is defined as follows • The agent looks for similar cases to a current one by using a method for similarity assessment • Function in this project is method of weighed attributes. • Two kind of weight is used • user specific weights are given through user’s questionnaire • case specific weights are part of structured database
Example of similarity computation Similarity function can have two values Calculation of the case similarity
Example of similarity computation Similarities are computed as it was described So cases 1 and 2 are more similar as 1 and 3
Conclusion • System use Case Base reasoning • Structured database is build from tree independent block • The future work is concentrated on testing tuning CBR methods