A Frame Work for Dynamic Matchmaking between Messages and Services in Multi-Agent Systems Muhammed Al-Muhammed
Motivation Agents cooperate to achieve their goals. Cooperation needs communication. Current assumptions: 1- they share ontologies, * an ontology is a formal definition of relationships between concepts in a domain. 2- speak the same language, 3- pre-agree on a message format.
Limitations This research tackles only query messages. Deals only with messages that completely match services.
Agent1 LO1 Services Agent2 LO2 Services Global ontology Translations (MMS) Communication Matchmaking System The proposed matchmaking system
Local OntologyGlobal ontology WordSellingValueWordPrice Syn.PriceSyn.Value, SellingValue TypeIntegerTypeReal Values $100ValuesPrice value recognizer DomainComputerDomainComputer Rel.SellingValue ISFOR Mem. Rel.Price ISFOR Mem. ?
Agent1 Local O.1 Services Agent2 Local O.2 Services Global O. Translations MMS I need info about PCs Input:LowPrice=SP1 0, HighPrice=Sp 20 Output: String Make, String Model,int Price Constraint:None Communication Service In C.F =? MMS Global.O Translations S1 S2. Si. Sn MSG in Common Frame (C.F)
The ServiceThe Message serviceName GivePcInformationServiceNameNone actionTypeNoneActionTypeNone ServiceTypeQueryServiceTypeQuery InputReal LowPrice, Real HighPrice Inputinteger HighPrice=SP 20 Integer LowPrice=SP 10, OutputString Model, Real Price, String Make, outputString Make String Model, Integer Price, String Make InConstraintsLowPrice>0 $ LowPrice<HighPrice InConstraintsNone OutConstraints Output sorted(Price)OutConstraintsNone
Agent1 LO1 Services Agent2 LO2 Services Global ontology Translations MMS Communication MSG in C.F Service In C.F = No MMS Global.O Translations S1 S2. Si. Sn I need info about PCs Input: Output:.. Constraint : =Yes Si
Contributions Enhance the interoperability Simplify agents’ communication Simplify the developers’ task