{"title":"Developing coherent multiagent systems using JAFMAS","authors":"Deepika Chauhan, A. D. Baker","doi":"10.1109/ICMAS.1998.699236","DOIUrl":null,"url":null,"abstract":"The Java-based Agent Framework for Multi-Agent Systems (JAFMAS) provides a generic methodology for developing speech-act based multiagent systems (MAS) along with a set of classes to support implementing these agents in Java. The methodology follows five stages: (i) agent identification, (ii) definition of each agent's conversations, (iii) determining the rules governing each agent's conversations, (iv) analyzing the coherency between all the conversations in the system, and (v) implementation. JAFMAS provides communication, linguistic and coordination support through sixteen Java classes. Communication support is provided for both directed and subject-based broadcast communication. Linguistic support is provided for speech-act (e.g. KQML) based communication languages. Coordination support comes from conceptualizing agent plans and their coordination as rule-based conversations represented by automata models.","PeriodicalId":244857,"journal":{"name":"Proceedings International Conference on Multi Agent Systems (Cat. No.98EX160)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1998-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"9","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Proceedings International Conference on Multi Agent Systems (Cat. No.98EX160)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/ICMAS.1998.699236","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 9
Abstract
The Java-based Agent Framework for Multi-Agent Systems (JAFMAS) provides a generic methodology for developing speech-act based multiagent systems (MAS) along with a set of classes to support implementing these agents in Java. The methodology follows five stages: (i) agent identification, (ii) definition of each agent's conversations, (iii) determining the rules governing each agent's conversations, (iv) analyzing the coherency between all the conversations in the system, and (v) implementation. JAFMAS provides communication, linguistic and coordination support through sixteen Java classes. Communication support is provided for both directed and subject-based broadcast communication. Linguistic support is provided for speech-act (e.g. KQML) based communication languages. Coordination support comes from conceptualizing agent plans and their coordination as rule-based conversations represented by automata models.