摘要
以人工智能领域分布式任务处理中的黑板模型作为Agent通信模型的基础 ,以能够实现单播、组播、广播为度量标准 ,通过 5个假设对通信安全需求进行分析 ,构造了Agent通信单元模型、树形拓扑多Agent通信模型 ,提出了在Internet环境中不同MAS间Agent通信实现方法 .在通信语言环境的分析中 ,通过综合考虑具有不同通信协议、不同通信内容格式、不同实体概念以及不同通信服务质量和通信安全等问题 ,提出了异构MAS系统间的通信语言环境的分层模型 .研究结果表明 ,在Agent通信协议中设置通信服务质量的功能 。
The Agent communication model is based on blackboard architecture that comes from distributing problem solved in artificial intelligence. With the premises of five hypothesizes and the analysis of the targets or metrics of Agent communication, which are unicast, multicast, broadcast, anycast and requirements of communication security among Agents, the authors built Agent communication unit model, tree topology multi Agent communication model(TTMAS) and present an approach to making the communication between different MAS in unsecured internet environment. Then, with the discussion of communication language, which has different communication protocols, dissimilar content formats, varied ontology, different QoS for network and network security requirement, the authors offer a layer model for Agent communication language environment. At the same time, we give three future directions of our research, which are realizing mechanism of blackboard architecture in concrete computing environment, intelligent and dynamic config mechanism for distributed tree topology MAS and Agent communication protocols based on TTMAS. MAS is used in many fields.
出处
《中南工业大学学报》
CSCD
北大核心
2002年第2期209-213,共5页
Journal of Central South University of Technology(Natural Science)
基金
国家教育部重点课题资助项目 ([2 0 0 0 ] 15 6号 )