多主体会话策略研究

Research on the Conversation Policy for Multi-Agents

用交互协议对主体之间的交互进行抽象的研究已经好多年了。然而,主体间会话的效率很低,而且会话的语义冲突问题一直没有解决。本文分析了主体之间通信协议和会话策略的关系,利用引入了本体的Dooley图描述出了主体间会话策略,并通过实验对主体选择不同的会话策略做了比较,其意义在于提高主体之间消除语义冲突的会话的效率,降低会话成本。

Although lots of research on the interactive protocol for agents have been done, the conversation efficiencies among agents are very low. The semantics conflict problems of conversations have not been solved yet. By analyzing the relationship between the agent communication protocol and the conversation polices, this paper introduces ontology into the Dooley graph to describe the agent conversation policy. The paper also compares the different conversation policies that agents choose with experiments. This method can improve the efficiency, eliminate the semantic conflicts of agent conversations,and reduce the cost of conversation.