在为搜索服务的智能Agent(IARS)模型的基础上,进一步探讨了利用多Agent(Multi-Agent)概念完善该文系统的理论基础和方法。由于多智能Agent系统中的任一Agent都能在任一时刻参加或离开系统的检索工作,因此,由单个Agent所获得的知识应能在系统各Agent间通信和传输。为此,在描述了总体设计的基础上,文章重点研究了为获取用户所需的信息,多Agent系统应在知识的通信和传输中采取的方法。 Based on the IARS model, this paper further explores the theoretical basis and method of using Multi-Agent to improve the system. Since any agent in the multi-agent system can join or leave the retrieval system at any time, the knowledge acquired by a single agent should be able to communicate and transmit among the agents. Therefore, on the basis of describing the overall design, the article focuses on the method that multi-agent system should take in the communication and transmission of knowledge in order to obtain the information needed by users.