一个多Agent结构的个性化信息搜索系统设计  

Design of Personalized Information Search System Based on Multi-Agent

作　　者：刘声田[1] 翟玉庆[2] 

机构地区：[1]山东广播电视大学,山东济南250014 [2]东南大学,江苏南京210096

出　　处：《山东电大学报》2009年第2期18-22,共5页Journal of Shandong TV University

基　　金：山东省高等学校优秀青年教师国内访问学者项目经费资助

摘　　要：设计一个基于多Agent个性化信息搜索系统PISMA-Personalized Information Search System Base on Multi-Agent。系统包含5个Agent和其它2个组件,在设计上强调了获取信息的个性化和时效性,具有智能搜索和主动搜索特点。论文先介绍了系统结构模型,重点说明了5个Agent的设计,其中的移动A-gent通信采用了我们设计的超时即返算法RIOT-Return Immediately Out Time;然后从用户模型、系统学习模型和个性化信息模型三个方面讨论了个性化搜索模型实现;最后进行了总结和对未来工作的展望。In this paper, we design a personalized search system named PISMA （Personalized Information Search System Base on Multi--Agent）, which consists of 5 agents and 2 other components. The system emphasizes information searching personally and timely and realizes information retrieval actively and intelligently. Firstly, we present a model of the system, in which we probe all the 5 agents and present a novel algorithm named RIOT（Return Immediately Out Time） ; secondly, we discuss 3 models which include User_Model, System_Study_Model and Personalize_Information_Model and Mobile_A- gent_Communication_Model to implement personalized information searching; lastly, we share our conclusions and plan for future work.

关 键 词：AGENT 个性化信息搜索 智能检索 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]