节点文献
基于Agent的信息检索系统的若干问题研究
Study on Some Issues of Information Retrieve System Based on Agent
【作者】 汤妹军;
【导师】 蒋建国;
【作者基本信息】 合肥工业大学 , 信号与信息处理, 2003, 硕士
【摘要】 Agent技术是近年来研究的一个热点,它是从人工智能的一个分支DAI(Distributed Artificial Intelligence)发展起来的。Agent技术和检索系统的结合是近年来检索系统研究的重点内容,Agent以其自身特性弥补了传统检索系统的搜索精度不高、过分依赖于网络连接和返回冗余信息过分等缺点。由于是一门全新的研究领域,这种基于Agent的检索系统也会产生一系列问题。本文总结了检索系统发展的历史,将Agent技术引入检索系统,并对基于Agent的检索系统的若干关键问题进行了研究: 1.通过对传统检索系统搜索算法的分析,提出了一种基于深度受限的动态调整算法,并针对搜索Agent的两种搜索方式,分别提出了深度受限的顺序搜索算法和深度受限的并行广度搜索算法。 2.通过将Agent任务进行分解,以一种分层结构来组织任务Agent和搜索Agent之间的协同,从而提出了分层协同的思想;并就如何命名Agent的问题,提出了多层命名的思想。 3.提出了建立一个基于反应Agent的模型来避免对网址的重复性访问,并由此反应Agent来对在网站上生成的搜索Agent进行管理,并以此模型讨论了在检索过程中可能出现的问题及一些解决方案。 最后指出了本文的不足并对将来的基于Agent的检索系统作了展望。
【Abstract】 Agent technology developed from DAI (Distributed Artificial Intelligence) which is a branch of Artificial Intelligence and has been a hotspot for research. The integration of Agent technology and retrieve system has been the emphasis on the research field of retrieve system recently. Agent made up for the shortcomings in traditional retrieve system, such as poor accuracy, relying on internet , many useless information and so on. However, the retrieve system based on Agent has also brought many new questions because it is a completely new research filed. This paper summed up the development history of the retrieve system and applied the Agent technology into the retrieve system. Some key issues of the retrieve system based on Agent have been studied in this paper:We analyzed traditional retrieve system algorithm and brought up dynamically adjusting algorithm based on limited depth . Because retrieve agent has two retrieve ways, we individually brought up sequence retrieve algorithm base on limited depth and parallel breadth retrieve algorithm based on limited depth.We brought up a multi-layer coordination in which task is decomposed into many subtasks and the coordination among task agents is organized in multiple layers way. Furthermore, we brought up multi-layer naming in order to identify agent.We brought up a model based on reactive agent to avoid repeatedly visiting the same website. We discussed some questions that could occur in the retrieving process and the corresponding solvable strategies on the basis of this model.In the last, we pointed out the shortcomings in this paper and looked into the future of the retrieve system based on agent.
【Key words】 Retrieve system; Agent dynamically adjusting algorithm based on limited depth; Sequence retrieve algorithm based on limited depth; Parallel breadth retrieve algorithm based on limited depth; Multi-layer coordination; Multi-layer naming; Reactive agent;
- 【网络出版投稿人】 合肥工业大学 【网络出版年期】2003年 03期
- 【分类号】TP391.1
- 【被引频次】2
- 【下载频次】173