Research And Development Of Chinese Meta Search Engine Based On Agent

Since the birth of WWW in 1991, the technologies and applications of WWW have gone a tremendous increase in the world, and affected almost all fields of work and daily life. With so many WWW servers and so much Web information on Internet, how to effectively deal with the information searching becomes more and more important than before. As part of the key project of National 9th Five Year Plan, The pre- and post- process ing technology for Web information search, and also the task of developing the Web information search system for the olden Bridge?project, this paper presents the research of the meta search engine, by using agent technology. The main contains of this paper involves the following aspects. (1)By researching the characters of Internet information and the tendency of search engine development, this paper describes the principle of Directory Search Engine, Robot Search Engine, Meta Search Engine, and analyzes the advantage and disadvantage of the current search engines. (2)Introduce the concept of Agent in Al, and define it exactly. Moreover, carry through a further study on its constitution, function and basic structure. Based on agent, research the characteristic, classification and interaction of multi agent cooperation system. (3) Design and implement a Web information retrieval prototype system--- WebHunter based on MAS(Multi Agent System). The basic components of WebHunter are Agents. This system structure contains three layers: pre-processing layer of information retrieval, information retrieval agent layer and post-processing layer of information retrieval. (4)Analyze the syntax of each search engine in detail, extract the common format from the most search engines, define the syntax of WebHunter according to SOIF(Summary Object Interchange Format). (5)Propose an approach of selecting search engines in meta search engine by using multi agent cooperation. With the cooperation of multi agents and the support of information base, WebHunter can avoid sending user query to those search engines that returned with poor query result, so that resource consumption can be reduced. (6)Adopt the parallel search engine agents which can interact with search engines in order to decrease the response time of WebHunter. (7)Describe how to rank the query results of the current search system, present a simple, effective ranking method in order to improve the system efficiency...