Study Of Intelligent Search Engine Based On Multi-Agent Cooperation

Nowadays, with the great development of Internet, navigation service to information on Internet offered by search engine has become a very important net service. Search engine has been a necessary network tool for Internet users. Users can search and get information resource quickly from Internet on which there is a great deal of information. So many commercial search engines have become "doors of the network". However, the users' requirement to information search and the information resources on Internet are both more and more. Although the search engines which are designed in traditional model and used by many people can partly resolve the problem how users can locate the information resources on the internet, now they also show some limitations, such as ignoring users' interests, puzzling users for many information resources, discarding information, and they can't accord with the increasing users' demands to the intelligent and individual information search service completely.After analyzing the limitations of traditional search engines, I get a result that the traditional search engines have no intelligence and the ability to analyze and understand knowledge. When traditional search engines search the information, they only search the indexes created by some search model, and they can't learn users' interests and needs, distinguish the meanings of words in different fields, associate and extend different words that have the same meanings. The key to resolve these limitations is to combine the artificial intelligence technique that has been widely used in many fields with the search engine technique to make the search engine intelligent. So I propose that the intelligent search engine can be realized by the cooperation of Multi -Agent System (MAS ) composed by many intelligent Agents, and based on it, an Intelligent Assisted Translation Search Engine System Based on Multi-Agent Cooperation (IATSESMA) is designed. Then I describe the application that MAS completes the intelligent search engine based on IATSESMA that is looked as model. IATSESMA is an intelligent search engine which may stand out users' individual peculiarity, can learn, understand, extend and deal with knowledge, and includes the initiative search and Meta search. It combines intelligent Agent technique, search technique and translation technique very well. And it provides the Chinese and English bilingual web pages collected from Internet for users to search translated texts. The intelligence of search engine in IATSESMA is realized by the cooperation of MAS that is composed by many intelligent Agents. Recall rate and accuracy rate of the whole system are improved by making index search, initiative search and Meta search complement each other. At the same time, I use "User's Individual Database" to describe the users' individual peculiarity of search engine. After Learning Agent and User Agent in MAS analyze and learn the information about the users' individual peculiarity, the search engine system helps users to detail the search requires, distinguish words' meanings in different fields, associate and extend different words, merge information based on users' interests.