The Research Of Learning Resource Searching System Based On Agent Technology

Learning Resource Searching System is an important part of Network based Education. The traditional Searching Systems usually employ "Keywords" method to match all the related resources. Using this method, the learner can quickly get a large number of results, but it still has some defects: it ignores the different backgrounds and interests of learners; it can't present results by category; the learners can't share their searching experience and satisfied resources in the system.To solving these problems, in this paper, the author designed an Intelligent Learning Resource Searching System based on Agent technology and Search Engine technology. This new system created Learner Information Model for each learner, which can be used to store and express learner's interests and other characters, these characters can impact the searching results; To make the learning resources list by different category, a Knowledge Base Model is created with the functionality of mapping keywords to corresponding fields and branch subjects. Agent has the traits of Autonomy and Cooperation. As a Multi-Agent System, besides the timely keywords matching, this system designs 3 new searching methods: "Self-Help Search", "Mutual-Help Search" and "High-Frequency-Help Search". "Self-Help Search" enables learners make the most of historical searching information. "Mutual-Help Search" enables learners share their searching experience and satisfied results automatically by agent cooperation. "High-Frequency-Help Search" proactively processes the requests with high searching frequency, so as to enhance the efficiency of system. In this paper, the author also designs a "Black Board" model to make agents interchange messages and communicate with each others.This paper use an Agent Oriented methodology named Gaia to analyse and design this multi-agent system, and by comparing different CASE tools and environments, the author chose JACK to program the prototype system.