Online Learning System Based On Multi-agent

This paper is based on the E-Learning System for stuff training in company, which is supported by Beijing talent developing project (No. 20061D0501600220). In order to overcome the traditional E-Learning Systems' shortcomings, this paper tries to improve the system in two aspects, learning content sharable and system intelligent.For learning content sharable, we adopt the SCORM (Shareable Course Object Reference Model) standard to our system. With SCORM, the learning content can be sharable and reusable. In this paper, we design a SCORM courseware for chapter 4 of "Java Program Design", and edit two different teaching strategies. So our system can show different learning content for different learner.For system intelligent, we use Agent technique. This paper gives two Agents: UIA (User Interface Agent) and CCA (Content Collect Agent). By using relation metadata of SCORM we put forward an algorithm for content collecting and pushing.In addition, we develop E-learning prototype system Based on SCORM, and pay attention to the realization of course importing, course analysing and course viewing.