Towards a multi-agent framework for intelligent tutoring systems

Intelligent tutoring systems have been used as valuable educational tools for their adaptability in the computerized pedagogical process. They can vary tremendously according to the provided functions and architecture. In this paper we propose a generic framework for intelligent tutoring systems that learning profile analysis and personalized learning paths are performed through a community of intelligent agents.The basic categories of agents defined in the framework are learner Agents, instructor Agents, and coordinator Agent. The learner agents monitor and evaluate the individual learning processes to generate personalized course paths. The instructor agents provide student profile analysis for teachers to perform necessary course adjustment.A message brokering architecture is utilized in the coordinator agent to provide a better mechanism for dealing with failures of agent communication. The agent modules and the process flow are also designed to accomplish the adaptive learning mechanism.