A generic architecture for ITS based on a multi-agent approach

Abstract

This paper present a generic architecture for Intelligent Tutoring Systems. First, learning is described with three hierarchic level: curriculum, lesson and problem. We describe an agent model to represent knowledge at problem level. This agent model contain a definition of intern archictecture and a description of the different algorithms for collaboration between agents. After we see that this model is an autosimilar one. We can use the same model for different level. Multi-agent systems can be see as primitif agents in the upper level and a multi-agent can be described with the same model as its components.