An Intelligent Architecture for Generating Evolutionary Personalized Learning Paths Based on Learner Profiles

In this paper we present a solution that is a continuation of work done in the field of adaptive educational systems. It is an approach oriented objectives based on ontologies, multi-agent system and Bayesian networks to generate dynamically personalized learning paths. The dynamic aspect is essential for each learning session in the context of this solution, because the learning paths meet the objectives formulated by the learners will be generated to measure and after formulation the specific request. The operation consists of searching, filtering and composing dynamically hypermedia units of learning responding to the learner profile. For structuring and modeling information managed by our architecture, we used ontologies of Semantic Web. We designed the ontology of learners using the standard IMS-LIP to represent the learner profiles, some fields are added to include, in the model of the learner, learning styles according to the model of Felder and Silverman. And for representation of resources we designed the ontology of resources based on the LOM standard. Furthermore, the architecture is divided into three layers; each layer is managed by a number of agents. Agents exploit the Bayesian model and ontologies to provide learners with personalized learning paths.