Designing and Managing a Real-Time Collaborative Learning Paths by a Multi-agents Platform

Abstract

In this paper, we present a new architecture of learning path generation and managing the learner’s progression in multi-users context. By “multi-users context”, we hint at real time collaboration between different learners in a hyper-dynamic world; moreover, the architecture allows us to take into account the preceding learners’ appreciations about pedagogical resources and learning path during the new learning path generation. The generation implies two steps: (i) building a graph of learning objects according to their prerequisites and according to the learner progression in her/his learning path, (ii) the recommendation of pedagogical resources associated to each graph node. Different criteria are proposed to select the relevant pedagogical resources in the scope of the learner’s profile, learner’s communities, and the appreciations of other learners about the available pedagogical resources. These criteria are used to maximize a fitness function for the pedagogical resources. The progression of the learner in her/his learning path and the communities are managed by the multi-agent system based on JADE to ensure the smooth organization of the training session.