A Multi-Agent Based System for Activity Configuration and Personalization in a Pervasive Learning Framework

We have begun to define and explore an agent-based distributed computational environment to support pervasive learning. In this article, we describe the system for configuring learning scenarios enhanced with an educational networking infrastructure including different kinds of devices. Students carry out indoor and outdoor activities with learning mediation tools either using classroom computers or mobiles devices, such as PDAs, to be used in labs or in field trips. All these activities are interwoven and supported in a seamless way by a learning repository, which is the backbone of the infrastructure. The configuration system is ´intelligent´ enough to be able to deploy and initialize the tools for the student´s activity taking into account the context of the user and the activity-localization, current working device, tools available- and personalize it according to the situated requirements. The set of specialist agents and the models needed are subsequently described, together with the benefits of the AI techniques for coping this problem