Open Multi-Agent Systems for Collaborative Web-Based Learning

Abstract

Web-based learning plays an important role in the modern teaching environment. Many Web-based tools are becoming available in this huge marketplace. Agent technology contributes substantially to this achievement. One of the fundamental problems facing both students and education services providers is how to locate and integrate these valuable services in such a dynamic environment. In this paper, I present mediator-based architecture to build open multi-agent applications for eLearning. An agent services description language is presented to enable services advertising and collaboration. The language exploits ontology of service domain, and provides the flexibility for developers to plug in any suitable constraint languages. Multiple matchmaking strategies based on agent service ontology are given to help agents finding appropriate service providers. The series of strategies consider various features of service providers, the nature of requirements, and more importantly the relationships among services.