An Adaptive E-Learning Environment Architecture Based on Agents and Artifacts Metamodel

Abstract

In this paper, we proposed an adaptive e-learning environment architecture that supports personalization by utilizing Agents and Artifacts (A&A) Metamodel. A&A Metamodel focuses on environment modeling in multi agent system (MAS) design and models entities in agents' environments with artifacts as first class entities like the agents. From the perspective of MAS based e-learning systems, learner models and learning resources are part of the environment of the agents and agents interact with them constantly. Thus, we proposed an e-learning architecture that focuses on environment abstraction and models access to different learner models and learning resources with artifacts to support personalization. In MAS based e-learning systems with the same functionality, specific agents are responsible for modeling learner information and retrieving learning resources. However, in the proposed approach, by exploiting A&A Metamodel, this operations are performed by artifacts to provide a more flexible and scalable solution. The proposed adaptive e-learning environment architecture is developed as a prototype with CArtAgO framework. A MAS based e-learning system is also implemented with Jason agent framework as a case study exploiting the developed environment. To evaluate the proposed approach, learning objects (LOs) for Logic Design course are developed and learners are modeled according to their learning styles by using a learner ontology. Finally, we presented results of the evaluation and discussed current limitations and future work directions.