Designing and modeling of a multi-agent adaptive learning system (MAALS) using incremental hybrid case-based reasoning (IHCBR)

Abstract

Several researches in the field of adaptive learning systems has developed systems and techniques to guide the learner and reduce cognitive overload, making learning adaptation essential to better understand preferences, the constraints and learning habits of the learner. Thus, it is particularly advisable to propose online learning systems that are able to collect and detect information describing the learning process in an automatic and deductive way, and to rely on this information to follow the learner in real time and offer him training according to his dynamic learning pace. This article proposes a multi-agent adaptive learning system to make a real decision based on a current learning situation. This decision will be made by performing a hypride cycle of the Case-Based Reasonning approach in order to follow the learner and provide him with an individualized learning path according to Felder Silverman learning style model and his learning traces to predict his future learning status. To ensure this decision, we assign at each stage of the Incremental Hybrid Case-Based Reasoning at least one active agent performing a particular task and a broker agent that collaborates between the different agents in the system.