Abstract
In this paper, we describe the design and development of a fuzzy agent-based system for an elarning environment that is capable of effective delivery of personalized courseware to elearners. By personalization, we mean that two different persons registered at the same e-course with comparable learning paces will have two different behaviors from the system. For so doing, we want to address a specific issue. We mean that we are given a set of courses displayed on the Web, a set of disciplines to which these courses belong, a number of visiting times any prospective elearner makes to the Web site. It is desired to find the fuzzy membership of this specific elearner to the given disciplines, and a profile for this elearner capable of predicting his future behavior i.e. what this elearner is planning to do without letting him explicitly express his query. In order to address this issue, we propose to use fuzzy logic to express gradual and imprecise membership and agent paradigm to express autonomy for mapping percepts coming from the environment into correct actions under the control of the logic embodied within it. Furthermore, we use state of the art advanced tools for implementation based on two versions. The first version, a prototype, is implemented using Matlab (TM). A second version using MicroSoft (TM) NET Technology, such as C#.NET (TM). This results in the development of an advanced fuzzy agent-based system for elearner profile construction capable of predicting human behavior during the process of learning while offering an intelligent Graphical User Interface (GUI).