Adaptive course generation based on evolutionary algorithm

For increasing demands on e-learning, many topics ensue to provide learners a more personalized learning experience and more suitable course and learning objects which is called Adaptive Course Generation problem(ACG). This article addresses how to support users´ personalized learning resources based on evolutionary PBIL (Population Based Incremental Learning) algorithm. Here both the users´ preferences and learning resources´ intrinsic characteristics are considered here. Some kinds of learning preferences are discussed briefly. Basic courses, itinerary courses and compulsory courses are also taken as important constraints for ACG. Moreover the granularity is discussed from both the point of experts and the point of the learning objects´ attributes. The above are modeled simply as a Constraint Satisfaction Problem (CSP). The objective function is to minimize the penalty function designed to evaluate the sequencing. The designed program is applied to Visual Basic course e-learning teaching with some quantity of freshmen. And it is tested on true teaching data and the statistics shows good validity, stability performance and it is popular with the participants.