Toward effective group formation in computer-supported collaborative learning

Abstract

Group formation task as a starting point for computer-supported collaborative learning plays a key role in achieving pedagogical goals. Various approaches have been reported in the literature to address this problem, but none have offered an optimal solution. In this research, an online learning environment was modeled as a weighted undirected complete graph in which each learner was implied as a node and the relationship between them was denoted as a weighted arc. The weight of each link indicated the similarity degree between the corresponding individuals. The similarity between two students was measured as the mean of their absolute interest levels. The graph was also represented through a symmetric adjacency matrix. Then, a novel binary integer programming formulation was proposed to model the group formation problem and optimally assign each learner to the most appropriate group. The method was utilized to divide an online class of 32 learners into 8 groups of size 4. Findings indicated that the suggested model was successful in optimally solving the problem in 20.53 seconds, on average. The performance of the method was also compared with a modified version of K-means clustering algorithm. Although, the running time of the suggested technique was not as good as the clustering algorithm, it generated better outcomes in theory and in practice.