Educational Group Formation Problem Resolution Based on an Improved Swarm Particle Optimization Using Fuzzy Knowledge

In educational context, instructors usually partition students into collaborative learning teams to perform collaborative learning tasks. Indeed, one of the grouping criteria most utilized by instructors is based on the students’ roles and on forming similar teams according to the roles of their members which is costly and complex. In this paper, we address the optimization problem of forming automatically learning teams by minimizing the knowledge-difference cost among formed teams. The knowledge index of each group depends on the Belbin roles of their students’ members in the form of a sum of students’ fuzzy rating indexes. The proposed algorithm is called improved particle swarm optimization with multi-parent order crossover (IPSOMPOX). The multi-parent order crossover is used in IPSOMPOX in order to investigate new solutions in the search space and to accelerate the convergence of the proposed algorithm to the best global solution. To evaluate the performance of the proposed algorithm, we apply it on several different experiments with different numbers of teams and students. The proposed algorithm is compared with the standard PSO and has proved better performance are shown in our results.