Genetic Algorithms for Student Grouping Problems

In this paper, we cope with problems of forming optimal groups of students in a class for learning collaboratively. The objective of the problems is to maximize the marks of the students through the collaborative learning. We propose two genetic algorithms (GAs) in order to search for the optimal solution of the problems. In the first GA, only promising solution space is searched by using an effective individual description. In the second GA, new individuals are generated by using an effective crossover operation in which they can inherit good groups from the current individuals. The performance of the proposed GAs is evaluated through numerical experiments in which actual data on students are used. We also show actual results of learning in the case where students learn collaboratively according to groups determined by the proposed GAs.