Grouping learners for cooperative learning: Grouping genetic algorithm approach

The grouping genetic algorithm (GGA) is an effective metaheuristic algorithm that solves grouping problems by exploiting their group structures. Rather than working on individual genes (or items), the GGA operators, namely group crossover, group mutation, and group inversion, are designed to work on groups of genes, which gives the algorithm more effective and efficient search characteristics. By preserving the structure of the groups, which forms the basic building blocks of the algorithm, group similarity is maintained and improved with minimal disruptions, unlike when genetic operators work on single items. Based on the group encoding scheme, GGA encodes the cooperative learners as items that are assigned to groups, according tomutual scores.The population of chromosomes is iteratively evolved over generations, exploring new regions of the solution space via group crossover operation and exploiting visited regions through group mutation. The versions of the crossover, mutation, and inversion operators used in this research are efficient and easy to apply. To enhance the search process, the inversion operator helps to dynamically maintain population diversity at acceptable levels, until final convergence as the iterations progress toward the termination condition. Computational experiments show that the algorithm is efficient and effective, providing optimal or near-optimal solutions, even over large-scale problems. Further research application on closely related grouping problems, such as team formation and learners' grouping problem in a fuzzy environment, is highly recommended.