Dynamic Re-Composition of Learning Groups Using PSO-Based Algorithms

In collaborative learning contexts, the problem of automatically forming effective learning groups gets considerably complex with larger class sizes, e.g. in MOOCs. Additionally, group dynamics caused by high dropout rates currently observable on online open course platforms poses challenges to learning group formation strategies. To address these problems, this paper presents PSO-based algorithms to dynamically re-compose learning groups. In addition to static grouping criteria (such as MBTI personality types), the algorithms take into account factors of the group success rate and group satisfaction during re-composition. We carried out simulations based on randomly generated sample data. The experimental results show that the proposed approach performs better than traditional exhaustive or random methods.