Utilizing Particle Swarm Optimizations on student grouping problem

Particle Swarm Optimization (PSO), a meta-heuristic algorithm proposed by Eberhart and Kennedy for solving computational optimization problems, has been applied to several fields including scheduling problem, electromagnetism, constructing student groups in education, etc. Meanwhile, various versions of PSO with superior improvements have been brought up gradually with the goal of eschewing premature convergence. In this paper, couples of PSO for solving the student composition problem were examined. In addition, for the sake of verifying that optimal solutions are acquired accurately and easily compared with the PSOs, an integer programming optimizer named CPLEX was took advantage of. The results indicated that the PSOs did realize a better performance in fitness acquisition and time processing than CPLEX no matter both in student grouping of small cases and the real data of 61 students.