2015 年 31 巻 1 号 p. 10-14
The group organization problem is a combinatorial optimization problem. This paper attempts to explore the solutions to the problems. The following case is treated : Students are to be allocated to groups of which the subjects are shown in advance. The conditions are that they ought to be assigned to their groups as preferable as possible, and that their abilities should be almost in equilibrium among groups. It is a multi-objective problem, which is rather difficult to solve manually. The conception of Pareto optimal solutions is used to treat multi-objective problems. One of the techniques to solve that kind of problems is MOPSO (Multi-Objective Particle Swarm Optimization). This study tries to search the quasi-optimal solutions to the group organization problem, using MOPSO. The effect of applying MOPSO originally intended for continuous optimization problems to the above case is considered.