抄録
This paper studies the collision particle swarm optimizer (CPSO) for discrete multi-solution problems (DMSP). The dynamics of the CPSO is governed by a deterministic difference equation on a discrete search space. It is useful from viewpoints of reproducibility and implementation. The CPSO can use inter-particle collision that can be effective to avoid trapping into partial/local solutions. The CPSO is applied to a DMSP based on an exploring problem of periodic points of the Hénon map. Performing elementary numerical experiments, the basic capability of the CPSO is investigated. This application is the first step to develop PSO-based analysis tool of nonlinear dynamical systems.
