Particle Swarm Optimization Containing Plural Swarms
Masaki Sugimoto, Taku Haraguchi, Haruna Matsushita, Yoshifumi Nishio · 2009
In this study, we proposes a modified particle swarm opti-mization (PSO) called plural PSO (PPSO) for bust-out the lo-cal optimum solution. The feature of PPSO is that the swarm of PPSO is not one but plural. The plural swarms share an information of the best position in each swarm. Except the swarm including the best particle in whole swarms, all the particles are repositioned to escape the local optima. We in-vestigate behaviors of PPSO and confirm its efficiency in mul-timodal functions. 1.