In this paper, a stochastic coupling network topology is introduced in a particle swarm optimizer network. In the proposed method, each sub-swarm group of particles stochastically determines whether to communicate to the other sub-swarm groups. By controlling the communication rate, the diversity of the obtained solutions is generated, and the number of interprocess communications is reduced. Resulting of them, the search performances can be improved and the parallel-distributed computing with high parallelism can be realized. The proposed method is evaluated by computer simulations and the effectiveness of the proposed method is discussed.
展开▼