基于环形结构带缓存器模型的并行微粒群算法

Parallel Particle Swarm Optimization Based on Ring Structure Model with Buffer Storage

受并行遗传算法的启发,文章设计和实现了一种基于环形结构带缓存器模型的并行微粒群算法。它基于一种单向环结构的拓扑连接,可以保证优良粒子在子种群间的扩散,丰富种群的多样性。仿真实验的结果表明,该并行算法不仅有效地提高了求解效率,而且在一定程度上改善了早熟现象,算法的各项性能与微粒群算法相比有了很大提高。

A novel algorithm of parallel particle swarm optimization based on ring structure model with Buffer storage is designed and implemented. It is enlightened parallel genetic algorithm. This algorithm uses unilateral-ring structure, which can assure good particles diffuse between sub population and rich the diversity of swarm. The experimental results show that not only the solving efficiency is raised but also the restraining premature convergence is improved. Comparing with standard particle swarm optimization （PSO）, the performances of the proposed algorithm are greatly improved.