基于自组织邻域结构的微粒群算法

Particle Swarm Optimization Based on Self-organizing Neighborhood

通过构建自组织邻域结构来保持群体多样性,以克服微粒群算法(PSO)易局部收敛的缺点。模拟动物群体趋利避害的行为选择机制,以微粒的适应值择优建立自组织邻域结构的连接。实验结果表明,基于自组织邻域结构的微粒群算法(SONPSO)优于微粒群算法、基于环形结构和动态环形结构的微粒群算法。

A Self-organizing neighborhood was constructed in order to maintain the diversity of population and overcome the defect of easy local convergence of PSO. The preferential attachment of fitness was established, which sim- ulates individuals that are willing to interact with better but not worse animal groups. Simulation results show that particle swarm optimization based on self-organizing ne zation, particle swarm optimization based on ring lattice tice. ighborhood （SONPSO）is superior to particle swarm optimi- and particle swarm optimization based on dynamic ring lattice.