一种基于加权有向拓扑的改进粒子群算法

研究粒子群优化算法(PSO)的拓扑结构和信息流动,以提高算法性能是PSO的一个有意义的研究方向。RuiMendes等人提出的全联通型算法(FIPSO),其拓扑结构本质上是加权无向图,两个邻接点之间的相互影响是对等的,与社会人际网络的真实情况不符。提出了一种改进型算法,重新构造了加权函数,体现了粒子之间影响的不平衡性。仿真结果显示:该改进算法对收敛速度和稳定性均有非常好的改善。

A Modified Particle Swarm Optimization Based on Directional Weighting Topology

It makes sense to search on the PSO from its topology and information flow in order to improve its performance.The FIPSO proposed by Rui Mendes etc.,whose topology is weighting undirected that influence between two adjacent particles is equivalent,doesn't accord with interpersonal relations in real societal networks.Weighting functions are reconstructed in the paper which realizes asymmetric influence between particles.The simulation results have shown that the performance of modified algorithm is far better,faster and more stabile in convergence.