一种基于聚类的小生境微粒群算法

A Clustering-Based Niching Particle Swarm Optimization

  • 摘要: 在小生境微粒群算法中引入一种简单的聚类算法,替换了原算法中依赖于圆形拓扑领域的小生境产生方法,构建出一种基于聚类的小生境微粒群算法.该算法在对主微粒群进行1-best PSO寻优的同时对其中的微粒进行聚类,当聚类簇中的个体数目达到规定的子微粒群最小规模时形成一个小生境.用这种算法能够产生大小和形状不同的小生境,克服了NichePSO算法的不足.

     

    Abstract: A simple clustering algorithm is used in niching particle swarm optimization(PSO) to replace the circular topological field based method for creating niche,and a clustering-based niching PSO is constructed. In the new algorithm,l-best PSO and a clustering algorithm are processed synchronously.When the number of particles in a cluster reaches the defined least particle number of a subswarm,a niche is formed based on this cluster.This algorithm can overcome the disadvantages of NichePSO by forming niches with different sizes and shapes.

     

/

返回文章
返回