摘要： 分析并验证基于变惯性权重的粒子群优化(PSO)在粒子寻优过程中的有效性，论述类无标度网的特殊拓扑性质。将有向动态类无标度网作为粒子寻优邻域，提出一种基于变惯性权重及动态邻域的改进PSO算法。实验结果证明，与传统PSO算法相比，改进算法的寻优效果较好，可在一定程度上避免陷入局部最优。
关键词:
粒子群优化,
类无标度网,
惯性权重,
度分布,
邻域拓扑
Abstract: This paper analyzes and verifies the effectiveness of Particle Swarm Optimization(PSO) based on variety inertia weight in the particle optimization process, and discusses the special topological properties of scale-free like network. It uses the dynamic scale-free like network as the particle’s optimization neighborhood. It proposes an improved PSO algorithm based on variety inertia weight and dynamic neighborhood. Experimental results show that the improved algorithm performs better than the traditional PSO and may avoid falling into the local optimum instead.
Key words:
Particle Swarm Optimization(PSO),
scale-free like network,
inertia weight,
degree distribution,
neighborhood topology
中图分类号:
姚灿中, 杨建梅. 基于变惯性权重及动态邻域的改进PSO算法[J]. 计算机工程, 2011, 37(21): 20-22.
TAO Can-Zhong, YANG Jian-Mei. Improved PSO Algorithm Based on Variety Inertia Weight and Dynamic Neighborhood[J]. Computer Engineering, 2011, 37(21): 20-22.