Abstract:
Aiming at problem that Particle Swarm Optimization(PSO) algorithm falls into local optimum easily, this paper presents a PSO algorithm based on sub-region. It makes the search space some sub-region, uses the PSO algorithm to optimize in each region, compares these sub- region global optimums and finds out the search space global optimums. Results compared with standard PSO algorithm and adaptive mutation PSO algorithm show that this algorithm can reduce the probability of optimizing which falls into local optimum, and it has strong optimization ability.
Key words:
Particle Swarm Optimization(PSO) algorithm,
local optimum,
global optimum,
sub-region
摘要: 针对粒子群优化(PSO)算法在寻优时容易陷入局部最优的不足,提出一种基于子区域的PSO算法。将搜索空间划分成若干个子区域,在各个子区域中均使用标准PSO算法进行寻优,通过比较各个子区域的全局最优解,从而得出整个搜索空间的全局最优。与标准PSO算法及自适应变异PSO算法的比较结果表明,该算法能降低在寻优过程中陷入局部最优的概率,具有较强的寻优能力。
关键词:
粒子群优化算法,
局部最优,
全局最优,
子区域
CLC Number:
CENG Jia-Dun, LIU Zhi-Gang, HUANG Yuan-Liang, LIU Xin-Dong. Research of Particle Swarm Optimization Algorithm Based on Sub-region[J]. Computer Engineering, 2011, 37(14): 205-207.
曾嘉俊, 刘志刚, 黄元亮, 刘新东. 基于子区域的粒子群优化算法研究[J]. 计算机工程, 2011, 37(14): 205-207.