混合三群粒子群优化算法求解min-max-min问题  

Hybrid three sub-swarms particle swarm optimizer for min-max-min problem

作　　者：韦鹏[1] 曹德欣[1] 

机构地区：[1]中国矿业大学理学院,江苏徐州221008

出　　处：《计算机工程与应用》2010年第35期219-221,230,共4页Computer Engineering and Applications

基　　金：国家自然科学基金(No.60775044)~~

摘　　要：针对标准粒子群算法求解复杂优化问题时容易出现过早收敛的问题,提出了混合三群协同粒子群算法(HTSPSO),将粒子群分为3个协同优化的子群,保持迭代后期粒子群的多样性。在4个经典测试函数上的仿真实验表明,新算法较传统PSO算法收敛更快,精度更高。将粒子群算法应用于求解一类min-max-min问题,并给出了数值算例。The Standard Particle Swarm Optimizer（SPSO）may lead to premature convergence when optimizing complex opti-mization problems.A Hybrid Three Sub-Swarm Particle Swarm Optimizer（HTSPSO） is presented to improve the performance of PSO.The swarm is divided into three sub-swarms and the sub-swarms work cooperatively to preserve the diversity of the swarm in the late stage of iterations.Experiments are conducted on four benchmark problems.The results demonstrate signifi-cant improvement in performance over the traditional PSOs.Furthermore,the PSO method is applied to solve a class of min-max-min problems and the numerical examples are proposed.

关 键 词：粒子群优化算法 三子群协同 min-max-min问题 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]