基于搜索空间可调的自适应粒子群优化算法与仿真  

Adaptive particle swarm optimization based on search space adjustable and simulation

作　　者：段其昌[1] 张红雷[1] 

机构地区：[1]重庆大学自动化学院,重庆400044

出　　处：《控制与决策》2008年第10期1192-1195,共4页Control and Decision

摘　　要：针对收缩因子粒子群优化(CPSO)算法易陷入局部最优和发生过早收敛的问题,提出了基于搜索空间可调的自适应粒子群优化(APSO)算法.该算法根据种群早熟收敛程度和个体适应值,在CPSO算法停滞时,将全部粒子有效地划分在3类不同的搜索空间,使种群始终保持搜索空间的多样性,易于跳出局部最优,从而有效地改善了CPSO算法后期的寻优能力.An adaptive particle swarm optimization （APSO） algorithm based on the search space adjustable is proposed and applied to solve the problem that constriction-factor PSO （CPSO） algorithm easily fall into local optimal and occur premature convergence. When the CPSO algorithm stagnates, according to the extent premature convergence groups and individual fitness, the algorithm will divide particles to three different searching spaces, by which the swarm is kept to maintain the diversity of the searching space and easy to jump out of local optima. The late CPSO algorithm optimization capabilities availability are improved.

关 键 词：收缩因子粒子群优化 早熟收敛 搜索空间 自适应参数调整 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]