摘要
针对基本粒子群算法容易陷入局部最优点,进化后期速度慢等缺点,设计了一种新的粒子群算法,将基本粒子群算法粒子行为基于个体极值点转化为个体自身极值与其他某一个个体极值的加权平均值,而全局极值点转化为群体中优秀个体极值的加权平均值。数值仿真实验表明,新算法比PSO具有更好的收敛性,能更快地找到问题的最优解。
A modified particle swarm optimization（MPSO） is proposed which aims at the shortcoming of the standard PSO algorithm, that is easily plunging into local optimum, converging slowly in last period of evolution. The new algorithm alters particle swarm optimization based on the individual extreme point into the weighted average of its own individual extreme and other individual extreme value, while the global extreme point into the weighted average of a group of outstanding individuals＇ extreme values. Experimental results indicate that the MPSO has better optimizer ability than PSO, and can more quickly find the optimal solution of problem.
出处
《计算机与数字工程》
2009年第8期33-35,共3页
Computer & Digital Engineering
关键词
粒子群算法
极值
加权平均值
particle swarm optimization, extreme value, weighted average
