摘要
为了提高粒子群搜索范围,防止陷入局部最优,在当今社会中团体分工合作的启发下,提出了一种新的基于分工合作的粒子群优化算法。在该算法中,将一个大的粒子群分成几个子群,按不同参数进化,在迭代过程中不断计算各个子群的平均适应度,设定一个阈值X,当任意两个子群的平均适应度之差大于该阈值时,则根据先进带动落后的合作思想,对平均适应度差的粒子群进行参数优化,实验结果表明,该算法在设定合适阈值时,扩大了搜索范围,从而提高了寻优精度。
A new particle swarm optimization algorithm is presented based on the theory of grouping and cooperating to improve the searching swarm, it is divided into several area and prevent getting optimized partially. In the algorithm , as a big subgroups, which have different evolutions. X as a threshold and calculate particle the average fitness of each subgroup in the iterative process continuously is set, when the margin of every two subgroups＇ fitnesses is greater than the threshold X, optimize the parameter of the subgroup whose average fitness is less according to the principle that the advanced bring the behindhand along. The experimental results demonstrate that when an appropriate threshold is set , the algorithm can expand the searching area to improve the precision of optimization.
出处
《科学技术与工程》
2009年第16期4806-4808,共3页
Science Technology and Engineering
关键词
分工合作
粒子群
子群
阈值
grouping and cooperating particle swarm subgroup threshold
