摘要
通过把探索周期轨道的问题转化为一个非负函数的极小化问题,基于空间收缩的思想,反复重新初始化种群以提高群体的差异性,提高对可行域探测的有效性,提出一类精英子空间差异演化算法求解不同类型非线性映射的不稳定周期轨道,对典型离散动力系统的初步仿真结果表明该算法是一种计算不稳定周期轨道的稳健有效算法.
In this paper, through transferring computing periodic orbits into the problem of a non negative function's minimization, a new approach for it by using Differential Evolutionary Algorithms basing on Elite Subspace is proposed by regenerating the population to raise the difference of the population. Experimental results on well known and widely used various nonlinear mappings indicate that proposed algorithm is robust and efficient for unstable periodic orbit of discrete chaotic system.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
2005年第4期96-102,共7页
Systems Engineering-Theory & Practice
基金
武汉理工大学校基金(XJJ2004113)资助
关键词
非线性映射
周期轨道
精英子空间
差异演化
Nonlinear map
Period Orbit
Elite Subspace
Differential Evolution