一种新的进化策略及其全局收敛性  

A new kind of evolutionary strategy and its global convergence properties

作　　者：周东生[1] 李斌[2] 唐焕文[1] 

机构地区：[1]大连理工大学管理科学与工程研究所 [2]大连理工大学应用数学系,辽宁大连116024

出　　处：《大连理工大学学报》2007年第1期146-151,共6页Journal of Dalian University of Technology

基　　金：国家自然科学基金资助项目(10571018)

摘　　要：在求解单峰最优化问题算法的基础上,给出了一种新的进化策略.针对连续函数优化问题,利用中心极限定理,在较弱的条件下,首先证明了基于均匀分布的(μ+λ)-ES算法依概率收敛,然后给出了采用一般连续性随机变量作为变异算子的(μ+λ)-ES算法依概率收敛的证明.数值结果表明:采用基于均匀分布的进化策略求解维数较高的连续函数优化问题能够快速有效地收敛到全局最优解.On the basis of the algorithm solving the single peak optimization problem, a new evolutionary strategy is put forward. In view of continuous function problems, it is first proved that the （μ＋λ）-ES algorithm based on uniform distribution converges in probability in a weaker condition by means of the central limit theorem. Then it is proved that the （μ＋λ）-ES algorithm which adopts the general continuous random variables as its variable operators converges in probability. Numerical results indicate that adopting the evolutionary strategy based on uniform distribution to solve the optimization problems of the continuous function with comparatively higher dimensions can converge to its global optimum solution quickly and effectively.

关 键 词：进化算法 进化策略 全局收敛性 中心极限定理 

分 类 号：O22[理学—运筹学与控制论]