积分——水平集求全局最优的遗传算法实现  

Genetic Algorithm Implementation of Integral-level Set Global Optimization

作　　者：王延源[1] 

机构地区：[1]山东临沂师范学院数学系,山东临沂276005

出　　处：《数学的实践与认识》2006年第3期212-216,共5页Mathematics in Practice and Theory

摘　　要：Chew Soo Hong,Zheng Q uan提出了一个积分——水平集求全局最优的概念性算法及M on te-C ar-lo随机投点的实现途径,并在很多实际问题中得到了很好的应用,但这一实现算法的收敛性是个未解决的问题.利用近年来广泛应用的遗传算法,给出了这一算法的另一种实现途径,并从理论和数值两个方面验证了算法的可行性.A theoretical algorithm for the integral-level set global optimization is proposed and the Monte-Carlo implementation of the algorithm is discussed in [1]. However, the global convergence of the implementation approach has not been proved. In this paper, the author propose an another implementation approach——the genetic algorithm implementation of the algorithm, and have proved the global convergence of this approach. The numerical results also show that the approach is efficient.

关 键 词：全局收敛性 最优 水平集 遗传算法 

分 类 号：O224[理学—运筹学与控制论] TP302[理学—数学]