SGA(Simplex-Genetic Algorithm):一类求解Minimax问题的通用算法  

SGA(Simplex-Genetic Algorithm): a Universal Algorithm for Solving Minimax Problem

作　　者：郑泳凌[1] 马龙华[1] 钱积新[1] 

机构地区：[1]浙江大学系统工程研究所,浙江杭州310027

出　　处：《系统工程理论与实践》2002年第12期33-38,87,共7页Systems Engineering-Theory & Practice

摘　　要：　在指出一般的迭代法不能保证收敛性之后,将注意力投向基于Stackelberg-NashEquilibrium的遗传算法(GA)的解决方法,并根据Minimax问题的特点指出该方法的不足之处.在此基础上,提出了SGA(Simplex-GeneticAlgorithm).仿真实例表明,这种方法速度和精度较之GA都有了很大提高,是求解最小最大问题的有效通用方法.Minimax problem is one of the branches of multilevel programming, but unfortunately it lacks efficient algorithms. This paper discusses the convergence of implementing the Alternative Method at the beginning, then offers SGA(simplex\|genetic algorithm) ,which is a improving algorithm of GA for solving Stackelberg\|Nash Equilibrium. Examples are provided to illustrate that SGA is an efficient and universal approach for solving minimax problem.

关 键 词：SGA MINIMAX问题 通用算法 遗传算法 对策理论 

分 类 号：O225[理学—运筹学与控制论]