采用混沌变异的进化算法  

Evolutionary Algorithms with Chaotic Mutations

作　　者：骆晨钟[1] 邵惠鹤[1] 

机构地区：[1]上海交通大学自动化系,200030

出　　处：《控制与决策》2000年第5期557-560,共4页Control and Decision

摘　　要：根据混沌理论关于进化与混沌的关系 ,设计一种采用混沌变异算子的进化算法 ,并提出“尺度收缩”的变异策略。对极小值函数优化问题的仿真实例表明 ,混沌变异是实数编码进化算法变异算子的有效实现 ;而采用“尺度收缩”策略的混沌变异算子明显改善了群体平均适应值 ,提高了算法性能 。Based on chaotic theory′s understanding on relationship between chaos and evolution, chaotic mutation operator was proposed for implementing real coded evolutionary algorithms. A new strategy called “shrinking” mutation strategy was also designed. Simulations with minimum function optimization problems show that, chaotic mutation is an effective implementation of the mutation operator. And the “shrinking” chaotic mutation can dramatically improve the average individual fitness and the performance of the evolutionary algorithm.

关 键 词：函数优化 进化算法 混沌变异 实数编码 

分 类 号：O242.23[理学—计算数学] TP301.6[理学—数学]