摘要
针对标准二进制编码遗传算法的缺陷,提出一种基于实数编码技术的新型自适应混沌遗传算法用于求解优化问题.该算法利用信息熵理论产生较好的初始群体分布,并依据概率分布函数构造杂交算子,同时结合混沌动力学特性和人工神经网络理论,设计了一种自适应混沌变异算子,使算法能有效维持群体多样性,防止和克服进化过程中的“早熟”现象,算法操作简单、易于实现.最后通过对几个经典测试函数的数值实验,验证了该算法在提高解的精度和加快收敛速度方面都有显著改善,从而为解决函数优化问题提供了一种行之有效的新方法.
This paper presents a new real-value encoding self-adaptive chaotic genetic algorithm to solve optimization problem based on the analysis for shortcoming of standard binary-encoding genetic algorithm. It is used the entropy based on information theory to initialize population with better distribution and designed a crossover operator in light of probability distribution function and a self-adaptive chaotic mutation operator combined chaotic dynamic character with artificial neural network theory, which maintains population diversity to prevent and overcome premature phenomena in the evolutionary process. This algorithm is easy to implement with the simple operation. Several typical benchmark function numerical experiments demonstrate that it is improved on the solution precision and increased convergence speed. The proposed method provides an effective new method to solve the function optimization problems.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2006年第4期708-712,共5页
Acta Electronica Sinica
基金
国家自然科学基金(No.50409010
No.50309013
No.40572166
No.50539140)
湖北省自然科学基金(No.2005ABA228)
关键词
遗传算法
混沌
早熟收敛
熵
genetic algorithm
chaos
premature convergence
entropy
