摘要
文中在连续空间统一的随机过程框架下 ,分析了遗传算法群体的概率密度序列的演化过程 ,给出并证明了群体概率密度的递归公式 .分析了标准遗传算法中选择算子和变异算子的基本性质 .导出了选择和变异条件下平均适应度单调递增并收敛到全局最优解的条件 .这些结论在一定程度上为实现自适应调节变异算子的概率 。
This paper is to study genetic algorithm in the unified framework of stochastic processes in continuous space.The evolutionary processes of the probability density of the population in continuous space is analyzed and the recursive formulas of the probability density of the population are derived.The basic properties associated with selection and mutation are analyzed.Sufficient condition for monotonous increase of the average fitness converging to the global maximum value is derived.These conclusions provide a basis in theory for the adaptive mutation operator and the convergence of genetic algorithm to the global optimal solution.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2000年第2期31-34,48,共5页
Acta Electronica Sinica
基金
国家自然科学基金!(No .6960 1 0 0 3)
关键词
遗传算法
选择算子
变异算子
genetic algorithm
continuous space
selection operator
mutation operator
