自适应调整子种群个体数目的遗传算法及其应用

摘要

In order to coordinate the exploitation ability and development ability of the algorithm,a genetic atgorithm of individuals number of subpopulation with adaptively adjustment was proposed.A fine-point set method was employed to conduct population initialization in order to guarantee that the individuals were scattered uniformly over entire search space.The population was divided into three subpopulations according to individuals adaptability and different crossover operators and mutation operators were selected for them respectively.In the process of evolution,the individuals number of the subpopulation was dynamically and adaptively adjusted at different search phases.The experimental result of several benchmark functions showed that this algorithm exhibited a better ability of optimum searching.By using the new algorithm for parameter estimation of heavy oil thermal cracking model,a satisfactory result could be obtained.%为了协调算法的勘探和开采能力,提出一种自适应调整子种群个体数目的遗传算法.该算法首先采用佳点集方法初始化种群以保证个体均匀分布在搜索空间中.基于个体的适应度将种群分为3个子种群,并分别采用不同的交叉和变异算子.在进化过程中,根据不同的搜索阶段自适应动态调整各子种群个体的数目.几个标准测试函数的实验结果表明该算法具有较好的寻优性能.将新算法应用到重油热解模型参数估计中,可以获得满意的结果.