精英策略的改进非支配遗传算法  

Improved non-dominated sorting genetic algorithm Ⅱ based on the elitist strategy

作　　者：张晓菲[1] 张火明[1] 

机构地区：[1]中国计量学院计量测试工程学院,浙江杭州310018

出　　处：《中国计量学院学报》2010年第1期52-58,共7页Journal of China Jiliang University

基　　金：国家自然科学青年基金资助项目(No.10602055)

摘　　要：在研究NSGA-Ⅱ算法的基础上,提出了一种新的精英策略.它既可以使种群中优良个体保持到下一代,又可以维护种群的多样性,避免算法陷入局部最优解.此外,对NSGA-Ⅱ算法的交叉算子和变异算子进行了重新选择,交叉算子采用模拟二进制交叉算子,变异算子采用非均匀变异算子.将它们引入INSGA-Ⅱ算法中,在优化前期可以进行大范围搜索,而后期可以集中搜索某个重点区域,从而可以提高解的精确性.通过经典函数测试,并将INSGA-Ⅱ算法与NSGA-Ⅱ算法的结果进行比较.结果表明,基于精英策略的改进算法INSGA-Ⅱ不仅取得了较好的结果,而且分布性也得到了有效的提高.A new elitist strategy based on NSGA-Ⅱ was proposed,which could not only make excellent individuals in population maintained until the next generation,but also maintain the population diversity and avoid the algorithm into a local optimal solution.In addition,the crossover operator and mutation operator on NSGA-Ⅱ algorithm were re-selected.SBX and the non-uniform mutation operator were introduced in order to realize INSGA-Ⅱ search in a wide range in the early stage of optimization,but could realize a focus ...

关 键 词：多目标优化 精英策略 非支配排序遗传算法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]