进化算法鲁棒最优解研究综述  

Survey of Robust Optimal Solutions in Evolutionary Algorithms

作　　者：郑金华[1] 罗彪[1] 李晶[1] 文诗华[1] 李望移[1] 

机构地区：[1]湘潭大学信息工程学院,湘潭411105

出　　处：《计算机科学》2009年第2期30-34,58,共6页Computer Science

基　　金：国家自然科学基金(60773047);湖南省教育厅重点科研项目(06A074)资助

摘　　要：在实际应用中,环境往往是不稳定的且易受到噪声的影响。因此,对于许多现实优化问题,一个鲁棒性好的解具有重要的意义。然而,以往关于进化算法(EAs)的研究主要集中在寻找全局最优解,解的鲁棒性却没有得到重视。从单目标鲁棒最优解、多目标鲁棒最优解及效率等方面较全面地分析了目前EAs搜索鲁棒最优解的研究现状。最后对相关研究工作做了展望。The environments are often unstable and susceptible to noises in real application. Therefore, it is significant for a solution with good robustness. However,most studies about evolutionary algorithms were focused on finding global optimal solutions till now, little attention concerned on robustness of solutions. We analyzed the research situation about searching robust optimal solutions with evolutionary algorithm from a series of aspects （single objective robust optimal solutions,mutli-objeetive robust optimal solutions and efficiency, et al）. Some future works were given at the last of the paper.

关 键 词：进化算法 鲁棒性 鲁棒最优解 单目标 多目标 效率 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] O242.23[自动化与计算机技术—控制科学与工程]