基于Pareto的快速多目标克隆选择算法  

Fast Pareto-based multi-objective clonal selection algorithm

作　　者：李恒杰[1] 郝晓弘[1] 张磊[1] 

机构地区：[1]兰州理工大学电气工程与信息工程学院,兰州730050

出　　处：《计算机应用研究》2008年第5期1368-1371,共4页Application Research of Computers

基　　金：甘肃省科技攻关项目(GS044-A52-001-24);甘肃省自然科学基金资助项目(3ZS042-B25-012)

摘　　要：基于免疫系统中克隆选择原理,提出了一种多目标克隆选择算法MCSA。该方法只对部分当前所得到的Pareto最优解进行进化操作,所求得的Pareto最优解保留在一个不断更新的外部记忆库中,并选用一种简单的多样性保存机制来保证其具有良好的分布特征。实验结果表明,该方法能够很快地收敛到Pareto最优前沿面,同时较好地保持解的多样性和分布的均匀性。对于公认的多目标benchmark问题,MCSA在解集分布的均匀性、多样性与解的精确性及算法收敛速度等方面均优于SPEA、NSGA-II等算法。A multi-objective clonal selection algorithm （MCSA） was proposed based on the clonal selection principle in the immune system. Only some Pareto optimal solutions were selected for further evolutionary operation in the algorithm, The Pareto optimal solutions were reserved in an external memory set which was renewed in each generation, and a simple mechanism was used to maintain good spread of solutions. It is shown by experimental results that the method can reach the Pareto optimal front very quickly and retain the better diversity of the solutions, The proposed MCSA is superior to other algorithms such as SPEA, NSGA-II etc. in terms of the precision, the quantity, the distribution uniformity, the diversity of solutions and the convergence rate of algorithm in solving one kind of typical benchmark problems.

关 键 词：克隆选择原理 PARETO最优解 多目标优化 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]