改进的克隆选择算法与SPEA相结合的进化算法  

Evolutionary Algorithm Based on Improved Clonal Selection Algorithm and SPEA

作　　者：杨观赐[1] 马鑫[2] 李少波[1,2] 钟勇[1] 于丽娅[2] 

机构地区：[1]中国科学院成都计算机应用研究所,四川成都610041 [2]贵州大学教育部现代制造技术重点实验室,贵州贵阳550003

出　　处：《四川大学学报（工程科学版）》2011年第5期109-113,共5页Journal of Sichuan University (Engineering Science Edition)

基　　金：教育部新世纪优秀人才支持计划资助项目(NCET09-0094);贵州省科学技术基金资助项目(黔科合J字[2010]2095号)

摘　　要：为了使进化过程中子代的繁殖能够像生物繁殖那样继承进化信息,通过挖掘抗体中优秀决定基并生成记忆集、增加高斯变异、用变异抗体群中亲和度高的抗体按概率替换记忆抗体群中低亲和度抗体等策略,提出了一种改进的克隆选择算法(ICSA)。将ICSA与SPEA相结合,形成了一种改进的克隆选择算法与强度Pareto进化算法相结合的新型的进化算法(ICSA-SPEA)。ICSA-SPEA通过克隆选择替代选择、交叉、重组等遗传操作。用一组多目标0/1背包问题测试算法性能的统计结果表明,改进的算法可以有效保持种群多样性,具有良好的收敛精度与准确度。In order to inherit evolutionary information as living beings during offspring generation,a kind of improved clonal selection algorithm（ICSA） was put forward based on mining excellent gene schema to fill a memory pool from antibody set,applying Gaussian mutation operator,and replacing low affinity antibody with high affinity antibody with probability from mutation antibody population during updating memory antibody population.By combining ICSA with strength Pareto evolutionary algorithm（SPEA）,a new kind of evolutionary algorithm（ICSA-SPEA） was proposed,which replaces the genetic operation such as selection,crossover and recombinant with clonal selection.The testing of a multi-objective 0/1 knapsack problem showed that ICSA-SPEA has ability to maintain the diversity of population,and is capable of finding out the well distributed non-dominated solutions approximating to Pareto front.

关 键 词：多目标优化 进化算法 克隆选择 基因挖掘 遗传信息 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP273[自动化与计算机技术—控制科学与工程]