求解动态优化问题的改进原对偶遗传算法  

Dynamic Problem Optimization Using the Improved Primal-Dual Genetic Algorithm

作　　者：王洪峰[1] 汪定伟[1] 刘黎黎[1] 

机构地区：[1]东北大学信息科学与工程学院,辽宁沈阳110004

出　　处：《东北大学学报（自然科学版）》2007年第5期639-642,共4页Journal of Northeastern University(Natural Science)

基　　金：国家自然科学基金资助项目(704310003);国家创新研究群体科学基金资助项目(60521003);国家支撑计划项目(2006BAH02A09)

摘　　要：针对求解动态0-1优化问题的原对偶遗传算法(pri mal-dual genetic algorithm,PDGA)中一个关键的运算──原对偶映射(pri mal-dual mapping,PDM)进行改进,提出了一种新的适应性的PDM方法.在新的映射方法中,利用种群中染色体各个基因位点上取值的统计信息来计算该基因位点进行PDM运算的概率.在一组动态优化函数的仿真实验中,改进的PDGA算法表现出比原始算法更好的性能.The PDM （primal-dual mapping） as a key operation in PDGA （primal-dual genetic algorithm） that has successfully been applied to the dynamic 0-1 optimization problems is improved, and a new adaptive PDM scheme is proposed. Then, the statistical information on the allele distribution in each locus over the population is used to calculate the probability of PDM in the corresponding locus. Simulation results from a set of dynamic benchmark problems showed that the improved PDGA outperforms the original algorithm in dynamic environment.

关 键 词：动态 优化 原对偶映射 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]