基于随机扰动的多目标进化算法  

An Evolutionary Algorithm for Multi-Objective Optimization Problem Based on Random Distuibance

作　　者：郭修豪 陈勇[1] 

机构地区：[1]重庆师范大学,重庆401331

出　　处：《现代计算机（中旬刊）》2015年第12期3-7,42,共6页Modern Computer

基　　金：国家自然科学基金资助项目(60703035);重庆市教委基金资助项目(No.KJ070801);重庆市教委科技项目(No.KJ120622)

摘　　要：运用遗传算法解多目标问题,结果往往会陷入局部最优。引入传统算法求得的外部种群,提出基于随机扰动的RDMOGA遗传算法。将新算法用标准多目标测试函数进行测验,并与韩丽霞提出的NMOGA算法进行对比,实验结果表明,新算法表现出良好的搜索性能。Uses genetic algorithm to solve multi-objective problem, the result is often trapped in local optimum. Introduces the external population of the traditional algorithm, and proposes a genetic algorithm based on random perturbation of the RDMOGA. The new algorithm is tested by using the standard multi objective test functions, and compared with the NMOGA algorithm proposed by Han Lixia. The test results show that the new algorithm shows good performance.

关 键 词：多目标优化 随机扰动 进化算法 拥挤距离排序 C-measure U-measure 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]