《计算机应用研究》|Application Research of Computers

改进自适应多种群NSGA-Ⅲ算法的研究

Research on improved adaptive multi-population NSGA-Ⅲ

免费全文下载 (已被下载 次)  
获取PDF全文
作者 刘彬,王卫涛,武尤,杨有恒
机构 燕山大学 a.电气工程学院;b.信息科学与工程学院,河北 秦皇岛 066004
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2021)01-009-0048-05
DOI 10.19734/j.issn.1001-3695.2019.10.0595
摘要 针对第三代非支配排序遗传算法(non-dominated sorting genetic algorithm-Ⅲ,NSGA-Ⅲ)在处理高维多目标函数时存在收敛精度低和搜索性能差等问题,提出一种自适应多种群NSGA-Ⅲ算法。首先将传统算法的单一种群划分成四个亚种群,并为每个亚种群分配不同的交叉算子;其次提出外部最优解集(external optimal solution set,EXS)的概念,通过计算个体更新最优解集的参与量来自适应调节每个亚种群的大小;最后利用局部搜索策略提高EXS的局部搜索性能。采用四个不同的测试函数,与七种对比算法进行仿真验证,结果表明在处理高维多目标优化问题时,提出算法的性能指标整体优于其他对比算法,能够获得较好的算法收敛性和种群多样性。
关键词 高维多目标; 非支配排序遗传算法; 自适应; 多种群
基金项目 河北省自然科学基金资助项目(F2019203320,E2018203398)
河北省人才培养项目(A201903005)
本文URL http://www.arocmag.com/article/01-2021-01-009.html
英文标题 Research on improved adaptive multi-population NSGA-Ⅲ
作者英文名 Liu Bin, Wang Weitao, Wu You, Yang Youheng
机构英文名 a.School of Electrical Engineering,b.School of Information Science & Engineering,Yanshan University,Qinhuangdao Hebei 066004,China
英文摘要 Aiming at the problems of low-convergence accuracy and poor search performance of the third-generation non-dominated sorting genetic algorithm-Ⅲ(NSGA-Ⅲ) when dealing with high-dimensional multi-objective functions, this paper proposed an adaptive multi-population NSGA-Ⅲ algorithm. First it divided a single population of the NSGA-Ⅲ algorithm into 4 sub-populations, and assigned different crossover operators to each subpopulation. Secondly, it proposed concept of external optimal solution set(EXS), and adaptively adjusted the size of each subpopulation by calculating the amount of participation of the individual to update the optimal solution set. Finally, this paper proposed the local search strategy to improve the local search performance of the EXS. It used 4 different test functions for simulation verification with 7 comparison algorithms. The results show that when dealing with high-dimensional multi-objective optimization problems, the performance index of the proposed algorithm is better than other comparison algorithms, which can obtain better algorithm convergence and population diversity.
英文关键词 high-dimensional multi-objective; NSGA-Ⅲ; adaptive; multi-population
参考文献 查看稿件参考文献
 
收稿日期 2019/10/16
修回日期 2020/1/3
页码 48-52,87
中图分类号 TP301.6
文献标志码 A