早熟收敛
水准点(测量)
差异进化
突变
利基
健身景观
数学优化
进化算法
生态位
趋同(经济学)
进化计算
计算机科学
适应度函数
人口
集合(抽象数据类型)
人工智能
生物
数学
粒子群优化
遗传算法
遗传学
生态学
经济
经济增长
人口学
社会学
栖息地
程序设计语言
地理
大地测量学
基因
标识
DOI:10.1109/ispds58840.2023.10235594
摘要
To handle the problem of premature convergence and premature loss of diversity in differential evolution, a niching method fitness-based and an adaptive mutation strategy differential evolution is proposed for global optimization. This algorithm designs a niche partitioning strategy that divides the population into niches with different numbers of individuals based on their fitness and evolutionary stages. Adopting adaptive mutation strategies for the obtained niches enhances the exploitation of superior niches and the exploration of inferior niches, thereby supporting balanced evolutionary search. On the CEC'2015 benchmark function test set, the suggested method's performance is assessed and contrasted with related methods. According to the outcomes, the suggested fitness-based niching and adaptive mutation schemes hold promise for enhancing DE performance.
科研通智能强力驱动
Strongly Powered by AbleSci AI