一种混沌映射动态惯性权重的樽海鞘群算法  

Salp Swarm Algorithm Based on Chaotic Map and Dynamic Inertia Weight

作　　者：周密 王潇棠 闫河 谢敏 ZHOU Mi;WANG Xiao-tang;YAN He;XIE Min(College of Science,Chongqing University of Technology,Chongqing 400054,China;Liangjiang Artificial Intelligence Academy,Chongqing University of Technology,Chongqing 401135,China)

机构地区：[1]重庆理工大学理学院,重庆400054 [2]重庆理工大学两江人工智能学院,重庆401135

出　　处：《小型微型计算机系统》2023年第2期313-318,共6页Journal of Chinese Computer Systems

基　　金：国家重点研发计划"智能机器人"重点专项项目(2018YFB1308602)资助;国家自然科学基金面上项目(61173184)资助;重庆市自然科学基金项目(cstc2018jcyjAX0694,cstc2020jcyj-msxmX0879)资助.

摘　　要：为提升群海鞘群算法求解精度和收敛速度,提出了基于混沌映射动态惯性权重的群海鞘群算法.首先利用Tent混沌映射生成初始种群,计算种群适应度并保留最优个体作为初始食物源位置;将种群大小均分为领导者和追随者两部分以提高算法全局搜索能力,通过引入疯狂算子完成对领导者的位置更新;在追随者位置更新公式中,提出了基于精英保留及动态惯性权重的追随者位置更新策略,通过计算个体适应度值完成食物源的位置更新.实验结果表明,改进算法拥有更好的性能.In order to improve the precision and convergence velocity of the Salp Swarm algorithm,Salp Swarm Algorithm based on chaotic map and dynamic inertia weight is proposed.In this paper,the initial population′s position is generated by Tent chaotic map,the population fitness is calculated by objective function and the optimal individual is retained as the initial food source position.The population size is divided into leaders and followers evenly to improve the global search capability of the algorithm.The crazy operator is introduced to complete the position update of the leader.In the formula of the follower position updating,a follower position update strategy based on elite preservation and dynamic inertia weight is proposed.The position updating of the food source is completed by calculating the individual fitness.The experimental results show that the improved algorithm has better performance.

关 键 词：樽海鞘群算法 混沌映射 疯狂算子 精英保留策略 动态惯性权重 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]