长沙理工大学学报（自然科学版）2026，Vol.23Issue(1)：162-173,12.DOI:10.19951/j.cnki.1672-9331.20250314001
新型元启发式算法:碰撞算法及其改进算法
A novel metaheuristic algorithm:collision algorithm and its improved algorithm
摘要
Abstract
[Purposes]To enhance the performance of metaheuristic algorithms in solving complex engineering optimization problems,this study,based on the law of conservation of momentum,proposed a novel metaheuristic optimization algorithm:the collision algorithm(CA).To address its shortcomings in local search capabilities and population diversity degradation,the algorithm was improved with multiple strategies,and then a multi-strategy improved collision algorithm(AMICA)was developed.[Methods]First,the CA framework was established,and its performance was validated.Next,CA was refined through strategies such as an elite pool strategy,Circle chaotic mapping initialization,t-distribution mutation strategy,adaptive spiral search strategy,and a nonlinear convergence factor based on the cosine function,leading to the development of AMICA.Finally,the performance of AMICA was tested using the IEEE CEC2017 benchmark functions.[Results]A comprehensive evaluation based on the IEEE CEC2017 test function demonstrates that AMICA exhibits significant advantages in unimodal,multimodal,hybrid,and composite functions.Its convergence speed,accuracy,and robustness are notably superior to CA and other comparative algorithms.The Wilcoxon rank sum test and Friedman test further validate AMICAs global search capabilities and stability,showing significant superiority across all test functions and ranking it as the top performer in overall performance.[Conclusion]The unique exploratory approach of the novel metaheuristic CA provides new insights for the development of metaheuristic algorithms.Through multi-strategy collaborative improvements,the algorithms performance is significantly improved.Compared to existing advanced algorithms,AMICA demonstrates superior overall performance in solving complex optimization problems.关键词
元启发式算法/碰撞算法/精英池策略/混沌映射/自适应螺旋搜索Key words
metaheuristic algorithm/collision algorithm/elite pool strategy/chaotic mapping/adaptive spiral search分类
信息技术与安全科学引用本文复制引用
李奇奇,牛林风,胡林..新型元启发式算法:碰撞算法及其改进算法[J].长沙理工大学学报（自然科学版）,2026,23(1):162-173,12.基金项目
国家杰出青年科学基金项目(52325211) （52325211）
湖南省优秀青年基金项目(2023JJ20040) （2023JJ20040）
湖南省科技计划项目(2024RC3169) （2024RC3169）
