Lévy飞行的正余弦乌燕鸥混合算法及应用  

A Hybrid Sine and Cosine Algorithm and Lévy Flight Adaptive Sooty Tern Algorithm and its Application

作　　者：孙珂琪[1] 陈永峰[1] SUN Ke-qi;CHEN Yong-feng(School of Railway Equipment Manufacturing,Shaanxi Railway Institute,Shaanxi Weinan 714000,China)

机构地区：[1]陕西铁路工程职业技术学院铁道装备制造学院,陕西渭南714000

出　　处：《机械设计与制造》2023年第1期212-217,共6页Machinery Design & Manufacture

基　　金：省教育厅科研计划专项(2017JCYJ-3-2)。

摘　　要：为解决标准乌燕鸥算法(STOA)易陷入局部最优和收敛速度慢等缺点,提出一种混合正余弦算法(SCA)和Lévy飞行的自适应乌燕鸥算法(SLSTOA)。采用正余弦算法的搜索方式,同时采用非线性递减自适应正弦因子,改进乌燕鸥算法的攻击搜索方式,来增强STOA算法的全局与局部探索能力。乌燕鸥个体和最优个体通过Lévy飞行策略进行变异,来增加种群多样性和扩大搜索空间,以达到提高跳出局部最优和全局探索能力。与四种先进的元启发式算法比较,SLSTOA算法性能通过6个基准测试函数进行评价,结果表明,相比其他四种元启发式算法,SLSTOA算法精度高、稳定性好和鲁棒性强。同时为验证SLSTOA算法的科学性与实用性,将其应用于解决32t/22.5m桥式起重机主梁结构优化设计中。In order to solve the shortcomings of sooty tern algorithm(STOA),such as easy to fall into local optimum and slow convergence speed,a hybrid sine and cosine algorithm(SCA)and Lévy flight(LF)adaptive black tern algorithm(SLSTOA)was proposed. In order to enhance the global and local exploration ability of STOA algorithm,the combination operator of SCA algorithm and nonlinear decreasing adaptive sine factor was introduced into STOA algorithm in the attacking behavior. Sooty tern individuals and optimal individuals were mutated by the LF strategy to increase the population diversity and search space and improve the ability to jump out of the local optimal and global exploration. Compared with four advanced metaheuristic algorithms,the performance of SLSTOA algorithm was evaluated by six benchmark functions. The results show that SSTOA algorithm has high accuracy,good stability and strong robustness,compared with the other four metaheuristic algorithms. At the same time,the practicability of SLSTOA algorithm was evaluated by the optimization design problem of 32t/22.5m bridge crane girder structure.

关 键 词：乌燕鸥算法 正余弦算法 自适应正弦因子 Lévy飞行 桥式起重机主梁 

分 类 号：TH16[机械工程—机械制造及自动化] TB112[理学—数学]