采用多模式飞行的乌鸦搜索算法  

Crow search algorithm using multi-mode flight

作　　者：冯爱武 王勇[1,2] 付小朋 Feng Aiwu;Wang Yong;Fu Xiaopeng(College of Artificial Intelligence,Guangxi Minzu University,Nanning 530006,China;Guangxi Key Laboratory of Hybrid Computation&IC Design Analysis,Nanning 530006,China)

机构地区：[1]广西民族大学人工智能学院,南宁530006 [2]广西混杂计算与集成电路设计分析重点实验室,南宁530006

出　　处：《计算机应用研究》2022年第6期1710-1717,共8页Application Research of Computers

基　　金：国家自然科学基金资助项目(61662005);广西自然科学基金资助项目(2021JJA170094)。

摘　　要：针对乌鸦搜索算法(CSA)的不足,提出采用多模式飞行的乌鸦搜索算法(MFCSA)。算法基于觅食能力的强弱,将群体分成觅食能力较强和较弱两个组,觅食能力较强者采用尾随跟踪当前群体最优目标策略,在群体信息指引下飞到群体当前最优位置附近开展搜索活动,增强了算法的局部开发能力;觅食能力较弱者采用观察和学习强者的觅食方法、遇到危险迅速飞离两种策略,前者可提升算法的全局探索能力,后者可保持种群的多样性。通过15个基准测试函数和两个工程应用问题的数值实验仿真结果表明,MFCSA在优化精度、收敛速度等方面有更好的表现,增强了规避陷入局部最优的能力,稳定性更好。Aiming at the shortcomings of crow search algorithm(CSA),this paper proposed a crow search algorithm using multi-mode flight(MFCSA).Based on the strength of foraging ability,the algorithm divided the population into two groups:strong and weak foraging ability.Those with strong foraging ability adopted the strategy of trailing and tracking the optimal target of the current group,and flied to the vicinity of the current optimal position of the group under the guidance of the group information to carry out search activities,which enhanced the local exploitation ability of the algorithm.Those with weaker foraging ability adopted the two strategies of observing and learning the foraging methods of the strong,and flying away quickly when encountering danger,the former could improve the global exploration ability of the algorithm,and the latter could maintain the diversity of the population.Through the numerical experiment simulation of 15 benchmark test functions and 2 engineering application problems,the results show that the MFCSA has better performance in optimization accuracy,convergence speed,etc.,enhances the ability to avoid falling into the local optimum,and has better stability.

关 键 词：乌鸦搜索算法 多模式飞行 工程约束问题 智能计算 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]