基于嵌入干扰机制的改进人工鱼群算法及其应用  

An Improved Artificial Fish Swarm Algorithm Based on Embedded Jamming Mechanism and its Application

作　　者：陆凤仪[1] 任重 徐格宁[1] LU Feng-yi;REN Zhong;XU Ge-ning(School of Machinery and Electronics Engineering,Taiyuan University of Science and Technology,Shanxi Taiyuan030024,China)

机构地区：[1]太原科技大学机械工程学院,山西太原030024

出　　处：《机械设计与制造》2020年第11期69-72,77,共5页Machinery Design & Manufacture

基　　金：“十三五”国家重点研发计划-港口等领域典型起重机械设计制造与服役过程风险防控关键技术研究(2017YFC0805703)。

摘　　要：针对人工鱼群算法(AFSA)中收敛精度不足、早熟现象明显、稳定性较差及结果精度受参数影响较大的问题,提出基于嵌入机制的改进人工鱼群算法。以参数统计分析为基础,设定步长与视野数值成反比的条件下,构建一种模仿生物视觉特性的步长、视野因子模型,并将差分进化算法中的变异行为作为人工鱼群完成觅食、聚群、追尾行为后的干扰行为。根据参数之间的关系、视野与步长的关系,采用五种类型测试函数分别对AFSA、DAFSA、DE进行对比验证,结果表明:DAFSA在稳定性、计算精度与早熟现象改进方面较AFSA有明显提升。同时为验证所提方法的实用性,将其应用在50t/22.5m桥式起重机主梁金属结构轻量化设计中。In the artificial fish Swarm algorithm(AFSA),the convergence accuracy is insufficient,the premature phenomenon is obvious,the stability is poor and the result precision is influenced by the parameter,an improved artificial fish swarm algorithm based on embedding mechanism is proposed.On the basis of the parameter statistical analysis,a step and visual factor model is constructed to imitate the biological visual characteristics,and the mutation behavior of the differential evolutionary algorithm is used as the interfering behavior of the artificial fish in the foraging,clustering and rear-ends behavior.According to the relationship between the parameters,the field of vision and the step size,five types of test functions are used to compare the AFSA,DAFSA and DE,and the results show that the DAFSA has a significant improvement in stability,computational accuracy and early-maturing phenomenon improvement.In order to verify the practicability of the proposed method,it is applied to the lightweight design of the metal structure of the main girder forbridge crane in 50 t/22.5 m.

关 键 词：人工鱼群算法 生物视觉特性 差分进化算法 干扰行为 

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