摘要
文章以多头拱架式贴片机为研究对象,建立贴装优化数学模型,给出保证贴装序列无循环的充要条件并证明;提出适合贴片机分组贴装的离散狼群算法,在喂料槽分配已知的情况下,优化贴装路径;为加快收敛速度、提高求解精度,引入面向近邻概率生成初始解的方法;设计狼群分组行为实现并行搜索,避免陷入局部最优;提出头狼调整策略进一步加快收敛速度;采用Taguchi正交实验优化算法关键参数,使参数选取有理论支撑。对不同规模的印刷电路板(printed circuit board,PCB)进行测试,实验结果表明:改进离散狼群算法是有效的,其求解精度更高,求解速度更快;改进离散狼群算法效率比未改进的平均提高3.18%,比蜂群算法平均提高16.67%。
Taking the multi-head arch placement machine as the research object,a mathematical model of placement optimization is established,and the necessary and sufficient conditions to ensure that the placement sequence has no cycle are given and proved.A discrete wolf pack algorithm is proposed to optimize the placement path of the placement machine under the known distribution of feeder slots.In order to accelerate the rate of convergence and improve the precision of solution,the method of generating initial solution based on nearest-neighbor probability is introduced.The group behavior of wolves is designed to realize parallel search and avoid falling into local optimum.The head-wolf adjustment strategy is proposed to further accelerate the rate of convergence.Taguchi orthogonal experiment is used to optimize the key parameters of the algorithm,thus making the selection of parameter has theoretical support.The test results of printed circuit boards(PCB)in different scales show that the improved algorithm is effective,with higher precision and faster speed.The efficiency of the improved wolf pack algorithm is 3.18%higher than that of the unimproved one,and 16.67%higher than that of the bee colony algorithm.
作者
倪郁东
曾志超
杨晨罡
杨航
高丽杰
夏仕伟
NI Yudong;ZENG Zhichao;YANG Chengang;YANG Hang;GAO Lijie;XIA Shiwei(School of Mathematics, Hefei University of Technology, Hefei 230601, China)
出处
《合肥工业大学学报（自然科学版）》
CAS
北大核心
2020年第8期1082-1089,共8页
Journal of Hefei University of Technology：Natural Science
基金
合肥工业大学大学生创新性实验计划基金资助项目(2018CXCYS205)。
关键词
多头拱架式贴片机
贴装优化
改进离散狼群算法
Taguchi正交实验设计
充要约束条件
multi-head arch placement machine
placement optimization
improved discrete wolf pack algorithm
Taguchi orthogonal experimental design
necessary and sufficient constraint conditions
