基于循环配送的混流装配线物料配送优化  

Optimization of Material Delivery Problem for Mixed-model Assembly Line Based on Milk-run Delivery

作　　者：周晓晔[1] 马小云[1] 朱梅琳 崔瑶 ZHOU Xiao-ye;MA Xiao-yun;ZHU Mei-lin;CUI Yao(School of Management,Shenyang University of Technology,Shenyang 110870,China)

机构地区：[1]沈阳工业大学管理学院,辽宁沈阳110870

出　　处：《系统工程》2023年第1期93-101,共9页Systems Engineering

基　　金：国家自然科学基金青年规划基金资助项目(71801160);辽宁省社科规划基金重点资助项目(L18AGL005);沈阳工业大学服务地方经济社会发展决策专项课题(FWDFGD2021017)。

摘　　要：针对混流装配线物料循环配送次数多、线边库存积压的问题,考虑多种零部件循环配送并储存至同一工位的实际生产情况,权衡循环配送次数与线边库存量之间的背反关系,建立基于循环配送次数与线边库存量的优化模型。在此基础上,提出单亲遗传蚁群算法对模型求解,根据问题的特点对蚁群算法的编码方式和选择概率操作进行改进,并通过构造基因交换和协同变异两种单亲遗传算子,克服蚁群算法收敛速度慢、易陷入局部最优的不足。最后,通过某汽车制造企业生产实例和一组实验验证了模型的正确性,证明了本文构建的算法在求解质量、运算速度及收敛性方面具有优势,且能够减少循环配送次数,有效控制线边库存量,提高牵引车满载率。For material milk-run delivery problem in mixed-model assembly line with multiple tours,large inventories,a mathematical programming model is presented with consideration of actual production and trade-off conflict between the number of tours and inventories.Subsequently,a partheno-genetic ant colony optimization(PG-ACO)algorithm by improving the coding mode and possible selection is proposed.Both the gene exchange and cooperative mutation operators are constructed to optimize the algorithm,respectively.It overcomes deficiencies of the traditional ant colony algorithm,such as slow convergence speed and tendencies to trap into local optima.Finally,an example of an automobile manufacturing enterprise and a group of experiments are used to verify the correctness of the model.The results reveal that PG-ACO algorithm has the superiority for computing the speed,quality and good performance of coping with the material milk-run delivery problem in a mixed-model assembly line and can reduce the number of tours,effectively control the inventory level and improve the full load rate of tow train.

关 键 词：循环配送 混流装配线 物料配送 单亲遗传蚁群算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]