物流配送中心分区自动分拣系统品项分配方法研究

发布时间：2018-06-13 04:07

本文选题：物流配送中心 + 自动分拣系统　；参考：《兰州交通大学》2017年硕士论文

【摘要】：随着物流配送中心的快速发展和人们对物流服务高时效的要求,现代物流配送中心对订单分拣作业的要求不断增高。为了提高分拣作业的效率,降低分拣作业时间,各个物流配送中心逐渐引进快速高效的分区自动分拣系统。由于自动分拣系统的物流成本较高,在不明显增加物流成本的前提下,如何进一步提高自动分拣系统的效率是现在物流配送中心研究的重点和难点。由于物流配送中心品项较多,订单结构复杂,如何进行合理的品项分配,提高设备的分拣效率,一直是物流配送中心亟待解决的问题。品项分配是将货物按品项分配到适合位置上的过程,它决定了各订单在各分拣区和各分拣通道的分拣量和各品项并行分拣程度,并且对订单分拣总时间有很大影响。基于此,论文以分区自动分拣系统为研究对象,从分区之间品项分配方法和分区内部各通道之间品项分配方法两方面进行研究,为物流配送中心分区自动分拣系统进行合理品项分配提供了理论指导。在分区之间品项分配方法研究中,对分区自动分拣系统的分拣作业流程进行了深入的分析,得出延迟时间是订单处理总时间的唯一变量。由于延迟时间的计算是一个复杂的递推过程,论文将学者Jane在人工分拣系统中提出的品项相似系数(表示任意两品项间的相关性)概念引入到自动分拣系统中,并根据自动分拣系统自身的特点加入分拣量因子对其进行改进。通过理论证明了延迟时间与品项相似系数和呈正相关性,将优化目标从减少订单处理总时间简化为减少延迟时间,进一步转化为减少品项相似系数和。论文基于优化目标,建立了基于品项相似系数的品项分配模型,并利用动态聚类算法和改进的动态聚类算法进行求解。通过实验证明了两种算法求解结果均优于品项顺序分配的结果,基于禁忌搜索算法改进的动态聚类算法结果优于动态聚类算法的求解结果。在分区内部品项分配方法研究中,在串行分拣策略基础上论文设计了混合分拣策略:先对能够进行并行分拣的品项进行逐批次分拣,最后不能进行并行分拣的品项进行串行分拣。基于此分拣策略,论文建立了分拣作业时间模型和品项分配模型,运用改进的小生境遗传算法对模型进行求解,通过算例仿真证明了混合分拣策略分拣时间优于串行分拣策略分拣时间;在品项分配一样的情况下,改进的小生境遗传算法求得的结果优于基本遗传算法求得的结果。
[Abstract]:With the rapid development of logistics distribution center and the requirement of high efficiency of logistics service, the demand of modern logistics distribution center for order sorting is increasing. In order to improve the efficiency of sorting operations and reduce the time of sorting operations, each logistics distribution center gradually introduced a fast and efficient automatic sorting system. Due to the high logistics cost of the automatic sorting system, how to further improve the efficiency of the automatic sorting system is the focus and difficulty of the current logistics distribution center research on the premise of not obviously increasing the logistics cost. Because the logistics distribution center has many items and the order structure is complex, how to distribute the items reasonably and improve the sorting efficiency of the equipment has always been an urgent problem to be solved in the logistics distribution center. The distribution of items is the process of distributing goods according to items to a suitable position. It determines the sorting quantity of each order in each sorting area and each sorting channel and the degree of parallel sorting of each item, and it has a great influence on the total time of order sorting. Based on this, this paper takes the automatic sorting system as the research object, from two aspects: the distribution method of items between partitions and the methods of distribution of items between channels within partitions. It provides theoretical guidance for rational distribution of items in automatic sorting system of logistics distribution center. In the study of the method of item allocation among different partitions, the sorting process of automatic sorting system is analyzed in depth, and it is concluded that the delay time is the only variable of the total order processing time. Because the calculation of delay time is a complicated recursive process, the concept of similarity coefficient of items (representing the correlation between any two items) proposed by scholar Jane in the manual sorting system is introduced into the automatic sorting system. According to the characteristics of the automatic sorting system, the sorting quantity factor is added to improve it. It is proved by theory that the delay time is positively correlated with the similarity coefficient of product items, and the optimization objective is simplified from reducing the total processing time of order to reducing the delay time, and further to reducing the similarity coefficient of items. Based on the objective of optimization, the model of item assignment based on item similarity coefficient is established and solved by dynamic clustering algorithm and improved dynamic clustering algorithm. The experimental results show that the results of the two algorithms are better than the results of the sequential distribution of items, and the improved dynamic clustering algorithm based on Tabu search algorithm is better than the dynamic clustering algorithm. On the basis of serial sorting strategy, a hybrid sorting strategy is designed in this paper: first, batch by batch sorting is carried out for items that can be sorted in parallel. Finally, the items that can not be sorted in parallel are sorted serially. Based on this sorting strategy, the sorting time model and item allocation model are established, and the improved niche genetic algorithm is used to solve the model. The simulation results show that the sorting time of hybrid sorting strategy is better than that of serial sorting strategy, and the result of improved niche genetic algorithm is better than that of basic genetic algorithm.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F253.9;TP311.13

【参考文献】