阵列式自动拣选系统配置优化研究

发布时间：2019-06-04 03:32

【摘要】：随着电子商务和连锁零售行业的快速发展,配送中心的商品拣选日益呈现小批量、多品种、多批次、高时效的发展趋势,对订单拆零拣选作业提出了更高的要求。拆零拣选是指小于一箱数量的货物拣选,由于每个订单包含货物数量较少,但需要从成千上万种品项(the stock keeping unit, SKU)中快速拣选出货物的最小包装单位,且订单数量庞大,所以拆零拣选是影响配送中心作业成本和订单履行效率的重要因素之一。为有效提高拆零拣选效率的同时降低人员劳动强度,阵列式自动拣选系统设计研发并得到成功应用。阵列式自动拣选系统是一类由大量水平倾斜式拣选通道在空间中排列组合而成的新型自动化拆零拣选系统。该系统的使用虽然可以提高拣选效率,但同时需要增加较多的人工补货成本,因此,管理者希望针对现场已有阵列式自动拣选设备找到一种优化的配置方式,实现阵列式自动拣选系统总节省人工成本最大、订单处理总时间最小的目标。目前国内外对于自动化拣选系统的研究多集中在A字机系统,该系统的设备结构和拣选控制方式与阵列式自动拣选机差别较大,多适用于拣选量集中于有限品项的订单拣选。此外,现有关于自动化拣选系统配置优化领域的文献,多是以系统订单处理总时间最小或节省人工成本最大为目标的优化,综合考虑系统成本与效率的多目标优化研究较少。基于此,本文提出以系统总节省人工成本最大、订单处理总时间最小为目标的阵列式自动拣选系统配置优化问题,通过对系统拣选品项选择与通道配比、品项分配以及品项货位分配的优化,实现系统性能的综合提高。本文主要内容与成果如下：(1)系统拣选品项选择与通道配比优化子问题中,以系统总节省人工成本最大为目标的建立数学模型,提出启发式算法求解。以同时采用阵列式自动拣选系统和人工拣选系统的双分拣区为工程应用背景,对该双分拣区内人工成本进行全面分析；在假设拣选品项确定的条件下,以阵列式自动拣选系统总节省人工成本最大为目标建立设备通道配比优化数学模型,设计贪婪算法得出拣选通道合理配比方案；在此基础上,将该问题推广到人工和自动化双分拣区系统品项分配中,归结为一类特殊的背包问题,并给出启发式算法。通过实例仿真从多角度分析了算法的有效性。(2)单机品项分配优化子问题中,以串行合流下订单处理总时间最小为目标建立品项分配数学模型,设计改进品项相似系数与聚类算法进行求解。将每个通道列视为一个拣货区,则阵列式自动拣选机属于分区自动化拣选系统。在串行合流下,设备订单处理总时间等于拣货区拣选时间总和,以最小拣货区拣选时间总和为目标建立品项分配模型。为求解该模型,将不同品项并行作业节省的独立拣选时间作为品项相似系数,提出基于该系数的搜索式层次聚类算法。该算法的核心思想是在满足拣货区内空间约束的条件下,通过将相关性强的品项分配至同一拣货区,增加各拣货区内并行拣货量,同时减少各订单内参与拣选的拣货区个数,实现系统分拣效率的提高。最后实例分析证明了改进品项相似系数的优越性和搜索式层次聚类算法的有效性。(3)单机列品项货位分配子问题中,以并行合流下订单处理总时间最小为目标建立数学模型,在假设品项分配方案确定的条件下,设计聚类算法对模型求解。针对阵列式自动拣选机内拣货区串行拣选、并行合流的工作特点,分析拣货区作业时序,建立单机订单处理总时间数学模型,在假设品项分配方案确定的条件下,依据拣货区内品项集合生成列品项,将其归结为以最大总虚拟视窗时差为目标的列品项货位分配问题。为求解该问题,设计聚类算法,将相关性强的列品项分配至间隔距离远的拣货区存储,以增大总虚拟视窗时差。实例仿真显示算法可有效减少单机订单处理总时间,提高作业效率。(4)根据各子问题讨论结果,研究以系统总节省人工成本最大、订单处理总时间最小为目标的串行合流下系统配置优化问题及综合求解方法。采用主要目标优化法将双目标优化问题转化为以系统总节省人工成本最大为目标、订单处理总时间为约束的单目标优化问题。提出启发式迭代算法求解,该算法首先以系统总节省人工成本最大为目标确定拣选品项和通道配比,然后应用搜索式层次聚类算法将品项在拣货区间分配,确定系统初始配置方案。在此基础上,设计配置方案迭代改进策略对初始配置方案进行调整,迭代优化至恰好满足系统拣选效率要求。实例分析证明了该方法的有效性和优越性。
[Abstract]:With the rapid development of e-commerce and the chain-chain retail industry, the commodity selection of the distribution center increasingly presents the development trend of small-batch, multi-product, multi-batch and high-time-time. The removal and sorting means that the goods are sorted less than a box of goods, and since each order contains fewer goods, it is necessary to quickly pick out the minimum packing unit of the goods from the stock keeping unit (SKU), and the quantity of the order is large, Therefore, it is one of the important factors that affect the operation cost and the order fulfillment efficiency of the distribution center. In order to effectively improve the efficiency of sorting and sorting, the labor intensity of the personnel and the design and development of the array automatic sorting system are reduced, and the successful application is obtained. The array type automatic sorting system is a new type of automatic sorting and sorting system which is formed by arranging a large number of horizontal inclined sorting channels in the space. The use of the system can improve the sorting efficiency, but at the same time, more manual replenishment cost is required, and therefore, the manager wants to find an optimized configuration mode for the existing array automatic sorting equipment on the site, so that the overall saving labor cost of the array automatic sorting system is the largest, The objective of the order processing total time is the minimum. At present, the research of the automatic sorting system at home and abroad is mainly focused on the A-word system, and the device structure and the sorting control method of the system are different from the array type automatic sorting machine, so that the sorting quantity is applicable to the order picking of the limited items in the picking quantity. In addition, the existing literature on the optimization field of automatic picking system configuration is the optimization of the goal of minimizing the total time or saving the labor cost with the system order, and the multi-objective optimization of the system cost and efficiency is less. Based on this, this paper presents an array-type automatic picking system configuration optimization problem with the largest total cost of the system and the minimum total time of the order processing. The optimization of the selection and channel matching, the item allocation and the distribution of the goods item in the system is proposed. And the comprehensive improvement of the system performance is realized. The main contents and results of this paper are as follows: (1) The selection of the system's picking item and the optimization of the channel ratio, and the establishment of the mathematical model with the maximum total labor cost of the system as the target, and the heuristic algorithm to solve the problem. the double-sorting area of the array type automatic sorting system and the manual sorting system is simultaneously adopted as the engineering application background, and the labor cost in the double-sorting area is comprehensively analyzed; and under the condition that the sorting item is determined, in that method, an array-type automatic sorting system is used for automatically saving the labor cost and the target establishment equipment channel ratio optimization mathematical model, and a greedy algorithm is designed to obtain a reasonable proportioning scheme of the sorting channel; on the basis of that, the problem is generalized to the product item distribution in the manual and automatic double-sorting area, It boils down to a class of special knapsack problem and gives a heuristic algorithm. The effectiveness of the algorithm is analyzed from the multi-angle through the example simulation. (2) In the problem of single-machine item allocation and optimization, the mathematical model of the item allocation is set up with the minimum total time of the order processing at the serial confluence, and the similarity coefficient of the product item and the clustering algorithm are designed to solve the problem. Each channel column is considered as a pick-up area, and the array automatic sorting machine belongs to the partition automatic sorting system. In that serial confluence, the total time of the equipment order proces is equal to the sum of the picking time of the picking area, and the sum of the picking time of the minimum picking area is the target establishment item distribution model. In order to solve the model, a search-based hierarchical clustering algorithm based on this coefficient is proposed. The core idea of the algorithm is to improve the sorting efficiency of the system by distributing the item items with strong correlation to the same picking area under the condition that the space constraints in the picking area are met, increasing the parallel picking quantity in each sorting area, and simultaneously reducing the number of picking areas participating in the sorting in each order. At last, the advantage of the similarity coefficient of the improved item and the effectiveness of the search-based hierarchical clustering algorithm are proved. (3) In the problem of the allocation of the single-stand column item, the mathematical model is built with the minimum total time of the order processing under the parallel merging, and the model is solved by the design clustering algorithm under the condition that the item allocation scheme is determined. aiming at the working characteristics of serial sorting and parallel merging of a picking area in an array type automatic sorting machine, the working time sequence of the picking area is analyzed, a single-machine order processing total time mathematical model is established, It is attributed to the problem of the allocation of the item items with the maximum total virtual window time difference as the target. In order to solve the problem, a clustering algorithm is designed, and the relevant strong column items are distributed to the picking area far from the interval to increase the total virtual window time difference. An example simulation display algorithm can effectively reduce the total time of single-machine order processing and improve the operation efficiency. (4) Based on the results of each sub-problem, the system configuration optimization problem and the comprehensive solution method of the serial confluence with the maximum total cost of the system, the minimum total time of the order processing and the minimum total time of the order processing are studied. The main objective optimization method is used to transform the double objective optimization problem into a single objective optimization problem with the maximum total cost of the system as the target and the total time of the order processing as the constraint. In this paper, a heuristic iterative algorithm is proposed to solve the problem. The algorithm first determines the picking item and the channel ratio with the maximum of the total cost of the system. Then, the search-type hierarchical clustering algorithm is applied to allocate the product items in the picking interval, and the initial configuration scheme of the system is determined. On this basis, the design and configuration scheme iteration improvement strategy is used to adjust the initial configuration scheme, and the iteration is optimized to just meet the system picking efficiency requirements. The example analysis shows the effectiveness and superiority of the method.
【学位授予单位】：山东大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：TH692

【参考文献】