并行分区拣选的自动分拣机系统拣选作业建模及优化研究

The stochastic order fulfillment of automated picking system is the key business for logistics distribution center. Its working efficiency directly affects the construction investment and operation costs. Along with widespread application in pharmaceutical, tobacco industry, the academic research on layout and picking efficiency of automated picking system gains popularity. In the context of determinated quantity of automated picking machines, this project focuses on how to increase efficiency of automated picking system by means of optimizing layout and picking task assignment based on parallel picking zone strategy, and then it analyzes relationship between principal influence factors and picking efficiency in order to gain performance of balance. In the first place, various kinds of automated picking system will be abstracted into minimized picking time models on account of queuing system. Then, minimized picking models will be transformed into optimal order assignment based on relaxation method for the purpose of researching on SKU (Stock Keeping Unit) assignment and SKU quantity splitting. Furthermore, solution space of picking zones will be reduced by using lower bound function. Finally, device parameters and number of picking zones will be selected as input parameters; improved ant systems will be presented for task scheduling to gain minimal picking time as results of SKU assignment and SKU quantity splitting; inner relationship between input parameters and picking efficiency will be analyzed. Therefore, this research will not only help system designers determinate layout and machine types for the automated picking system, but also has significant implications for design and improvement of picking strategy of automated picking system.

随机订单自动拣选作业是配送中心的核心业务，其作业效率直接影响其运行成本，随着医药、烟草等多行业自动分拣机系统的广泛应用，分拣机设备参数选型、布局设计、拣选策略等优化问题成为研究的重点。本课题通过设计排队系统抽象各类并行分区拣选的自动分拣机系统，建立可变虚拟视窗的分区拣选作业时间模型；为降低模型求解复杂度，将目标模型转化为最佳订单拣选量分配，提出利用模型下界函数与分区数量的单调性缩小分区数量的解空间，设计静态和动态相结合的复合聚类求解算法进行品项分配优化；将设备参数及分区数量作为输入参数，利用品项分配和拆分的结果，设计最大最小蚁群算法调度分拣机拣选顺序，求解设备作业效率，分析设备参数与拣选效率的内在规律，获得各个参数的平衡点。该课题的研究既能指导自动分拣机系统的布局规划和设备选型，又为并行分区拣选的自动分拣机系统拣选策略的优化研究奠定基础。

配送中心的订单拣选是一个随机服务系统，整个流程的服务时间是衡量配送中心作业效率的指标，其中拣选服务时间是关键。配备输送系统的货到订单的自动分拣机系统，因其订单作业时间不包含拣货员或拣选设备寻找货位的时间，其作业效率明显高于货到人的作业模式，在大型配送中心中得到广泛的使用。随机订单自动拣选系统的作业效率直接影响其运行成本，随着医药、烟草等多行业自动分拣机系统的广泛应用，分拣机设备参数选型、布局设计、拣选策略等优化问题成为研究的重点。本课题通过简化合并次要影响因素，依据订单拣选系统中的订单队列特点，设计了排队系统描述并行分区拣选下分拣机系统工作流程,建立了可变虚拟视窗的分区拣选作业时间模型。为降低模型求解复杂度，将目标模型转化为最佳订单拣选量分配，利用模型下界函数与分区数量的单调性缩小分区数量的解空间，设计静态和动态相结合的复合聚类求解算法进行品项分配和品项拆分优化。将设备参数及分区数量作为输入参数，利用品项分配和拆分的结果，设计最大最小蚁群算法调度分拣机拣选顺序，求解设备作业效率。利用正交实验设计的方法对自动拣选系统的运行参数进行分析，分析各参数因素水平高低对拣选效率的影响规律，利用级差分析法定量确定各参数水平的影响；在各参数及各参数水平的不同搭配设计中，找出优化的组合方案。针对自动分拣机系统补货缓存为研究对象，通过对货物在补货缓存中的作业过程分析，建立货物在输送设备上达到输入输出平衡的数学模型，在计算排队系统的设备服务率和平均作业时间的基础上，求解最大补货长度和各个临界补货长度，从而确定补货缓存的最优长度。该课题的研究既能指导自动分拣机系统的布局规划和设备选型，又为并行分区拣选的自动分拣机系统拣选策略的优化研究奠定基础。