Abstract

The design and control of effective sortation systems has become more complex as both the volume of parcels and also the number of time-definite service options offered by parcel carriers have grown. In this article, we describe approaches for planning two-stage parcel sort operations that explicitly consider time deadlines and sorting capacities. In two-stage sorting, parcels are sorted into groups by a primary sorter and then parcels from these groups are dispatched to secondary stations for final sort. We define a sort planning optimization problem in this setting using mixed-integer programming, where the primary objective is to minimize operational cost subject to machine capacity and parcel deadline constraints. Since a detailed optimization problem for sort planning based on flows in a time–space network is difficult to solve for realistically-sized instances, we develop an alternative formulation that is easier to solve and shares the same feasible region of first-stage sorting decisions with the detailed model; for many practical objective functions, this simpler model can be used to find cost-optimal solutions to the detailed model. We illustrate the proposed modeling approach and its effectiveness using real-world instances obtained from a large parcel express service provider.

摘要

随着包裹数量和包裹承运商提供的限时服务选项数量的增长，有效分拣系统的设计和控制变得更加复杂。在本文中，我们描述了明确考虑时间期限和分拣能力的两阶段包裹分拣操作的规划方法。在两阶段分拣中，包裹由一级分拣机分拣成组，然后将这些组中的包裹分派到二级站进行最终分拣。我们使用混合整数规划在此设置中定义了分类规划优化问题，其中主要目标是在机器容量和包裹截止日期限制下最小化运营成本。由于基于时空网络中的流的排序规划的详细优化问题难以解决实际大小的实例，因此我们开发了一种更容易解决的替代公式，并与第一阶段排序决策共享相同的可行区域详细模型；对于许多实用的目标函数，这种更简单的模型可用于为详细模型找到成本最优的解决方案。我们使用从大型包裹快递服务提供商处获得的真实实例来说明所提出的建模方法及其有效性。这个更简单的模型可用于为详细模型找到成本最优的解决方案。我们使用从大型包裹快递服务提供商处获得的真实实例来说明所提出的建模方法及其有效性。这个更简单的模型可用于为详细模型找到成本最优的解决方案。我们使用从大型包裹快递服务提供商处获得的真实实例来说明所提出的建模方法及其有效性。