关键词: 品类拆解, 数量拆解, 合并打包能力, 经验法则, 深度优先搜索

Abstract: The rapid development of the digital economy has led to the prosperity of online retail industry. To meet the individual needs of customers, it causes a lot of order fulfilment problems, especially the problem of order splitting and combining delivery. The existing literature usually only optimizes the two problems independently, ignoring the intrinsic connection that exists between them. And joint research of the two are an effective means to reduce the cost of order fulfilment. The inventory layout of “multiple warehouses in one place” is the key to solving the problem of order fulfilment and through the horizontal transfer between warehouses to achieve splitting of orders and combining delivery. It can meet the individual needs of customers, reduce the number of order deliveries and the delivery cost of orders, and realize the optimal order fulfilment process.
In this paper, we consider the complete order fulfilment process, which is to pick out the order customer's demand by splitting the category and quantity according to the warehouse situation (warehouse location, product category and quantity, etc.) and the order characteristics (customer location, order product category and quantity, order delivery time, etc.) under the warehouse layout model of “multiple warehouses in one place”. Then, the splitting orders are assigned to the appropriate warehouse, and the sub-orders belonging to the same customer are transferred horizontally between warehouses. Finally, the packing operation is performed at the consolidation warehouse, and the combined parcels are delivered to the customer. In addition, this paper constructs a mixed integer planning model by considering the combining and packing capacity and time constraints to minimize order fulfilment cost from three perspectives: order forwarding cost, order combining and pack cost, and order delivery cost.
In order to solve the model, the artificial rule of thumb is applied to construct the order-warehouse priority sequence matrix, and the depth-first search algorithm DFS is designed to quickly generate the initial feasible solution that meets the reality and is better, and the optimization scheme with the minimum cost is obtained by iterating based on the Improved Genetic Algorithm (IAPGA) framework. Since order splitting and consolidated delivery considered is a new combinatorial optimization problem, there are no benchmark calculations in the literature. Therefore, the validity of the model and algorithm are verified by adding data on the ability to construct combined packages and the latest time allowed for delivery of orders. And we set the data in the context of the reality that there may be a mass ordering of products based on two numerical experiments in the relevant literature.
The results show that: (1)It can be obtained that as the number of orders increases, the total cost of order fulfilment increases. At a certain number of orders, with an increase in the number of warehouses and with the order splitting and combining delivery solutions becoming more, the algorithm solution time increases, but the total order fulfilment cost can be reduced to a certain extent. (2)Under different consolidations of inventory size and order quantity, the order fulfilment cost increases as the number of orders increases. And when the number of orders is certain, the order fulfillment cost decreases as the inventory size increases. This is because the increase in inventory size reduces order splitting, and packing costs, transshipment costs, and distribution costs all decrease. (3)Under different consolidations of combined packing capacity and different order quantities, the order fulfilment cost increases as the number of orders increases. When the number of orders is certain, the order fulfilment cost decreases as the combination and packing capacity increases due to the relaxation of the combination and packing capacity constraint, which allows more order packing operations to be completed in the same consolidated warehouse, thus reducing the order transfer distance and delivery distance.

Key words: category splitting, quantity splitting, consolidated packaging capacity, rule of thumb, depth-first search