Assortment Allocation to Distribution Centers to Minimize Split Customer Orders

Operations and Information Management Department; Faculty Adviser: Marshall Fisher

Yihaodian is a key player in the Chinese online retail space and one of China’s fastest growing retailers. In collaboration with the company’s management, this research explores split orders, which occur whenever an order is fulfilled in more than one shipment. This entails additional costs to the retailer. In the context of electronic commerce retail, we study the problem of assigning stock-keeping units to distribution centers (DCs) to minimize split orders. We show that this problem is NP-hard, propose several heuristics and test them against other benchmark algorithms using actual transaction data from a Chinese online retailer. We evaluate the results of the test along three dimensions: the average number of shipments per order, workload balance across the DCs, and the number of DCs that are capable of fulfilling a particular order.