Order Fulfillment in Online Retailing: What Goes Where

We present three problems motivated by order fulfillment in online retailing. First, we focus on one warehouse or fulfillment center. To optimize the storage space and labor, an e-tailer splits the warehouse into two regions with different storage densities. One is for picking customer orders and the other to hold a reserve stock that replenishes the picking area. Consequently, the warehouse is a two-stage serial system. We investigate an inventory system where demand is stochastic by minimizing the total expected inventoryrelated costs subject to a space constraint. We develop an approximate model for a periodic review, nested ordering policy. Furthermore, we extend the formulation to account for shipping delays and advance order information. We report on tests of the model with data from a major e-tailer. Second, we focus on the entire network of warehouses and customers. When a customer order occurs, the e-tailer assigns the order to one or more of its warehouses and/or dropshippers, so as to minimize procurement and transportation costs, based on the available current information. However, this assignment is necessarily myopic as it cannot account for any subsequent customer orders or future inventory replenishments. We examine the benefits from periodically re-evaluating these real-time assignments. We construct nearoptimal heuristics for the re-assignment for a large set of customer orders by minimizing the total number of shipments. Finally, we present saving opportunities by testing the heuristics on order data from a major e-tailer. Third, we focus on the inventory allocation among warehouses for low-demand SKUs. A large e-tailer strategically stocks inventory for SKUs with low demand. The motivations are to provide a wide range of selections and faster customer fulfillment service. We assume the e-tailer has the technological capability to manage and control the inventory globally: all warehouses act as one to serve the global demand simultaneously. The e-tailer will utilize its entire inventory, regardless of location, to serve demand. Given we stock certain units of system inventory, we allocate inventory to warehouses by minimizing outbound transportation costs. We analyze a few simple cases and present a methodology for more general problems. Thesis Supervisor: Stephen C. Graves Title: Abraham J. Siegel Professor of Management Science & Engineering Systems