Content area
Full Text
Abstract
In this work, we propose an approach that leverages historical customer purchase conversion rates when designing a middle-mile consolidation network that aims to maximize the profit of large e-commerce retailers. We embed lead-time dependent sales volume predictions into a new mixed-integer program (MIP) that simultaneously determines shipment lead times and consolidation plans to maximize profit, i.e.. sales revenue net logistics cost. To find high-quality consolidation plans for large, practically-sized instances, we build an adaptive IP-based local search solution approach. Preliminary results from a U.S.-based e-commerce partner show that incorporating historical customer purchase conversion rates into the consolidation network design increase profit by over 20% depending on the retailer's desired flexibility when selecting lead times.
Keywords
Service network design, e-commerce logistics, middle mile, local search
Georgia Institute of Technology, Atlanta, GA, USA
1. Introduction
New e-commerce retailing models such as ship-to-home and ship-to-store require that retailers implement new proaches for directly to consumers on demand. Some high-demand stock-keeping units (SKUs) may be held at fulfillment centers (FCs), while others may be shipped directly from vendors. Middle-mile consolidation network design problems arise when large retailers jointly coordinate shipments from vendors and FCs into last-mile distribution (LMD) facilities, seeking consolidated truckloads when possible to reduce high less-than-truckload (LTL) and parcel freight charges. In the U.S., Amazon, Wayfair, and The Home Depot are shippers who coordinate such networks. In
this paper, we consider the problem where the retailer must ship orders over tüne from known origin stocking locations (FCs or vendor locations) to known destination LMD facilities. Examples of such LMD facilities may be those operated by package transportation companies or postal services (e.g., UPS), branded delivery subsidiaries (e.g., Amazon Prune), and/or less-than-truckload (LTL) carriers or local large-and-bulky delivery companies. Each customer is quoted a lead time for their order; thus, shipments must move from their origin to their LMD destination to meet a tüne deadline. The retailer also has lead-time dependent sales volume predictions that they can use to select the lead tüne shown to customers. Because there is competition among e-commerce retailers to serve customers rapidly, one way to maximize sales revenue could be to offer and meet shorter lead tüne promises for products with time-sensitive customers. To minimize the cost of meeting these deadlines,...