Same-Day delivery with pickup stations and autonomous vehicles

Abstract

Same-day delivery has become a very important challenge for e-commerce providers. However, effective delivery concepts are still not established. In this paper, we analyze the potential of combining parcel pickup stations and autonomous vehicles for same-day delivery. This combination may allow for consolidation of goods, automated delivery processes, and reduced operation costs. We consider the problem on the operational planning level. A depot, a set of pickup stations, and an autonomous delivery fleet are given. Customers dynamically order goods to a preferred pickup station and expect fast service. The goods need to be delivered from the depot to the preferred pickup station or another pickup station in the close neighborhood. Autonomous vehicles are directly dispatched between the depot and the stations. To decide where and when to dispatch a vehicle and about the corresponding goods to load, we present a policy function approximation (PFA). Our intuitive PFA allows real-time decision making while balancing the tradeoff between fast delivery and consolidation. We conduct an extensive computational study for the pickup station network of the city of Braunschweig. We show how our delivery concept allows fast delivery for up to 100 delivered goods per vehicle and day. We further pair the PFA with a value function approximation to account for heterogeneity in demand. We finally derive important managerial insights for strategical and tactical planning and present an extensive outlook on promising future research directions.