Operations Research 2025
Abstract Submission

2302. The dynamic same-day delivery problem with walking, parking and stochastic customer requests

Invited abstract in session WE-1: Delivery problems, stream Mobility, Transportation, and Traffic.

Wednesday, 16:30-18:00
Room: Audimax

Authors (first author is the speaker)

1. Ninja Scherr
University of Vienna
2. Sara Reed
University of Kansas
3. Ann Campbell
Management Sciences, The University of Iowa
4. Barrett Thomas
Management Sciences, University of Iowa

Abstract

Due to the rise in e-commerce sales, logistics service providers need to transport increasing volumes of goods to customers’ homes. In addition, the trend for same-day delivery leads to shorter, more fragmented delivery tours as well as dynamic decision making on when to dispatch vehicles. These shorter delivery tours typically exhibit smaller consolidation potential leading to inefficient use of the delivery driver. Thus, delivery companies seek ways to improve routes to best use drivers’ available time. One aspect still not considered widely in the same-day delivery research is the difficulty of parking. Often, same-day delivery models assume that the delivery driver can park at each customer location to deliver the goods. However, in many cities, parking spaces are scarce and congestion leads to longer parking search times. When planning delivery tours in such settings, a more efficient approach may be to park once and deliver packages to several customers by foot from that parking spot. Yet, most models in the same-day delivery literature also do not consider walking by the driver. In this talk, we present the same-day delivery problem with walking, parking, and stochastic customer requests. In this problem, customers request service and expect delivery on the day of their order. The service provider may aim to maximize the overall profit or the number of customers that can be accepted over the day with given resources. We describe the problem and model it as an MDP. Then, we propose different policies for the problem that differ in how they delay the dispatching of the delivery vehicle. In a computational study, we show the impact of the different policies on the number of accepted customers, tour lengths, and overall costs for different parking search times.

Keywords

Status: accepted


Back to the list of papers