EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

2233. Feasibility Prediction in Attended Home Delivery with Uncertainty: A Machine Learning Approach

Invited abstract in session MB-29: Combinatorial Optimization models and applications in Logistics and Transportation II, stream Combinatorial Optimization.

Monday, 10:30-12:00
Room: 157 (building: 208)

Authors (first author is the speaker)

1. Liana van der Hagen
Zero Hunger Lab, Tilburg University
2. Niels Agatz
Decision and Information Science, RSM University
3. Remy Spliet
Econometric Institute, Erasmus University Rotterdam

Abstract

In attended home delivery, retailers usually let customers select a delivery time slot for receiving their orders. The delivery capacity, i.e., the number of vehicles and drivers, is often fixed and inflexible in the short term. To effectively use their delivery capacity, e-grocers may dynamically close time slots for certain new customers given the already accepted customer orders. One complicating factor is that for example online grocers allow customers to change their order basket at any time before the cut-off. Consequently, the e-grocer is uncertain about how much vehicle capacity should be reserved for each of the customers during the booking process. We study the challenges that arise with this order size uncertainty and propose strategies to deal with it. For each of the time slots to be evaluated, we use machine learning to quickly predict the probability that a feasible route plan exists that visits all accepted customers in their selected time slot.

Keywords

Status: accepted


Back to the list of papers