Together with my colleagues Uğur Arıkan, Baris Cem Sal, Severin Schmitt and Jonas Witt from DHL Data & Analytics, I published a paper in the INFORMS Journal on Applied Analytics (volume 53, issue 5, September–October 2023). It describes how Deutsche Post plans the daily routes of its parcel couriers and predicts delivery time windows for its customers.
With this work, our team was a finalist for the 2022 Daniel H. Wagner Prize for Excellence in the Practice of Advanced Analytics and Operations Research, and the paper appeared in the journal’s special issue for the prize.
The problem
Estimated delivery time windows and live tracking of shipments play a key role in the customer experience of last-mile delivery. Both depend on two things: reliable knowledge of the order in which a courier will deliver the parcels in a tour, and precise predictions of delivery times.
Our approach
- Courier-centric routing: our routing algorithm combines operations research with statistics and machine learning. It gathers the tacit knowledge of experienced couriers by deducing precedence constraints from historical delivery data, and uses them when generating tours. This lets us provide custom routes for individual couriers.
- Delivery time windows: a machine learning model predicts the delivery time windows. The courier-centric routes proved to be a main driver for making these predictions accurate and for enabling live tracking.
Impact
Deutsche Post uses the solution to plan the daily routes of couriers for approximately 13,000 parcel delivery districts in Germany, and to provide live tracking and estimated delivery time windows for 1.6 million parcels every day.
I also talked about this work in The Erium Podcast.
Citation: Arıkan, U., Kranz, T., Sal, B. C., Schmitt, S., & Witt, J. (2023). Human-Centric Parcel Delivery at Deutsche Post with Operations Research and Machine Learning. INFORMS Journal on Applied Analytics, 53(5), 359–371. https://doi.org/10.1287/inte.2023.0031