New Routing Approaches for Bike Deliveries
Status
This project started in 2019 and has been successfully completed.
Researchers
Prof. Rico Zenklusen (ETH)
Simon Bruggmann (ETH)
Martin Nägele (ETH)
Description
In this project, ETH and the Swiss Post/notime AG aim at addressing novel challenges appearing in routing problems for same day deliveries.
The business to customer delivery market has experienced significant changes in recent years. In particular, regarding parcel logistics, same day delivery has become a major component, enjoying continuously increasing demand and generating extensive extra efforts for delivery companies in solving challenges that come along with this growth. A key aspect among these challenges that is of particular importance for same day deliveries is the last mile, where parcels have to be distributed from a depot to customer addresses in the same city within a short time window. The last mile in same day deliveries is often done by bike couriers, which are faster and more flexible in urban areas compared to delivery trucks. Typically, a bike courier loads his cargo bike at the depot once or twice during one delivery time window, and follows a pre-computed delivery route that is provided to him.
From an algorithmic point of view, both the tasks of assigning parcels to bike couriers on the one hand, and defining strong and on-time delivery routes on the other, are highly nontrivial. Needless to say, the two questions are highly dependent, thus forcing a global approach in order to obtain solutions of good overall quality. An additional challenge stems from the need to support unforeseen real-world changes in the problem instance, like potential delays on some of the routes or new/cancelled orders, that have to be anticipated on short notice. These aspects of the problem require the design of very fast routing algorithms that can be run many times during the same day to be able to react to changes and revise decisions taken earlier.
Our goal is to leverage modern techniques from Combinatorial Optimization to design fast routing algorithms addressing the challenges sketched above that return solutions for which we can provide strong guarantees in terms of solution quality.