A mixed truck and robot delivery approach for the daily supply of customers

Abstract

Innovative last-mile logistics solutions are needed to reduce delivery costs, traffic congestion, and pollution in cities. A promising concept in this context are truck-and-robot systems, as they enable significant cost and traffic reduction compared to classic truck deliveries. The system relies on small autonomous delivery robots to cover the last meters to a customer. Existing truck-and-robot concepts to date consider home deliveries by robots, while trucks are only used to transport robots and not for deliveries. This assumption disregards the fact that regular truck deliveries are still needed for some delivery requests, such as for the delivery of bulky items, or for customers who do not accept robots. Our research addresses this issue and proposes a mixed truck and robot delivery concept in which both robots and the delivery truck can visit customers. Our tailored solution approach is based on a General Variable Neighborhood Search that efficiently solves the routing problem and outperforms existing truck-and-robot routing algorithms. The numerical experiments show that this approach enables cost reductions of up to 43% compared to classical truck deliveries and up to 22% compared to a truck-and-robot system that does not allow deliveries by both truck and robots on the same tour. Further analyses reveal additional benefits of such mixed tours and the robustness of our approach for different problem settings.