Connected and Autonomous Shuttles for Optimal Passenger Transportation and Last-Mile Parcel Delivery
Auteurs
Stolfi D.H., Chaalal E., Faye S.
Référence
Gecco 2025 Companion Proceedings of the 2025 Genetic and Evolutionary Computation Conference Companion, pp. 907-910, 2025
Description
This article proposes the use of connected and automated shuttles (CAS) as a transformative approach to public transportation, which can be also used for delivering parcels to pick-up/drop-off (PUDO) points, located close to bus stops. A multi-objective optimisation framework for CAS systems is designed, focusing on three critical dimensions: Quality of Service (QoS), Transit Network Design (TND), and User Travel Demand (UTD). Additionally, we aim to share passenger transport and last-mile delivery, taking advantage of empty places on shuttles to transport mail baskets optimally filled by parcels. A genetic algorithm is proposed to assign parcels to shuttles taking into account the parcels’ size and destination, different shuttles’ stops, and space available on them. Our results, obtained from the optimisation of urban transit scenarios, show the framework’s applicability and its effectiveness, understanding the trade-offs between competing objectives as well as a reduction in the number of parcels to be delivered by the postal service.
