A Decentralised Auction-Based Framework to Proactively Rebalance Ride-Hailing Fleets under Demand Uncertainty

On-demand mobility services are transforming urban mobility. They can provide individual and collective benefits when managed optimally, and their successful integration within the existing urban transport system can enhance its performance. In contrast, inadequate fleet management can inflict high pick-up waiting times and passenger drop-out rates. One of the main challenges for on-demand mobility service operators is to proactively rebalance their fleets to ensure that the spatial distribution of supply matches the demand. This paper proposes to address this problem with a distributed auctioning approach. We design an architecture that relies on local controllers interacting with idle vehicles, encouraging them to relocate to their service area. We conduct simulations on the city of Lyon in France, which reveal a substantial increase in the number of passengers served compared to a scenario without rebalancing