Ride-Sourcing Vehicle Rebalancing with Service Accessibility Guarantees via Constrained Mean-Field Reinforcement Learning

2025-03-31

AI Summary

To address spatiotemporal supply–demand mismatch in ride-hailing—manifesting as prolonged passenger wait times, low vehicle utilization, and spatial inequity in service coverage—this paper proposes a Mean-Field Control Reinforcement Learning (MFC-RL) framework incorporating explicit accessibility fairness constraints. Unlike conventional mean-field approaches that assume strict supply–demand equilibrium, our method explicitly models geographic service fairness as a hard constraint. It jointly optimizes three concurrent operational decisions: ride matching, driver cruising, and vehicle rebalancing. The framework employs continuous-state/continuous-action modeling, stochastic dynamics simulation calibrated on real-world trajectory data, and distributed policy optimization. Evaluated on a city-scale simulation with over ten million orders in Shenzhen, the approach reduces average passenger wait time by 23.7%, increases vehicle utilization by 19.4%, improves service coverage in underserved areas by 31.2%, and enables millisecond-level online dispatch for fleets exceeding 10,000 vehicles.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Search and Optimization: Learning to Search
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Economics, Online Markets and Human Computation: The sharing economy

Abstract

The rapid expansion of ride-sourcing services such as Uber, Lyft, and Didi Chuxing has fundamentally reshaped urban transportation by offering flexible, on-demand mobility via mobile applications. Despite their convenience, these platforms confront significant operational challenges, particularly vehicle rebalancing - the strategic repositioning of thousands of vehicles to address spatiotemporal mismatches in supply and demand. Inadequate rebalancing results in prolonged rider waiting times, inefficient vehicle utilization, and inequitable distribution of services, leading to disparities in driver availability and income. To tackle these complexities, we introduce scalable continuous-state mean-field control (MFC) and reinforcement learning (MFRL) models that explicitly represent each vehicle's precise location and employ continuous repositioning actions guided by the distribution of other vehicles. To ensure equitable service distribution, an accessibility constraint is integrated within our optimal control formulation, balancing operational efficiency with equitable access to the service across geographic regions. Our approach acknowledges realistic conditions, including inherent stochasticity in transitions, the simultaneous occurrence of vehicle-rider matching, vehicles' rebalancing and cruising, and variability in rider behaviors. Crucially, we relax the traditional mean-field assumption of equal supply-demand volume, better reflecting practical scenarios. Extensive empirical evaluation using real-world data-driven simulation of Shenzhen demonstrates the real-time efficiency and robustness of our approach at the scale of tens of thousands of vehicles. The code is available at https://github.com/mjusup1501/mf-vehicle-rebalancing.

Problem

• Addressing vehicle rebalancing in ride-sourcing platforms to reduce rider waiting times.
• Ensuring equitable service distribution across geographic regions via accessibility constraints.
• Modeling real-world stochasticity in transitions and rider behaviors for scalable solutions.

Innovation

• Scalable continuous-state mean-field control
• Accessibility constraint in optimal control
• Real-world data-driven simulation validation