Navigating the queue: A dynamic pricing framework for ride-sourcing platform with heterogeneous demand

Ride-sourcing platforms commonly employ dynamic pricing to regulate travel demand and dispatch or relocate vehicles to manage supply distribution. However, existing strategies often overlook the heterogeneity in spatial-temporal demand distribution and diverse customer behaviors. To address this gap, this paper proposes an origin-destination (OD)-based dynamic pricing framework that integrates link-node-level vehicle dispatching and relocation decisions, while accounting for heterogeneous demand and customer queue-jumping behavior. In this framework, customers, characterized by varying values of time (VOT) and willingness to pay (WTP), can choose to wait, abandon the trip, or pay a premium to jump the queue. To model this heterogeneous behavior, we introduce an integral-form demand function that captures elasticity with respect to both VOT and WTP. The dynamic pricing problem is formulated as a constrained optimization problem that jointly determines OD-based pricing and queue-jumping premiums. A rolling horizon approach, featuring a short decision interval with a longer planning horizon, is introduced to capture queueing dynamics that extend across periods, thereby supporting both real-time responsiveness and long-term optimization. An event-triggered control scheme is further incorporated to reduce computational complexity and stabilize pricing fluctuations. An Approximate Dynamic Programming (ADP) approach is adopted to manage vehicle dispatching and relocation at the link-node level. Empirical validation on the Manhattan road network confirmed the effectiveness of the proposed approach in improving platform profitability, increasing demand satisfaction, and enhancing overall system efficiency.