Rapidly advancing on-demand ridesharing services, including those with self-driving technologies, hold the promise to revolutionize delivery of mobility. Yet, significant imbalance between spatiotemporal distributions of vehicle supply and travel demand poses a pressing challenge. This paper proposes a multi-period game-theoretic model that addresses dynamic pricing and idling vehicle dispatching problems in the on-demand ridesharing systems with fully compliant drivers/vehicles. A dynamic mathematical program with equilibrium constraints (MPEC) is formulated to capture the interdependent decision-making processes of the mobility service provider (e.g., regarding vehicle allocation) and travelers (e.g., regarding ride-sharing and travel path options). An algorithm based on approximate dynamic programming (ADP), with customized subroutines for solving the MPEC, is developed to solve the overall problem. It is shown with numerical experiments that the proposed dynamic pricing and vehicle dispatching strategy can help ridesharing service providers achieve better system performance (as compared with myopic policies) while facing spatial and temporal variations in ridesharing demand.

包括自动驾驶技术在内的按需乘车共享服务迅速发展，有望彻底改变交通出行方式。然而，车辆供应的时空分布与旅行需求之间的严重失衡提出了紧迫的挑战。本文提出了一种多阶段博弈论模型，该模型解决了具有完全兼容的驾驶员/车辆的按需乘车共享系统中的动态定价和怠速车辆调度问题。制定了具有平衡约束的动态数学程序（MPEC），以捕获出行服务提供商（例如，关于车辆分配）和旅行者（例如，关于乘车共享和行进路线选项）的相互依赖的决策过程。一种基于近似动态规划（ADP）的算法，带有用于解决MPEC的定制子例程的解决方案，旨在解决整个问题。数值实验表明，提出的动态定价和车辆调度策略可以帮助拼车服务提供商在面对拼车需求时空变化的同时，实现更好的系统性能（与近视策略相比）。