Dynamic optimization strategies for on-demand ride services platform: Surge pricing, commission rate, and incentives

Abstract

On-demand ride services reshape urban transportation systems, human mobility, and travelers' mode choice behavior. Compared to the traditional street-hailing taxi, an on-demand ride services platform analyzes ride requests of passengers and coordinates real-time supply and demand with dynamic operational strategies in the ride-sourcing market. To test the impact of dynamic optimization strategies on the ride-sourcing market, this paper proposes a dynamic vacant car-passenger meeting model. In this model, the accumulative arrival rate and departure rate of passengers and vacant cars determine the waiting number of passengers and vacant cars, while the waiting number of passengers and vacant cars in turn influence the meeting rate (which equals to the departure rate of both passengers and vacant cars). The departure rate means the rate at which passengers and vacant cars match up and start a paid trip. Compared with classic equilibrium models, this model can be utilized to characterize the influence of short-term variances and disturbances of current demand and supply (i.e., arrival rates of passengers and vacant cars) on the waiting numbers of passengers and vacant cars. Using the proposed meeting model, we optimize dynamic strategies under two objective functions, i.e., platform revenue maximization, and social welfare maximization, while the driver's profit is guaranteed above a certain level. We also propose an algorithm based on approximate dynamic programming (ADP) to solve the sequential dynamic optimization problem. The results show that our algorithm can effectively improve the objective function of the multi-period problem, compared with the myopic algorithm. A broader range of surge pricing and commission rate and the introduction of incentives are helpful to achieve better optimization results. The dynamic optimization strategies help the on-demand ride services platform efficiently adjust supply and demand resources and achieve specific optimization goals.