Abstract Time-dependent order cancellation behaviour affects ridesourcing platform operation and overall system performance. This paper models impacts of time-dependent order cancellation behaviour on the driver–rider matching efficiency and examines ridesourcing platform operation strategies, where taxi is the alternative to the ridesourcing mode. A time-dependent matching model is developed to characterize the complex interaction among participants (drivers and riders) considering ridesourcing order cancellation behaviour, where the impact of the cancelled orders on the numbers of drivers and riders is explicitly modelled. In particular, we formulate the time-dependent order cancellation rates before and after the matching stage (i.e., cancellation of unconfirmed and confirmed orders). Under the proposed model, a platform profit maximization problem is formulated and three pricing strategies are examined. Our numerical studies demonstrate that the dynamic pricing (i.e., customer/rider fare and driver wage) can well accommodate time-dependent system inputs (e.g., demand rates) and thus enable the platform to increase profit via better market segmentation. We also investigate the objective of maximizing the number of completed trips and examine the trade-off between the platform profit and the number of completed trips. In addition, we show that the relaxation of upper bounds of the ridesourcing fare and order cancellation penalty can increase the ridesourcing platform’s profit and indirectly improve the utilization rate of taxis as well as the taxi company’s profit.

Highlights • This paper models impacts of time-dependent order cancellation behaviour. • Cancellations of unconfirmed and confirmed orders are both considered. • Pricing/penalty strategies are examined under different regimes. • Relaxation of pricing/penalty regulation can increase profits and taxi utilization.