Information-based taxi-passenger matching management in transportation hubs: A double-ended queuing perspective

To alleviate the imbalance between taxi supply and demand at transportation hubs, this paper proposes a double-ended queuing system to explore the taxi-passenger match mechanisms. Based on Markov theory, we construct a queuing model framework that incorporates different information visibility: unobservable (nothing provided), partially observable (expected waiting time provided), and fully observable (real-time waiting time provided). The dynamic evolution, release scenarios, and queuing optimization potential under each case are in detail explored. In the unobservable case, we analyze the birth-death process of the queuing system. In the other two cases, we derive the equilibrium strategies under varying conditions and study the impacts and applicability of information. Additionally, we propose socially optimal strategies to study the potential and directions for optimization. Finally, we carry out some case studies to validate the above results. The results show that partially observable information helps balance arrivals at both ends and enhances social welfare while leaving a small optimization margin. It is recommended to provide expected waiting time if the capacity is higher on the side with higher arrival rates, while full observable information can enhance social welfare and leave greater room for optimization. However, it is recommended to provide real-time waiting time when demand is high or when the capacity on the side with lower arrival rates is limited.