Research on Taxi Search Strategy based on Passenger Distribution and Traffic Conditions

Abstract

With the rising smart transportation and sharing economy, taxis play an increasingly important role in urban public transportations. While researches focus mostly on the operation of taxis based on the long-term traffic supply and demand equilibrium, it is essentially unknown how the real-time recognition of passenger demand and traffic conditions will influence the outcome of the taxi search strategies. Based on Beijing traffic data, we study the optimal search strategy of taxis through simulation analysis in real network, considering both the passenger distribution and the traffic conditions. The results show that the taxi driver's passenger empirical index will benefit the search efficiency. More important, it is suggested that while weighing between the passenger distribution and traffic conditions, the optimal search strategy at low passenger density is moving towards areas with more passengers, and the optimal strategy at high passenger density is moving towards areas with better traffic conditions. Our findings can help on-line taxi operators achieve better management efficiency with real-time forecasting of passenger distribution and traffic conditions.