Reverse-Auction-Based Competitive Order Assignment for Mobile Taxi-Hailing Systems

Mobile Taxi-Hailing (MTH) is one of the most attractive smartphone applications, through which passengers can reserve taxis ahead for their travels so that the taxi service’s efficiency can improve significantly. The taxi-hailing order assignment is an important component of MTH systems. Current MTH order assignment mechanisms fall short in flexibility and personalized pricing, resulting in an unsatisfactory service experience. To address this problem, we introduce a Competitive Order Assignment (COA) framework for the MTH systems. The COA framework mainly consists of the Multi-armed-bandit Automatic Valuation (MAV) mechanism and the Reverse-auction-based Order Assignment (ROA) mechanism. The taxis apply the MAV mechanism to automatically generate the transport service valuations for orders. The platform applies the ROA mechanism to complete each round of order assignment. Then, we analyze the online performance of MAV, and prove that ROA satisfies the properties of truthfulness and individual rationality. Finally, we also demonstrate the significant performances of MAV and ROA through extensive simulations on a real trace.