Ride-matching for the ride-hailing platform with heterogeneous drivers

Abstract

We study a ride-matching problem in a ride-hailing platform with heterogeneous drivers. Drivers are split into two categories, part-time drivers and full-time drivers , based on their level of reliance on the platform. Part-time drivers are less patient than full-time drivers and more sensitive to waiting time in the matching process, which causes uncertainty in the supply for ride-hailing platforms. Motivated by this uncertainty, we propose a priority-based matching policy (PMP), where the platform gives part-time drivers matching priority over full-time drivers when making matches. We prove that the ride-hailing platform can always benefit from PMP on measures, such as the number of matches (passenger–driver pairs) or total revenue. We identify the principal factors in determining the optimal matching priority ratio to be applied in PMP; these include the passenger arrival rate, the arrival rate of part-time drivers, and the average patience level of part-time drivers. We also show that under certain market conditions, PMP is a negative policy for full-time drivers, e.g., reducing their average income. Using a subsidy, the platform can compensate for the lost income of full-time drivers. This creates a triple-win situation for the platform, drivers and passengers.