213 / 2024-04-23 12:39:06

Probabilistic Services on a Ride-hailing Platform

Ride-hailing platform,congestion effect,probabilistic selling,probabilistic pricing

Motivated by the "simultaneous call" service option Didi offers, we study probabilistic services on a ride-hailing platform offering high- and low-type services. Specifically, we consider two types of probabilistic services: Probabilistic Pricing (PP) and Probabilistic Selling (PS). Riders choosing either service option do not know ex-ante which service they will ultimately receive. Under PP, riders receive a particular type of service and then pay accordingly. In contrast, under PS, riders pay the same amount in advance irrespective of their received service type ex-post, as in PS in the retail industry. Using a rational expectation framework, we characterize the optimal probabilistic service strategies. Our results show that the introduction of probabilistic service allows the platform to segment riders more effectively and better balance the supply and demand, mitigating the congestion level in the market. Interestingly, more demand from the low-type riders can be fulfilled, albeit at a higher price for the low-type service. {Moreover, whether to introduce PP or PS depends on riders' aversion to congestion and the platform's cost of offering the probabilistic service. Remarkably, providing probabilistic service can be a "win-win-win" policy for riders, drivers, and the platform when riders are not highly averse to congestion and the computational cost of the probabilistic services is not too high. Our findings provide practitioners with valuable guidelines for designing probabilistic services in ride-hailing markets.