The Benefits of Willingness-to-Pay-Based Incentive-Driven Rider Repositioning in Ride-Hailing Systems

Modern ride-hailing systems are facing great challenges in improving the rider-driver matching rate due to several essential reasons: i) supply and demand are highly mismatched both temporally and spatially during peak hours; ii) order cancellations caused by hard-to-meet dispatching are difficult to avoid because of the complicated topology of the road network; iii) pick up time are often inaccurately estimated, primarily due to the uncertain nature of traffic conditions, thereby elevating the risk of losing customers. In this work, we investigate the benefits of optimal rider repositioning in ride-hailing systems, which incorporates riders' willingness in response to monetary incentives. In particular, a nonlinear mixed-integer programming model is developed to address the joint optimization of the pickup location selection, discount setting, and rider-driver matching, where a Willingness-to-Pay (WTP) model is employed to capture riders' preferences. We show the NP-hardness of the studied problem and propose an efficient exact column generation algorithm to accelerate the solution. To test the performance of our method in real-world road networks under dynamic traffic conditions, we leverage Manhattan road map and taxi trip data to predict travel times in real-time for extensive numerical experiments and sensitivity analysis. The computational results exhibit an average of 7.59% increase in matched riders and a 5.56% improvement in total revenue, and a 2.35% reduction in pickup time for experiments over 10 days (120,000 trip requests in total). By embracing our proposed method in real ride-hailing systems, the mismatch between supply and demand can be accommodated effectively. Specifically, if riders are willing to walk a short distance and sensitive to monetary incentives, a win-win-win outcome for all stakeholders (riders, drivers and platforms) can be achieved: cheaper and easier-to-access mobility service for riders, less detours and pickup times for drivers and higher system-wide profit for platforms.