Online Relocating and Matching of Ride-Hailing Services: A Model-Based Modular Approach

Chang Gao,Xi Lin,Fang He,Xindi Tang

2023-10-13

Abstract:This study proposes an innovative model-based modular approach (MMA) to dynamically optimize order matching and vehicle relocation in a ride-hailing platform. MMA utilizes a two-layer and modular modeling structure. The upper layer determines the spatial transfer patterns of vehicle flow within the system to maximize the total revenue of the current and future stages. With the guidance provided by the upper layer, the lower layer performs rapid vehicle-to-order matching and vehicle relocation. MMA is interpretable, and equipped with the customized and polynomial-time algorithm, which, as an online order-matching and vehicle-relocation algorithm, can scale past thousands of vehicles. We theoretically prove that the proposed algorithm can achieve the global optimum in stylized networks, while the numerical experiments based on both the toy network and realistic dataset demonstrate that MMA is capable of achieving superior systematic performance compared to batch matching and reinforcement-learning based methods. Moreover, its modular and lightweight modeling structure further enables it to achieve a high level of robustness against demand variation while maintaining a relatively low computational cost.

Machine Learning,Systems and Control

What problem does this paper attempt to address?

The paper attempts to address the problem of dynamically optimizing order matching and vehicle dispatching in ride-hailing platforms. Specifically, the paper focuses on ride-hailing platforms operating under a dispatching model, where these platforms use online algorithms to achieve real-time connections between passengers and drivers, reducing information asymmetry and improving service efficiency. However, in actual operations, ride-hailing demand often exhibits significant spatiotemporal heterogeneity, leading to imbalances in supply and demand within the network, resulting in adverse consequences such as increased passenger waiting times, order losses, and increased idle time for drivers.

To this end, the paper proposes a Model-based Modular Approach (MMA) aimed at addressing the aforementioned issues by optimizing order matching and vehicle dispatching strategies. MMA adopts a two-layer modular modeling structure: the upper layer determines the spatial transfer patterns of vehicle flows within the system to maximize the total revenue of the current and future stages; the lower layer, guided by the upper layer, quickly performs vehicle-to-order matching and vehicle dispatching. MMA is interpretable and equipped with customized polynomial-time algorithms capable of handling large-scale online matching and dispatching problems. The paper also theoretically proves and empirically validates the effectiveness and superiority of MMA through numerical experiments.