A Study on Self-Rebalancing in Inter-Region Vehicle Distribution for Ride-Hailing Services

This study investigates the dynamic characteristics of self-rebalancing, where vehicle drivers independently determine their repositioning based on perceived conditions. Unlike conventional rebalancing methods controlled by operators, self-rebalancing reflects decentralized decision-making. A utility-based approach incorporating three key factors was employed: attraction, cost, and competition. A simulation-based analysis was conducted to examine the impact of different combinations of these parameters on service performance under two distinct demand scenarios. The first scenario represents a system with a large demand size and highly im-balanced patterns characterized by main flows, while the second scenario features a smaller demand size with randomly generated requests. A parameter study demonstrated that the influence of each parameter differs depending on the size and distribution characteristics of the demand. In the large-demand scenario, low weights of attraction and high weight of cost maximized the service success rate, while high competition weights reduced idle time ratios. Conversely, in the small-demand scenario, higher weights of attraction and competition increased rebalancing frequency to meet distributed demand more effectively. These findings highlight the need for scenario-specific optimization of utility parameters to achieve balanced service performance and operational efficiency. Rebalancing strategy of vehicle drivers should be optimized according to the unique characteristics of the demand patterns. This study provides a foundational framework for adaptive rebalancing strategies and suggests directions for future research, including real-world validation and addressing uncertainties in demand prediction.