Hierarchical Vehicle Scheduling Research on Tide Bicycle-Sharing Traffic of Autonomous Transportation Systems

Abstract

To facilitate the intelligent and automated provision of mobility services by autonomous transportation systems, bike-sharing can be a supplement to public transport for addressing their point-to-point issue, namely, “last mile” service. However, according to the different nature of land use, the uneven spatio-temporal distribution of travel demand can directly lead to difficult access to bikes with high travel costs for users and operating costs for operators. Based on this, this paper analyzes the user behavior patterns within different areas by using GeoHash encoding and proposes a hierarchical autonomous vehicle scheduling model based on tide bicycle-sharing traffic, namely, HATB. It minimizes operating costs and maximizes user satisfaction to dynamically optimize scheduling routes and required vehicles within each layered zone. As for extracting historical orders of Mobike in Beijing, an example analysis through the genetic algorithm of HATB is conducted to support effective and efficient scheduling. Compared to existing scheduling methods, HATB has faster convergence and lower time complexity, which improves bike turnaround efficiency and practical application ability, thus making it more convenient for the public to travel during peak hours.

Public Bicycle Dispatch Method Based on Spatiotemporal Characteristics of Borrowing and Returning Demands

“…From a spatial perspective, a hierarchical public bicycle dispatch method based on 'station-cluster' structures has been constructed by scholars [15][16][17]. A public bicycle rebalancing framework considering both dynamic rebalancing within each station and static rebalancing between stations was devised by Z. Tian [9].…”

“…In practical applications, it is essential to maintain the ratio of bicycle docks to available bicycles at stations within the range of [E min , E max ] after redeployment. The redeployment demand for station i should satisfy Equation ( 16), and Equation ( 16) can be further expressed as Equation (17).…”

“…The range of values for station redistribution can be determined by using Equation (17). When determining the specific amount of redistribution, stations with inbound redistribution demands are prioritized, followed by determining the outbound redistribution amount for stations based on proximity.…”

“…Some scholars have considered the issue of potentially damaged bicycles within public bicycle systems and have conducted research on bicycle redistribution methods under such circumstances [14]. To alleviate the complexity of solving the problem, scholars have approached the issue from a spatial perspective, constructing hierarchical public bicycle redistribution methods based on 'station-cluster' structures [15][16][17]. For instance, R. Hu developed a dynamic optimization and rebalancing model for bike-sharing systems based on demand prediction, aiming to minimize operational costs while maximizing user satisfaction [18].…”