A heuristic approach for bike-sharing system rebalancing problem with multiple time windows

In urban environments, the bicycle-sharing system (BSS) has emerged as a vital complement to public transportation, effectively bridging gaps in urban mobility. The successful operation of these systems is contingent upon ensuring the timely availability of bicycles and docking spaces, a challenge particularly pronounced during peak hours. Addressing this, our study introduces a novel approach for optimizing the rebalancing of BSS. We have developed a time-windowed rebalancing request generator, which, in conjunction with a route planning algorithm based on variable neighborhood search (VNS), provides a robust solution to this issue. This iterative evolutionary algorithm not only refines solutions progressively but also incorporates a shaking process to enhance the possibility of achieving global optimization. Our methodology further includes an algorithm for generating initial solutions, alongside two shaking processes designed to improve efficiency. We validate our approach using real-world data from Taipei’s Youbike system, demonstrating superior performance in rebalancing efficiency compared to existing algorithms. Our study contributes to the optimization of BSS rebalancing, offering practical insights for urban transportation planning.