With the expansion of the sharing economy, growing urban traffic, and increasing environmental pollution, bike-sharing systems (BSSs) are developing rapidly all over the world. A major operational issue in BSS is to reposition the bikes over time such that enough bikes and open parking slots are available to users. Especially during peak hours, it is essential to stabilize BSS in use. To cope with the issue, this article proposes a new approach integrating multiobjective optimization and a weighting factor based on the shortage event types of each station. In addition, the multiobjective artificial bee colony algorithm is modified according to the features of this work to find optimal solutions. The proposed approach is applied to the real-life repositioning of a BSS during peak hours to verify its feasibility and effectiveness. Also, the algorithm is compared with other frequently used multiobjective algorithms. For the comparative study, convergence metric and spacing are adopted to further measure the algorithm performance. The scalability of the proposed approach in addressing the multiobjective repositioning problems during peak hours is also verified by multiple trials. Note to Practitioners—This work deals with bike repositioning in bike-sharing systems (BSSs) during peak hours, which has major significance in the efficient operation of such systems. It builds a multiobjective optimization model and solves it through a modified multiobjective artificial bee colony algorithm. The existing single-objective optimization methods fail to solve the concerned problem. This work can find the optimal routes of the repositioning vehicles along with the number of desired parked bikes of corresponding stations. The experimental results indicate that the proposed method is highly effective and can greatly and readily help decision-makers better manage the BSS of a practical size.

中文翻译：

---

通过改进的人工蜂群算法在自行车共享系统中进行多目标自行车重新定位


随着共享经济的扩张、城市交通的增长和环境污染的加剧，自行车共享系统（BSS）在世界范围内迅速发展。 BSS 的一个主要运营问题是随着时间的推移重新定位自行车，以便为用户提供足够的自行车和空位停车位。特别是在高峰时段，稳定BSS的使用至关重要。为了解决这个问题，本文提出了一种结合多目标优化和基于每个站点的短缺事件类型的权重因子的新方法。此外，根据本工作的特点对多目标人工蜂群算法进行修改，以寻找最优解。将所提出的方法应用于现实生活中高峰时段 BSS 的重新定位，以验证其可行性和有效性。此外，该算法还与其他常用的多目标算法进行了比较。为了进行比较研究，采用收敛度量和间距来进一步衡量算法性能。多次试验也验证了所提出的方法在解决高峰时段多目标重新定位问题方面的可扩展性。从业者须知——这项工作涉及高峰时段自行车共享系统（BSS）中的自行车重新定位，这对于此类系统的高效运行具有重要意义。建立多目标优化模型，并通过改进的多目标人工蜂群算法进行求解。现有的单目标优化方法无法解决这一问题。这项工作可以找到重新定位车辆的最佳路线以及相应站点所需停放的自行车数量。 实验结果表明，该方法非常有效，可以极大方便地帮助决策者更好地管理实际规模的BSS。