A dynamic ridesharing system is a platform that connects drivers who use their personal vehicles to travel with riders who are in need of transportation, on a short notice. Since each driver/rider may have several potential matches, to achieve a high performance level, the ridesharing operator needs to make the matching decisions based on a global view of the system that includes all active riders and drivers. Consequently, the ride-matching problem that needs to be solved can become computationally expensive, especially when the system is operating over a large region, or when it faces high demand levels during certain hours of the day. This paper develops a graph partitioning methodology based on the bipartite graph that arises in the one-to-one ride-matching problem. The proposed method decomposes the original graph into multiple sub-graphs with the goal of reducing the overall computational complexity of the problem as well as providing high quality solutions. We further show that this methodology can be extended to more complex ride-matching problems in a dynamic ride-sharing system. Using numerical experiments, we showcase the advantages of the new partitioning method for different forms of ride-matching problems. Moreover, a sensitivity analysis is conducted to show the impact of different parameters on the quality of our solution.

动态乘车共享系统是一个平台，可在短时间内将使用私人车辆旅行的驾驶员与需要运输的驾驶员连接起来。由于每个驾驶员/驾驶员可能都具有多个潜在的匹配项，因此要达到较高的性能水平，共享乘车运营商需要基于系统的全局视图做出匹配决策，该全局视图包括所有活动的驾驶员和驾驶员。因此，需要解决的行驶匹配问题可能变得在计算上变得昂贵，尤其是当系统在较大区域内运行时，或者在一天中的某些小时内面临高需求水平时。本文开发了一种基于二分图的图划分方法，这种二分图是在一对一的乘车匹配问题中出现的。所提出的方法将原始图分解为多个子图，目的是降低问题的整体计算复杂度并提供高质量的解决方案。我们进一步表明，该方法可以扩展到动态乘车共享系统中更复杂的乘车匹配问题。通过数值实验，我们展示了针对不同形式的拼车问题的新分区方法的优势。此外，进行了敏感性分析，以显示不同参数对我们解决方案质量的影响。我们展示了针对不同形式的拼车问题的新分区方法的优势。此外，进行了敏感性分析以显示不同参数对我们解决方案质量的影响。我们展示了针对不同形式的拼车问题的新分区方法的优势。此外，进行了敏感性分析，以显示不同参数对我们解决方案质量的影响。