The problem of dispatching shuttles to serve trip requests can be mathematically formulated as a dial-a-ride problem (DARP). With on-demand mobility services gaining more popularity due to the recent developments in the gig economy, communication technologies, and urbanization, the real-time application of DARP is attracting ever more interest. However, the fact that the size of DARP grows exponentially with number of requests and number of available seats renders the current solution methodologies inadequate for online applications. In order to tackle this issue, we propose a general framework that shifts much of the computational burden of the optimization problems that need to be solved into an offline phase, thereby addressing on-demand requests with fast and high-quality solutions in real time. Using numerical experiments, we demonstrate the benefits of the proposed method. Furthermore, we conduct sensitivity analyses to show the performance of our methodology under different parameter settings.

调度班车来满足行程请求的问题可以在数学上表述为拨号问题 (DARP)。由于零工经济、通信技术和城市化的最新发展，按需移动服务越来越受欢迎，DARP 的实时应用越来越引起人们的兴趣。然而，DARP 的规模随着请求数量和可用席位数量呈指数增长这一事实使得当前的解决方案方法不适合在线应用程序。为了解决这个问题，我们提出了一个通用框架，将需要解决的优化问题的大部分计算负担转移到离线阶段，从而以快速和高质量的解决方案实时解决按需请求。使用数值实验，我们展示了所提出方法的好处。此外，我们进行敏感性分析以显示我们的方法在不同参数设置下的性能。