In the Netherlands, it is expected that passenger activities on railway networks will double by 2050. To manage the passenger demand, railway capacity planning needs to be adapted. One fundamental aspect of the railway capacity planning is the scheduling of large maintenance projects. These maintenance requests should not be scheduled during major events to avoid the disruption of service. To do so, passenger operators can submit event request, i.e., a time period and location in which no maintenance project should be scheduled. Currently, these event requests are considered as hard constraints and the flexibility in maintenance scheduling decreases resulting in conflicts. In this study, the focus is on scheduling maintenance projects to minimize passenger delays while relaxing the hard constraints of event requests. This problem is addressed by introducing a Mixed Integer Linear Program (MILP) that minimizes passenger delays while scheduling maintenance projects, which includes capacity constraints for alternative services in event request areas during maintenance projects. The computational complexity of the proposed model is reduced by adding valid inequalities from the Single Machine Scheduling problem and using a simulated annealing meta-heuristic to find a favorable initial solution guess. Then, the MILP is solved to global optimality with Branch-and-Bound. A case study on the Dutch railway network shows improvements when event requests are not considered as hard constraints and an increase in the flexibility to schedule maintenance projects. This allows decision makers to choose from a set of optimal maintenance schedules with different characteristics.

中文翻译：

---

根据铁路运营商的活动要求，最大限度地减少乘客延误，改善铁路维护项目的调度

在荷兰，预计到 2050 年，铁路网络上的客运活动将增加一倍。为了管理客运需求，需要调整铁路运力规划。铁路容量规划的一个基本方面是大型维护项目的调度。这些维护请求不应安排在重大活动期间，以避免服务中断。为此，客运运营商可以提交事件请求，即不应安排维护项目的时间段和地点。目前，这些事件请求被视为硬约束，维护计划的灵活性降低，导致冲突。在这项研究中，重点是安排维护项目，以最大限度地减少乘客延误，同时放宽活动请求的严格限制。通过引入混合整数线性程序 (MILP) 可以解决此问题，该程序可以在安排维护项目时最大限度地减少乘客延误，其中包括维护项目期间事件请求区域中替代服务的容量限制。通过添加单机调度问题中的有效不等式并使用模拟退火元启发式来找到有利的初始解猜测，可以降低所提出模型的计算复杂性。然后，使用分支定界法将 MILP 求解为全局最优。对荷兰铁路网的案例研究表明，当事件请求不被视为硬性约束时，情况会有所改善，并且安排维护项目的灵活性也会增加。这使得决策者可以从一组具有不同特征的最佳维护计划中进行选择。