Timetable-based determination of the capacity of railroad networks under consideration of the service quality

在欧洲运营密集的铁路网中，铁路交通的战略规划正朝着时刻表驱动的规划方法发展，其中基础设施的设计符合要求并反映网络时刻表的结构。随着规划模式的转变，出现了根本性的挑战，这需要详细分析时间表和基础设施之间的相互依存关系。为此，提出了一种新的铁路网能力和基础设施规划方法，该方法反映了时刻表和运力分配的变化条件，并在网络层面上考虑了时间和空间上连贯的列车路径。该方法的核心是将数学最优化与随机(极大，+)方法相结合。虽然优化用于列车调度和列车路径分配，但(max，+)-系统允许评估与特定时刻表相关的运行质量。通过将这两个组成部分结合在一起，实现了宏观交通分配模型，该模型允许评估基础设施适应特定需求的能力，以及可以以特定服务质量运营的列车数量。分配模型是网络容量规划过程的基础。在微观基础设施的基础上，构建列车路径段。这些区段是预先安排的系统列车路径的空间构建块，是运力分配的基础。将需求结构转换为基于网络的交通流关系，并使用预先安排的系统列车路径在宏观基础设施上对其进行布线。混合整数规划用于根据预先安排的列车路径的选择来确定单独的、完整的和最优的列车路径。结果，确定了容量最优路径分配以及剩余容量。迭代地调整预先安排的系统列车路径并求解系统列车路径分配的优化模型允许同时、在网络范围内确定容量最优的列车路径。使用随机(max，+)方法将服务质量和稳健性结合在一起，该方法允许基于先前带有随机变量的确定性容量最优解中的到达、出发和列车运行时间的识别来建模延误传播和正点率。个别列车运行是按照一系列活动来描述的。列车之间的相互依赖产生了列车之间的连接，这允许递归地捕获不同列车之间的延迟传播。结果，可以在本地和网络级别上评估与特定时刻表和列车需求相关联的服务质量，从而实现用于考虑时刻表连接并与服务质量相关的能力评估的连贯模型。

Strategic planning of railway traffic in Europe’s densely operated railway network is developing towards timetable-driven planning approaches, where infrastructure is designed to meet the requirements and to reflect the structure of the network timetable. With shifting planning paradigms, fundamental challenges are arising, which require detailed analysis of the mutual dependencies of timetable and infrastructure. To this end, a new methodology for capacity and infrastructure planning in railway networks is developed that reflects the changing conditions in timetabling and capacity allocation and accounts for temporally and spatially coherent train paths on the network level. The core of the developed methodology is formed by the connection of mathematical optimization and a stochastic (max,+)-approach. While the optimization is used for train scheduling and train-path assignment, the (max,+)-system allows to assess the operational quality associated with a specific timetable. By combining these two components, a macroscopic traffic assignment model is achieved, which allows to assess the infrastructures capabilities to accommodate a specific demand, as well as the number of trains that can be operated at a specific service quality. The assignment model forms the base of the network capacity planning procedure. Based on a microscopic infrastructure, train path segments are constructed. These segments are the spatial building blocks for pre-arranged system train paths, which are the foundation of capacity allocation. The demand structure is transformed into network-based traffic flow relations which are routed on the macroscopic infrastructure using pre-arranged system train paths. A mixed-integer program is used for determining the individual, complete and optimal train paths based on a selection of pre-arranged train paths. As a result, the capacity-optimal path assignment, as well as the residual capacity is determined. Iteratively adjusting pre-arranged system train paths and solving the optimization model for system train path assignment allows for the simultaneous, network-wide determination of capacity-optimal train paths. Service quality and robustness are incorporated using a stochastic (max,+)-approach which allows to model delay propagation and punctuality based on the identification of arrival, departure and train running times in the previous deterministic capacity-optimal solution with random variables. Individual train runs are described in terms of a sequence of activities. The mutual dependencies between trains yield connections between trains, which allows to recursively capture delay propagation between different trains. As a result, the service quality associated with a specific timetable and train demand can be assessed both locally, and on the network level, such that a coherent model for capacity assessment accounting for timetable connections and relating to the service quality is achieved.