Traffic controllers regulate railway traffic by sequencing train movements and setting routes with the aim of ensuring smooth train behaviour and limiting, as much as possible, train delays. In this paper, we describe the implementation of a real-time traffic management system, called ROMA (Railway traffic Optimization by Means of Alternative graphs), to support controllers in the everyday task of managing disturbances. We make use of a branch-and-bound algorithm for sequencing train movements, while a local search algorithm is developed for rerouting optimization purposes. The compound problem of routing and sequencing trains is approached iteratively, computing an optimal train sequencing for given train routes and then improving this solution by locally rerouting some trains. An extensive computational study is carried out, based on a dispatching area of the Dutch railway network. We study practical size instances, and include in the model important operational constraints, including rolling stock and passenger connections. Different types of disturbances are analysed, including train delays and blocked tracks. Comparison with common dispatching practice shows the high potential of the system as an effective support tool to improve punctuality.

During rail operations, unforeseen events may cause timetable perturbations, which ask for the capability of traffic management systems to reschedule trains and to restore the timetable feasibility. Based on an accurate monitoring of train positions and speeds, potential conflicting routes can be predicted in advance and resolved in real time. The adjusted targets (location-time-speed) would be then communicated to the relevant trains by which drivers should be able to anticipate the changed traffic circumstances and adjust the train's speed accordingly. We adopt a detailed alternative graph model for the train dispatching problem. Conflicts between different trains are effectively detected and solved. Adopting the blocking time model, we ascertain whether a safe distance headway between trains is respected, and we also consider speed coordination issues among consecutive trains. An iterative rescheduling procedure provides an acceptable speed profile for each train over the intended time horizon. After a finite number of iterations, the final solution is a conflict-free schedule that respects the signaling and safety constraints. A computational study based on a hourly cyclical timetable of the Schiphol railway network has been carried out. Our automated dispatching system provides better solutions in terms of delay minimization when compared to dispatching rules that can be adopted by a human traffic controller