ON-TIME

Abstract

A wide literature is available on models and tools for the optimal real-time management of railway traffic, but the knowledge of their effects on real operations is still blurry and very limited due to the scarce implementation of these systems in practice. This paper analyses how these tools perform when interfaced in a closed-loop setup with a realistic traffic environment. A framework is developed that couples the rescheduling tool ROMA with the microscopic simulation model EGTRAIN. Railway traffic is managed for different perturbed scenarios using a rolling horizon scheme where optimal plans are periodically computed based on current traffic information and implemented in the simulation model. The closed-loop setup is investigated for different combinations of its parameters relatively to quality and stability of rescheduling plans. A comparison is performed against a typical open-loop approach that implements only the plan computed on the basis of expected train entrance delays. Both the closed-loop and the open-loop approaches are evaluated against the case in which no rescheduling is considered and trains keep on following the original timetable. Results obtained for the Dutch corridor Utrecht-Den Bosch show that the closed-loop always outperforms the open-loop in terms.