Railway traffic disturbance management

Abstract: With the increasing traffic volumes in many railway networks and reports on capacity deficiencies that result in insufficient punctuality and reliability, the need for efficient disturbance management solutions becomes evident. This thesis focuses on solutions that aim to minimise the consequences of disturbances for the various stakeholders and specifically on methods for re-scheduling the traffic. Railway traffic re-scheduling is a complex task with many influencing factors to consider and multiple stakeholders with sometimes conflicting interests. This problem is typically handled manually by traffic dispatchers that have a very limited access to support systems to facilitate their decision-making. This limitation hampers the possibilities to achieve sustainable and system-optimal decision-making and to provide the stakeholders with reliable traffic prognoses. We first study how railway traffic system users experience and are affected by the way the disturbances are communicated and handled by the traffic dispatchers. The results indicate that the disturbance-related information provided by the dispatchers is currently insufficient. The stakeholders need to acquire improved prognoses of their traffic and immediate part of the network to internally be able to minimise the negative effects of the disturbances. Furthermore, an analysis of the disturbance management problem structure and how the problem can be modelled is provided. The analysis shows that there exist fundamental restrictions in the traffic system that bounds the traffic flow but also a large number of context-dependent considerations such as sustaining certain connections or prioritising specific trains. The prevalence and feasibility of such considerations are difficult to identify and model. Moreover, the objectives of the disturbance management are vague and partly unclear, and therefore it is also difficult to measure and evaluate the outcome of the corresponding decision-making. Finally, a number of optimisation-based solution approaches with the purpose to facilitate for the dispatchers and their decision-making has been developed. The performance and applicability of the approaches have been evaluated for various disturbance settings using data for parts of the Swedish railway network that currently experience capacity deficiencies. The evaluation has identified certain disturbances characteristics that have a significant influence on the disturbance propagation, and which in some cases complicate the re-scheduling procedure. Furthermore, the significance of applying certain re-scheduling objectives and their correlation with performance measures has been analysed. The analysis shows e.g. that a minimisation of accumulated delays has a tendency to delay more trains than a minimisation of total final delay or total delay costs. An experimental study of the long-term effects when applying a limited planning perspective has also been conducted. The results indicate that solutions which are good on longer-term can be achieved despite the use of a limited planning horizon. In parallel to the optimisation-based approaches, an agent-based conceptual model with emphasis on the interplay between the different components in the railway traffic system has been proposed.