Optimization of Operations in Public Transportation

Railway planning is a long and complex process that is broken down and solved through multiple planning phases that are interconnected. Most of these phases involve optimizations problems, problems such as line planning, timetabling or rolling stock scheduling. All these problems have been subject of studies in the scientic literature and several mathematical formulation have been proposed, as well as several solution methods (from exacts to heuristics).

Railway operators have started including decision support tools in their planning processes but still, in many cases these problems are tackled and solved with manual methodologies. In Denmark, railway transportation, during the past years has seen a growth in terms of passengers demand and it is expected that this growth will continue, especially considering that with Danish Green Transport Agreement signed in 2009 it is required by the State that more of the transport demand has to be served by public transportation.

In this context the goal for all the railway stakeholders (in particular railway operators and infrastructure managers) is to provide a reliable and ecient service taking into account an increased passengers demand. To achieve this goal more sophisticated decision support tools are needed; tools that can be based on operations research models and methods since these methodologies allow to have a more ecient use of the given resources and to balance operational costs, passenger service quality and robustness of the proposed plans.

In this thesis dierent railway problems are studied, in particular timetabling and rolling stock scheduling problems. The major contributions are summarized in the introductory chapter and the single projects developed are all presented in the separated chapters; each of the chapter is a paper that has been submitted, or is about to be submitted, to an international journal within the operations research eld.

The problems studied are all optimization problems; in the various chapters for each of them new mathematical formulations and new solution methods, exact and/or heuristic, are proposed. For all of the projects the solution methods have been tested on real life data and, when not available, on realistic data provided by DSB, the main Danish railway operator, which has been an active partner of this PhD project.