European air and rail transport networks between main cities : a comparative analysis

Details

Supervisors

Delvenne, Jean-Charles

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] : ingénieur civil en mathématiques appliquées

Abstract

International transport networks are at the core of our economies and societies as they shape the way we travel and the way we trade. At European level, the global air and rail transport networks allow passengers to travel between most of the cities in a short time. Low-cost airlines offer affordable flights between most of the large European cities and the high-speed rail (HSR) infrastructure is continuously expanding. The last decade has however seen a growing concern in the way we travel and the impact it has on the environment, as flying remains one of the most carbon-intensive activity one can do on an hourly basis. This phenomenon has been reflected by policy initiatives proposing to reduce the number of domestic flights, by the Swedish movement of flygskam (shame of flying) or by the recommissioning of some night train routes. In this thesis, we propose to study the European Air Transport Network (EATN) and the European Rail Transport Network (ERTN) using complex networks theory. Transport systems can indeed be modelled as networks with nodes (airports, train stations) and edges linking them (travel routes). The particular feature of this thesis is to study both networks in parallel with essentially the same nodes representing main cities in Europe. Using measures of complex networks theory, we show that the EATN has the properties of small-world networks, with the average number of flights required to travel between any two cities being as low as 1.81. On the other hand, the ERTN is characterized by much longer paths and travelling between two cities requires taking 5.77 trains on average. Using directed graphs, we model door-to-door journeys between cities taking into account transfer times to the airport/station, checkin-in and connecting times. We find that for 6% of the journeys between main cities in the EU, the door-to-door travel time is shorter by train than by plane, and that in June 2017 at least 4 million passengers took a flight on a journey where the train is faster. In addition, we evaluate that the train is usually competitive with the plane up to a distance of 444 km, and up to 704 km when the journey can be made on high-speed trains. We then assess the impact on train lines following a potential shift of those 4 million passengers from air to rail. Our analysis shows that high-speed rail will have to carry the biggest share of additional passengers resulting from a modal shift, and that the main hubs are not always the same in the air and rail networks. Our network model is then used for two case studies. Firstly, we evaluate the feasibility of night train services between main cities in Europe. We find that night train routes could in theory be operated between 22% of the cities, and up to 38% if we implement high-speed night trains as in China. Secondly, we assess the potential in terms of passenger traffic on the future high-speed line between Lyon and Turin and find that the market share that could be gained from the air transport sector is rather modest. Finally, we consider the problem of the optimal network design for the HSR network in Western Europe using methods developed by Micheal Gastner and Mark Newman. Several optimal network candidates are generated with some of them exhibiting up to 75% similarity with the actual network.