| Summary: | Incidents in the road transport system can have large negative consequences for the society and the business community. The basic aim of vulnerability analysis is to identify scenarios that i) would lead to severe consequences, and ii) have some likelihood of being realized in the future. The Thesis proposes two main approaches to vulnerability analysis. The first significant component of the analysis is to identify important links in the road network, i.e., links where a disruption would lead to severe consequences. The second component is to identify exposed users, i.e., users for which the consequences of a disruption would be particularly severe. Paper I introduces the concepts of importance and exposure and how they can be operationalized in terms of increased travel time when road links are closed. The measures are applied to the road network of northern Sweden. Among other things, we find that the most important road links from a socio-economic efficiency perspective are sections of the main roads in the region going through the main population centres. The most exposed users, on the other hand, live in the sparsely populated municipalities in the northwest along the Norwegian border. Paper II studies the geographic patterns of exposure and importance in Sweden and identifies properties of the geography, road network and travel patterns that to a large extent explain the observed spatial differences. We find that the municipalities around Stockholm have the most important road networks, and that people in the southern parts of Sweden are considerably less exposed than in the northern parts. We also find that the sparsity of the road network, the travel times of the users and the traffic load on the links provide good explanatory variables for the regional variations in exposure and importance. Paper III proposes a link importance measure that incorporates both efficiency considerations, i.e. the total increase in travel time, and equity considerations, i.e. the unevenness of the distribution among users. ...
|
|---|