2014
Vulnerability analysis methods for road networks
BÍL, Michal, Rostislav VODÁK, Jan KUBEČEK, Tomáš REBOK, Tomáš SVOBODA et. al.Základní údaje
Originální název
Vulnerability analysis methods for road networks
Autoři
Vydání
European Geosciences Union General Assembly 2014 (EGU 2014), 2014
Další údaje
Typ výsledku
Prezentace na konferencích
Utajení
není předmětem státního či obchodního tajemství
Příznaky
Mezinárodní význam, Recenzováno
Změněno: 28. 5. 2014 10:36, RNDr. Tomáš Rebok, Ph.D.
Anotace
V originále
Road network belongs to important lifelines of the modern society. It may be harmed by random or intentional events. Roads are often affected also by natural hazards, whose impacts are both direct and indirect. Whereas the direct impacts (e.g. road damage by a landslide) are localized close to the natural hazard occurrence, the indirect impacts can embrace widespread service disabilities and considerable travel delays. The change of flows in the network may affect population living far from places originally hit by a natural disaster. These effects are possible primarily due to an intrinsic nature of this system. The consequences and extent of indirect costs also depend on what set of road links was damaged, because the road links differ in their importance.
The more robust (interconnected) the road network is, the less time is usually needed to secure the serviceability of an area hit by a disaster. Such networks also prove higher degree of resilience. The evaluation of the road network structure is therefore essential in any type of vulnerability and resilience analysis.
There are many approaches used for evaluation of vulnerability of a network and for identification of the weakest road links. Only few of them are, however, capable to simulate impacts of simultaneous closure of many links, which often happens during a disaster. The main problem is that in cases of a disaster, which usually has a large regional extent, the road network may remain disconnected. Majority of commonly used indices use direct computation of shortest paths or time between OD (origin – destination) pairs and therefore cannot be applied when the network breaks up to two or more components. However, since extensive break-ups often happen in case of major disasters (e.g., natural flooding), it is important to study the network vulnerability in these cases as well, so that appropriate actions can be applied to make it more resilient.
Performing such an analysis of network break-ups requires considering the network as a whole, identifying ideally all the cases generated by simultaneous closure of multiple links, and evaluating them using various criteria. Spatial distribution of settlements, important enterprises and overall population in nodes of a network are several factors, besides the topology of the network, which could be taken into account when computing vulnerability indices and identifying the weakest links and/or weakest link combinations.
However, even for medium-sized networks (i.e., hundreds of nodes and links), the problem of break-ups identification becomes very hard to resolve – the combinations of multiple concurrently broken links result in huge state space, which has to be examined. The naive approaches of brute force examination thus fail and more elaborated algorithms have to be applied.
In our work, we address the problem of evaluating the vulnerability of road network by simulating impacts of simultaneous closure of multiple roads/links. We present an ongoing work on a sophisticated algorithm focused on an identification of network break-ups and evaluating them by various indices. Making use of several principles from graph theory, the proposed algorithm is able to cope with medium-sized networks and identify the break-ups of requested size for further evaluation in reasonable time.
The more robust (interconnected) the road network is, the less time is usually needed to secure the serviceability of an area hit by a disaster. Such networks also prove higher degree of resilience. The evaluation of the road network structure is therefore essential in any type of vulnerability and resilience analysis.
There are many approaches used for evaluation of vulnerability of a network and for identification of the weakest road links. Only few of them are, however, capable to simulate impacts of simultaneous closure of many links, which often happens during a disaster. The main problem is that in cases of a disaster, which usually has a large regional extent, the road network may remain disconnected. Majority of commonly used indices use direct computation of shortest paths or time between OD (origin – destination) pairs and therefore cannot be applied when the network breaks up to two or more components. However, since extensive break-ups often happen in case of major disasters (e.g., natural flooding), it is important to study the network vulnerability in these cases as well, so that appropriate actions can be applied to make it more resilient.
Performing such an analysis of network break-ups requires considering the network as a whole, identifying ideally all the cases generated by simultaneous closure of multiple links, and evaluating them using various criteria. Spatial distribution of settlements, important enterprises and overall population in nodes of a network are several factors, besides the topology of the network, which could be taken into account when computing vulnerability indices and identifying the weakest links and/or weakest link combinations.
However, even for medium-sized networks (i.e., hundreds of nodes and links), the problem of break-ups identification becomes very hard to resolve – the combinations of multiple concurrently broken links result in huge state space, which has to be examined. The naive approaches of brute force examination thus fail and more elaborated algorithms have to be applied.
In our work, we address the problem of evaluating the vulnerability of road network by simulating impacts of simultaneous closure of multiple roads/links. We present an ongoing work on a sophisticated algorithm focused on an identification of network break-ups and evaluating them by various indices. Making use of several principles from graph theory, the proposed algorithm is able to cope with medium-sized networks and identify the break-ups of requested size for further evaluation in reasonable time.