Joint Failure Vulnerabilityof Transportation Infrastructure

Joint Failure Vulnerabilityof Transportation Infrastructure


A Erath, K Axhausen, ETH Zurich, CH


This paper proposes an approach to assess joint failure vulnerability of transportation infrastructure for a national transport network. Applied for the Swiss case, several methods to detect the most relevant joint failures are presented and compared


Infrastructure disruption caused by natural hazards is a reality for transportation networks and might become even more important with an increase of such incidents due to climate change. However, only little research has been carried out so far to integrate vulnerability into the planning process and the management of transport infrastructures. After we presented in earlier papers how vulnerability due to single link failures can be quantitatively integrated in the management of large road infrastructure networks, this paper focuses on the development of methodologies to assess joint failure vulnerability.

For intensely used road networks in densely populated countries, like Switzerland, a reliable assessment of the additional user costs has to be based on traffic assignments which consider congestion effects. Depending on the network topology, the demand distribution and the availability of alternative destinations, also mode and destination choice shifts need to be considered when computing link failure consequences. Since the failure of every link represents a single scenario, the network-wide assessment of single link failures is computationally burdensome, despite different, successfully implemented speed-up strategies such as the usage of subnetworks. However, for single link failures the number of scenarios is restricted to the number of links in the network and the network-wide assessment remains feasible. Assuming independence of link failures, the number of scenarios to be considered scales exponentially with the considered number of jointly failed links within one scenario. Given typical planning network sizes of several thousand nodes and links the full enumeration of possible scenarios becomes infeasible.

Regarding the use of failure assessment in infrastructure management systems (IMS) which optimize the allocation of resources for maintenance and natural hazard protection and mitigation measures, two problem approaches seem feasible:
First, the assumption of independence of link failure can be relaxed since the threat of natural hazards usually is locally restricted. Depending on the considered natural hazard, the interdependence of natural hazard threats to infrastructure could be evaluated based on natural hazard maps. However, natural hazard maps are not available throughout all relevant regions and do usually not consider joint risk probabilities.
Second, optimization algorithms can be employed to detect the combination of link failures that lead to the highest failure induced cost whereby the considered hazard area and the number of failed links can be restricted according to the type of natural hazard.

In this paper we focus on the later approach and present different optimization heuristics to detect those combinations of link failures which lead to the highest failure induced cost. In addition, the effectiveness of the considered heuristics with respect to their ability to detect the effectively worst case scenario will be assessed: for several different smaller networks of different network topology and demand characteristics (i.e. subnetworks that are cut out of the Swiss planning network) full enumerations of joint link failure (with restrictions of the number of jointly failed links) are computed. The results of these full enumerations are then used as benchmarks against the results of different optimizations heuristics are evaluated. Thereby, special attention is paid on whether and how network topology and demand characteristics might influence the effectiveness of the tested optimization heuristics. The derived findings will then determine the heuristics used for the analysis of the entire Swiss planning network which is also covered by the paper.


Association for European Transport