Performance-driven vulnerability analysis of infrastructure systems

In this research we suggest a novel approach that reduces the complexity of vulnerability analysis of infrastructure systems, by addressing the vital information necessary for decision makers and avoiding an exhaustive analysis of data that is ultimately irrelevant. This objective is achieved by framing the vulnerability analysis as part of a decision-making process, thus focusing on clear and quantifiable requirements of specific essential end-users, and assessing under which particular circumstances these requirements are no longer satisfied by the infrastructure systems that support the end users' activities. Methodically, this is made possible by reversing the process through which the vulnerability of infrastructure systems is typically analysed and identifying specific minimum threshold states in which critical supplies will be interrupted to reveal the critical components in supply chains. This reduction in the complexity of vulnerability analysis also allows the consideration of highly relevant information that may currently be ignored to avoid an excessive computational burden. To test its feasibility, the approach was implemented in a synthetic, though realistic case study of an analysis of the vulnerability of a healthcare system to seismic hazards. This enabled the identification of several potential threats to the healthcare system's continued performance following future earthquakes.

Automated summary

This research introduces a new approach to reduce the complexity of vulnerability analysis in infrastructure systems. By framing vulnerability analysis as part of a decision-making process, the focus is shifted to clear and quantifiable requirements of specific end-users. The approach identifies critical components in supply chains by reversing the analysis process and reveals minimum threshold states in which critical supplies will be interrupted. The method also allows the consideration of highly relevant information currently ignored to avoid excessive computational burden.