From probabilistic risk analysis to resilience with network science: lessons from the literature and best practice

Natural, engineered and social systems are growing ever more complex and interconnected. Rapid and often seemingly erratic change, including that caused by humans, is becoming the new normal. Conventional risk analysis methods, such as probabilistic risk analysis (PRA), focus on the identification of the vulnerabilities of specific components to an expected adverse event and quantify the loss in functionality of the system because of these events. However, strong coupling that exists between social, technical and economic systems make this form of protection unrealistic and cost-prohibitive. Since interconnectedness is ubiquitous to these systems, a new paradigm of resilience enabled by network science has been proposed. In addition to acting as connective tissues across built, natural and social systems, network science tools provide us with a way to understand robustness and recovery of these systems in a hazard-agnostic way. However, translating this paradigm to address real-world problems is still a momentous challenge for decision makers, engineers and social scientists. Here we explore the applicability of the new paradigm to real life systems, such as Indian Railways and the US National Airspace, and compare its benefits and limitations with the conventional PRA framework. Insights thus obtained can inform how existing gaps may be bridged to enable integration of resilience in system design and regulatory structures.