Policy-based disaster recovery planning model for interdependent infrastructure systems under uncertainty

Due to continuous population expansion and the threat of climate change, the past century has witnessed increasing occurrences of natural hazards, leading to significant global losses and requiring substantial restoration efforts. This issue challenges decision makers to act in a timely and effective manner to protect infrastructure systems from future natural hazards. This study presents a policy-based decision model for restoration planning, as part of the PRAISys platform, to support informed disaster mitigation of interdependent infrastructure systems under uncertainty. Following the concept of disaster recovery priority used in practice, this model determines the priority rank of each recovery task from pre-defined policies and simulates the restoration accordingly. This model captures different types of interdependencies with rigorous models at the component and system levels and predicts possible system recoveries under a given damage scenario in a probabilistic manner. This model can quantitatively evaluate the effectiveness of decision strategies on system recovery and resilience under different disaster recovery policies. As a demonstration example, this study applies the proposed model to the post-earthquake recovery simulation of three interdependent infrastructure systems (i.e., power, communication, and transportation) in the Lehigh Valley, Pennsylvania, USA. A total of sixteen cases were considered to represent different restoration strategies. For every case, the uncertainties in the recovery steps are captured by probabilistic simulation, and system resilience is calculated for every recovery sample. Simulation results from different strategies are compared to evaluate the effectiveness of non-intuitive strategies on system recovery and resilience. The proposed model uses a simple and straightforward concept to mimic practical disaster recovery plans. It is easy to understand and implement for modelers, and it is also useful to compare outcomes from different recovery criteria and decision strategies for practitioners.