Recovery optimization of interdependent infrastructure: A multi-scale approach

Rapid post-disaster recovery of infrastructure is necessary for prompt societal recovery. Regional resilience analysis can promote mitigation and recovery strategies that reduce the spatial extent and duration of infrastructure disruptions. Three significant challenges in regional resilience analysis are 1) modeling the physical recovery of infrastructure; 2) modeling the associated service recovery; and 3) developing a computationally manageable approach for the recovery modeling and optimization. This paper presents a novel multi-scale approach for the post-disaster recovery modeling and optimization of interdependent infrastructure. The multi-scale approach facilitates the recovery modeling and enables developing recovery strategies that are feasible to implement and easy to communicate. To enhance regional resilience, the paper integrates the recovery modeling into a multi-objective optimization problem. The optimization problem aims to schedule the required recovery activities such that disrupted services are restored as fast as possible, while minimizing the incurred cost. In the optimization problem, resilience metrics are introduced to monitor and quantify service recovery. The optimization problem is subject to recovery scheduling and network flow constraints, where each is formulated as a nested optimization. The multi-scale approach to the recovery optimization highlights the role of infrastructure at multiple scales to achieve selected recovery objective(s). As an illustration, the proposed approach is used to optimize the post-disaster recovery of interdependent infrastructure in a virtual community testbed.