ABSTRACT

A large number of natural disasters and the abundant literature dealing with them have shown that in the post-disaster recovery phase, the dependencies and interdependencies among infrastructure sectors and their components play a critical role. For this reason, an accurate assessment of structural and infrastructural resilience should account for such dependencies. Yet, accurate models to capture these dependencies and predict their effect are still in their infancy. This paper presents substantial advancements in modeling three crucial types of dependencies that may severely hinder the post-disaster recovery: resources sharing among multiple recovery tasks, precedence constraints among repair actions, and functionality prerequisites across multiple components. A coherent approach is proposed to handle the effect of all these dependencies, while also accounting for the substantial uncertainties involved, such as repair duration, resource demand, and unexpected availability of temporary resources that can relax some of the dependencies. As part of the PRAISys project, the approach is demonstrated for the case of a transportation network recovering from an earthquake that has damaged several of its bridges. The result is a scenario-based simulator that can predict (in a probabilistic sense) the post-disaster restoration phase and the system functionality recovery. This tool can inform and enhance pre-event decisions, allowing investments into very different options to be compared. Several tools have been developed to answer the question: “Which bridge should be retrofitted? The proposed approach allows one to answer more general questions for informed decision-making. For instance: “Will my network be more resilient if I allocate part of my retrofit budget to the purchase of portable generators that will systematically allow my repair crews to relax some dependencies?” or “How many mutual assistance agreements for road clearing should I establish, to be reasonably sure to have enough crews under a certain seismic scenario?”