Graduate researchers: Camilo Gomez, Nirmal Jayaram, Mahalia Miller, Jason Wu
Project sponsor: National Science Foundation, Stanford UPS Foundation, Google
Thanks to Caltrans, the Metropolitan Transportation Commission and the San Francisco Public Utilities Commission for providing data
Figure adapted from Miller and Baker, 2014
This work aims to combine new probabilistic and multiscale modeling approaches to provide insights into the performance of spatially distributed infrastructure subject to component disruptions from natural disasters or other sources. Current infrastructure risk assessment approaches are often limited by simplified treatment of the complex network effects or of the variations in disruption over a large spatial region. A common way to overcome these limitations is to estimate disruption given a single scenario disaster, but the probability of occurrence of the scenario is rarely incorporated, hindering risk management based on cost-benefit analysis. These limitations are caused by the impossibility of computing disruptions to complex networks under the huge number of possible disaster scenarios that might occur. This effort aims to overcome current challenges through insights in spatial correlation of disruptions, multiscale infrastructure systems modeling, and advanced reliability algorithms.