Bayesian-based dynamic forecasting of infrastructure restoration progress following extreme events

Following an extreme event, efficient restoration of infrastructure functionality is a paramount task for sustaining community lifelines. With the overall goal of improving the rapidity of infra-structure restoration, the objective of this research is to forecast up-to-date infrastructure restora-tion progress considering associated uncertainties through integrating Bayesian inference and earned schedule. In this research, beta cumulative distribution function is assumed to represent infrastructure restoration progress. As infrastructure restoration progresses, Bayesian inference with Markov chain Monte Carlo is applied to update the planned restoration progress. Based on the updated progress, earned schedule and Monte Carlo simulation are specialized to forecast the future restoration progress as well as considering restoration-associated uncertainties. A case study on power infrastructure restoration during Hurricane Irma at Miami-Dade County was pre-sented. The results demonstrate the updating capability of the proposed approach and underline the importance of plan updating during infrastructure restoration. In practice, a reliable and up-to-date progress forecasting better informs practitioners to understand the latest infrastructure restoration status, thereby enhancing restoration operations to impacted communities in a timely manner.