Resilience-based post-disaster response and recovery strategies for a transportation-community system

Abstract

Highway networks are critical lifelines for supporting emergency response and recovery activities in post-earthquake circumstances. Highway bridges are seismically vulnerable components of a highway network, and damaged bridges can disrupt road connections between cities, leading to severe delays in rescue operations in the emergency response phase and reconstructions of communities in the recovery phase. Therefore, efficient restoration of damaged bridges is of paramount importance to quick recovery of highway networks from earthquakes, thereby enhancing highway network resilience. Post-earthquake restoration of highway bridges consists of two phases: the emergency restoration and the long-term restoration. Emergency restoration is performed in the emergency response phase and is just sufficient to support disaster relief activities. Prior studies on scheduling emergency bridge restoration activities assumed that all information on bridge damages and the corresponding restoration methods were known before conducting emergency restoration activities. However, in practice, emergency restoration is conducted with gradually revealed bridge damage information that is collected by emergency inspection activities. Given that emergency inspection and restoration activities can be performed simultaneously on a highway network, complex interactions among these activities can occur and may significantly affect the inspection routes and restoration schedules. Specifically, emergency inspection routes may change due to the blockage of highways for the emergency restoration of bridges on them; meanwhile, emergency inspection routes could also affect emergency restoration schedules because only the bridges that have been inspected will be scheduled for emergency restoration. Meanwhile, given that bridge damage information is revealed gradually via inspection efforts, such real-time damage information may affect the following inspection routes and restoration schedules. How to account for such interactions and the real-time bridge damage information in emergency inspection routing and restoration scheduling remains a challenge. On the other hand, long-term bridge restoration is performed in the recovery phase and aims to fully restore all damaged bridges to their pre-earthquake conditions. Existing methods on long-term bridge restoration-scheduling problems used monotonically increasing functions to model the recovery processes of highway networks' functionality while neglecting the decrease of network functionality resulting from the restoration-downtime impact. The failure to take into account the impact of restoration downtime on networks' functionality may lead to the overestimation of highway network resilience, thereby resulting in inefficient bridge restoration schedules with significant restoration downtime of highways. To address these challenges, this thesis aims to facilitate the post-earthquake recovery of highway networks by developing efficient bridge restoration strategies for a post-earthquake highway network. The specific objectives in this thesis are: (1) to understand the impact of interactions among emergency bridge inspection and restoration activities on the optimal emergency inspection routes and restoration schedules; (2) to develop a real-time decision-making tool for continuously updating emergency bridge inspection routes and restoration schedules based on the real-time bridge damage information collected by bridge inspection activities; and (3) to investigate the impact of restoration downtime on highway network resilience. To achieve these objectives, mathematical optimization tools, including integer programming and decomposing techniques, are developed. Moreover, hybrid genetic algorithms are developed to efficiently solve these mathematical programs. These optimization models were tested on real highway networks in Sichuan, China, using data from the 2008 Wenchuan Earthquake. This thesis concludes three key findings. First, the inspection-restoration interactions can considerably increase the complexity of the emergency inspection routes and restoration schedules, and simultaneously performing emergency inspection and restoration activities can aid in the significant improvement in highway network resilience. Second, updating the emergency inspection routes and restoration schedules in real-time by employing the proposed real-time inspection-routing and restoration-scheduling model can ensure the effectiveness and efficiency of emergency inspection and restoration operations. Third, taking into account the restoration-downtime impact on decreasing the highway network functionality can help to establish efficient long-term bridge restoration schedules, while neglecting such an impact can lead to the overestimation of highway network resilience.