| dc.description.abstract | Mitigation planning in many disaster-prone areas has shown success in helping the community to withstand hazardous events, reducing the recovery time and costs, and preventing life losses. This research proposes a multi-objective optimization framework to enhance decision-making to mitigate risk from potential hazards in an integrated and quantitative manner.
First, this study introduces an optimization framework that can integrate different dimensions of community resilience in one model as competing objectives to measure the potential impacts and damage from hazard events. To the best of our knowledge, this framework is the only framework that can provide flexibility on some major components. The decision makers can apply the proposed framework to various hazards without changing the mathematical formulation. The framework's objectives can be determined by the people who are involved in decision-making. Moreover, the number of objectives also can vary according to the actual needs of decision makers.
Second, the proposed framework is applied to tornado mitigation in the city of Joplin, Missouri, USA, to demonstrate how the retrofitting strategies reduce the potential impacts of direct economic loss (economic dimension), population dislocation (social dimension), and building functionality (physical infrastructure). The results analyses illustrate how the decision makers can utilize the information from the optimal solutions to determine the appropriate retrofitting solution for the community.
Finally, a machine learning (ML) model is developed to predict potential economic damage on domestic supply, employment, migration, and household income by using input data of the computable general equilibrium (CGE) model. This ML model can act as a surrogate model to help the non-CGE expert to interpret the relationship between the capital shock by sector and economic impact from hazards shock on capitals. The predicted impact on domestical supply, employment, migration, and household income from this ML model can act as coefficients of objectives functions (domestical supply, employment, migration, and household income) of the proposed multi-objective optimization model. | en_US |
