Optimal probabilistic placement of facilities using a surrogate model for 3D tsunami simulations

Abstract

Abstract. This study proposes a framework for utilizing results obtained from advanced numerical simulations and performing probabilistic tsunami hazard assessment for investigating optimal facility placement. A set of numerical simulations of the tsunami off the Pacific coast caused by the 2011 Tohoku earthquake are performed considering uncertainties in fault parameters. Both inundation depths and tsunami forces acting on buildings are numerically simulated and defined as tsunami hazard indices. Proper orthogonal decomposition is then applied to numerical results for extracting characteristic spatial modes, which can be used to construct surrogate models. Monte Carlo simulations (MCSs) were performed at a low computational cost using surrogate models. The optimal placement of facilities was probabilistically investigated with the help of genetic algorithms using the MCS results along with the concept of system failure probability. The results indicate that the proposed framework allows determining the optimal placement of facilities by applying different strategies at low computational costs while effectively reflecting the results of advanced tsunami simulations.