Using metaheuristic algorithms to optimize a mixed model-based ground-motion prediction model and associated variance components


In this study, two metaheuristic optimization algorithms are employed to estimate a mixed model-based ground-motion model (GMM) with several variance components. Two optimization algorithms, particle swarm optimization (PSO) and teaching–learning-based optimization (TLBO), are employed to compute regression coefficients and uncertainties of the GMM by considering a one-stage maximum likelihood estimation framework. These optimization models are applied to a complex predictive equation in a way that the results best fit a ground-motion dataset. Uncertainties and the regression coefficients of a functional form of a predictive equation are estimated and compared with other models to show the strengths and the limitations of the proposed approaches. The obtained results provide reliable solutions and demonstrate good accuracy compared to previous search algorithms based on a given dataset of ground-motions. We estimated error metrics for the predicted data, and the results show that using the proposed algorithms provides better results compared to the existing search algorithms.

Publication Title

Journal of Seismology