AI AND MACHINE LEARNING IN SEISMIC DESIGN OF BUILDINGS

Abstract

This article presents an analytical review of the application of artificial intelligence (AI), machine learning (ML), and deep learning (DL) in seismic design, assessment, and optimization of buildings. Five interrelated areas are considered: machine learning for seismic response prediction, deep learning of nonlinear time histories, damage and seismic risk assessment, the direct application of ML in design and multi-criteria optimization, and physically enhanced ML approaches. It is shown that research is evolving from individual regression and classification models to surrogate models capable of reproducing the complex dynamic response of a structure and participating in an iterative search for design solutions. Of particular importance is the ability of ML models to significantly reduce the number of costly nonlinear calculations in multivariate design.