Adaptive Metamodels for Structural Reliability

This book focuses on adaptive metamodel-based methods that dramatically reduce computational costs in structural reliability analysis and reliability-based design optimization (RBDO). The metamodel-based methods significantly reduces the computational burden of reliability analysis and design, and has therefore received widespread attention and application in the reliability community. As for structural reliability, metamodels are mainly used to construct low-cost approximations of performance functions to support subsequent analysis and design. The prerequisite for this method is that the original performance function is a computationally expensive black-box model. Therefore, by establishing high-precision and low-cost approximations, the computational cost can be effectively reduced. This book mainly introduces the basic principles of metamodel-based methods, and demonstrates the flexibility and efficiency of metamodel-based methods in reliability engineering, as well as their potential in practical applications. This book is aimed at graduate students, scholars, and reliability engineers interested in structural/mechanism reliability analysis and design optimization. MATLAB/Python pseudo-codes for key algorithms are available online.