关键词: 可靠度, 高斯过程, 响应面法, 机器学习

Abstract:

Gaussian Process (GP) is a machine learning method based on strict statistical learning theory, which is highly precise in solving the highly nonlinear problem with small samples and high dimensions. Aiming to the problem of low precision in using traditional response surface method for structural reliability analysis with high nonlinear implicit performance function, Gaussian process regression model is used to reconstruct the implicit performance function, and combined with the improved traditional response surface method, a new structural reliability analysis by Gaussian process response surface method is proposed. The research results show that the presented method has advantages of high accuracy and high efficiency in solving the reliability of implicit performance function.

Key words: reliability, Gaussian process, response surface method, machine learning