Entropy-based adaptive design for contour finding and estimating reliability

TL;DR

This paper introduces an entropy-based adaptive Gaussian process design that enhances failure probability estimation in reliability analysis, especially near failure regions, by improving contour detection and enabling efficient batch data acquisition.

Contribution

The paper presents a novel entropy-based adaptive GP sampling method that improves failure region identification and integrates batch selection for reliability analysis.

Findings

Better identification of multiple failure regions.

Higher confidence in failure probability estimates.

Effective batch selection without loss of accuracy.

Abstract

In reliability analysis, methods used to estimate failure probability are often limited by the costs associated with model evaluations. Many of these methods, such as multifidelity importance sampling (MFIS), rely upon a computationally efficient, surrogate model like a Gaussian process (GP) to quickly generate predictions. The quality of the GP fit, particularly in the vicinity of the failure region(s), is instrumental in supplying accurately predicted failures for such strategies. We introduce an entropy-based GP adaptive design that, when paired with MFIS, provides more accurate failure probability estimates and with higher confidence. We show that our greedy data acquisition strategy better identifies multiple failure regions compared to existing contour-finding schemes. We then extend the method to batch selection, without sacrificing accuracy. Illustrative examples are provided on benchmark data as well as an application to an impact damage simulator for National Aeronautics and Space Administration (NASA) spacesuits.