Reliability Analysis Using Adaptive Physics-Informed Neural Network with Cross-entropy Optimization

doi:10.1016/j.apm.2025.116627

Reliability Analysis Using Adaptive Physics-Informed Neural Network with Cross-entropy Optimization

2025-11-20

0

PRE

AI

B

Boming Liu

陈振中 (Zhenzhong Chen) *

DOI：10.1016/j.apm.2025.116627

Original

Original request for help

Save

Abstract

En 中文

• A framework to improve efficiency of failure probability estimation is proposed. • Cross-entropy sampling is introduced to better capture complex limit state surfaces. • New convergence criteria are introduced to prevent overfitting and reduce iterations. • It is verified superior to traditional methods in terms of accuracy and efficiency.

Reliability Analysis Using Adaptive Physics-Informed Neural Network with Cross-entropy Optimization

Abstract

Journal

Organization