Title: A novel enhanced kriging-based method for reliability analysis integrating Bayesian optimisation with ensemble strategy

Authors: Yongbo Cheng; Manman Dong; Liangqi Wan

Addresses: College of Economics and Management, Nanjing University of Aeronautics and Astronautics, Nanjing, 211106, China; School of Management Science and Engineering, Nanjing University of Finance and Economics, Nanjing, 210046, China ' School of Economics and Management, Zhejiang University of Water Resources and Electric Power, Hangzhou, 310018, China ' School of Management Science and Engineering, Nanjing University of Finance and Economics, Nanjing, 210046, China

Abstract: This paper introduces an innovative and adaptive enhanced kriging-based approach for reliability analysis, combining Bayesian optimisation with posterior probability. The proposed methodology aims to overcome the shortcomings of inaccuracy of traditional kriging methods. The enhanced kriging model offers a more accurate prediction by an ensemble of different individual kriging models. Bayesian optimisation enhances the reliability assessment by providing a probabilistic measure of the model's accuracy. Furthermore, the posterior probability is employed to calculate the optimal weights for each kriging model. The approach improves the accuracy and efficiency of reliability analysis in complex systems. To demonstrate the effectiveness of the proposed approach, two illustrative examples including a high-nonlinear problem and a small failure probability problem are presented, showcasing its ability to provide accurate and robust failure probability estimates. This novel integration of kriging, Bayesian optimisation, and posterior probability offers great instruction in improving product quality and engineering applications.

Keywords: quality management; kriging; reliability analysis; Bayesian optimisation; ensemble strategy.

DOI: 10.1504/IJPQM.2026.152590

International Journal of Productivity and Quality Management, 2026 Vol.47 No.3, pp.361 - 378

Received: 25 Dec 2023
Accepted: 03 Jan 2024
Published online: 30 Mar 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article