Adaptive CUSUM Control Chart Utilizing Supervised Learning for Monitoring the Process Location Parameter: A Case Study Application
Quality and Reliability Engineering International · 2025-07 · 41 (5) · 2032-2050
DOI: 10.1002/qre.3760 ↗Kurz gefasst
Fachartikel in der Zeitschrift Quality and Reliability Engineering International, online erschienen am 10.03.2025. Verfasst von Babar Zaman und 1 weiteren Autor:innen. Die Arbeit wurde bisher 15-mal zitiert (Crossref) und nennt 36 Literaturangaben.
Abstract (Originalsprache)
The classical cumulative sum (CUSUM) control chart (CC) is effective in detecting small‐to‐moderate shifts in the process location parameter. It excels at identifying predefined shifts based on a specific parameter (known as ). However, in real‐world applications, shifts in the process may vary in magnitude, which can degrade the performance of this CC. To address this limitation, this study proposes an adaptive CUSUM (ACUSUM) control chart that leverages machine learning techniques. Specifically, a supervised machine learning model, support vector regression (SVR), is employed to enhance its adaptability. The key innovation lies in using the SVR model to train inputs (i.e., characteristics of interest) and outputs (i.e., the parameter ) to predict dynamically, based on the current process conditions and the magnitude of shifts. This newly developed ACUSUM CC is referred to as the CC. Performance evaluation of the CC is conducted using the following two metrics: the average run length and the standard deviation of the run length. These metrics are compared against those of other control charts through a MATLAB algorithm that incorporates Monte Carlo simulation techniques. The results demonstrate that the proposed CC outperforms some existing CCs in detecting shifts. To validate its practical application, the CC is tested on real‐life data from the manufacturing industry, illustrating its effectiveness and efficiency in a practical context.
Zitieren
APA
Zaman, B., & Khan, N. (2025). Adaptive CUSUM Control Chart Utilizing Supervised Learning for Monitoring the Process Location Parameter: A Case Study Application. Quality and Reliability Engineering International, 41(5), 2032-2050. https://doi.org/10.1002/qre.3760
Vancouver
Zaman B, Khan N. Adaptive CUSUM Control Chart Utilizing Supervised Learning for Monitoring the Process Location Parameter: A Case Study Application. Quality and Reliability Engineering International. 2025;41(5):2032-2050. doi:10.1002/qre.3760
BibTeX
@article{Zaman2025,
author = {Zaman, Babar and Khan, Naveed},
title = {Adaptive CUSUM Control Chart Utilizing Supervised Learning for Monitoring the Process Location Parameter: A Case Study Application},
journal = {Quality and Reliability Engineering International},
year = {2025},
volume = {41},
number = {5},
pages = {2032--2050},
doi = {10.1002/qre.3760}
}
EUVIMED