A new phase II change-point detection based EWMA chart for monitoring and diagnostics of linear profiles

ABSTRACT

A new Phase II control chart, which is based on the change-point model and combined with the exponentially weighted moving average (EWMA) mechanism, is proposed to monitor general linear profiles. The new control chart can be used to monitor general linear profiles when the true in-control parameters are unknown, and only a few historical data are available. In addition, when the chart triggers an out-of-control signal, not only can it estimate the location of the change point, it can also identify which of the parameters have changed as well as the change directions. Using Monte-Carlo simulations, the proposed chart is shown to be effective and has good diagnostic performance. Furthermore, the simulation results show that the proposed chart has better performance than the existing charts in most of the out-of-control scenarios considered. An example is used to illustrate how the proposed chart can be implemented in practical applications.

KEYWORDS:

• Change point
• diagnostics
• EWMA
• linear profiles
• Phase II monitoring

Acknowledgements

We thank the Editor and two anonymous reviewers for their insightful comments and suggestions which improve the presentation of the paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Longcheen Huwang

Dr. Longcheen Huwang received his PhD in Statistics from Cornell University in 1991. After that, he has been a regular faculty of National Tsing Hua University in Taiwan. Currently, he is a full professor of Institute of Statistics at National Tsing Hua University. His current research interests lie in industrial statistics, especially statistical process control. He has published many research papers in the leading journals, such as Annals of Statistics, Technometrics, Journal of Quality Technology, Journal of Multivariate Analysis, Statistica Sinica, Canadian Journal of Statistics, IIE Transactions, Naval Research Logistics, etc. He is an elected member of International Statistical Institute.

Arthur B. Yeh

Dr. Arthur B. Yeh is Professor of Statistics at the Department of Applied Statistics and Operations Research, Allen W. and Carol M. Schmidthorst College of Business, Bowling Green State University. He received his Ph.D. in statistics from Rutgers University. He previously served as Chair of the Department and Associate Dean of the College. He was the Owens-Illinois Professor in the College from 2015 – 2018. Dr. Yeh has published peer-reviewed journal articles in topic areas including optimal experimental designs, univariate and multivariate control charts, multivariate process capability indices, univariate and multivariate run-by-run process control, and statistical profile monitoring. He has served as an associate editor for The Statistical Papers and The American Statistician.

Yi-Wen Wang

Yi-Wen Wang received her MS degree in Statistics from National Tsing Hua University, Hsinchu, Taiwan in 2022. She is currently an engineer at Taiwan Semiconductor Manufacturing Company in Tainan, Taiwan.