Multivariate Cuscore control charts for monitoring the mean vector in autocorrelated processes

In many systems, quantitative observations of process variables can be used to characterize a process for quality control purposes. As the intervals between observations become shorter, autocorrelation may occur and lead to a high false alarm rate in traditional Statistical Process Control (SPC) charts. In this article, a Multivariate Cuscore (MCuscore) SPC procedure based on the sequential likelihood ratio test and fault signature analysis is developed for monitoring the mean vector of an autocorrelated multivariate process. The MCuscore charts for the transient, steady and ramp mean shift signal are designed; they do not rely on the assumption of known signal starting time. An example is presented to demonstrate the application of the MCuscore chart to monitoring three autocorrelated variables of an online search engine marketing tracking process. Furthermore, the simulation analysis shows that the MCuscore chart outperforms the traditional multivariate cumulative sum control chart in detecting process shifts.

[1]  Stefan H. Steiner,et al.  A Multivariate Robust Control Chart for Individual Observations , 2009 .

[2]  Gyo-Young Cho,et al.  Multivariate Control Charts for Monitoring the Mean Vector and Covariance Matrix , 2006 .

[3]  J. Healy.  A note on multivariate CUSUM procedures , 1987 .

[4]  G. Jenkins,et al.  Some Statistical Aspects of Adaptive Optimization and Control , 1962 .

[5]  Francisco Aparisi,et al.  Techniques to interpret T 2 control chart signals , 2006 .

[6]  Daniel W. Apley,et al.  Autocorrelated process monitoring using triggered cuscore charts , 2002 .

[7]  Douglas M. Hawkins,et al.  Self-Starting Multivariate Exponentially Weighted Moving Average Control Charting , 2007, Technometrics.

[8]  Wei Jiang,et al.  A Weighted CUSUM Chart for Detecting Patterned Mean Shifts , 2008 .

[9]  Arthur B. Yeh,et al.  A multivariate exponentially weighted moving average control chart for monitoring process variability , 2003 .

[10]  Seongkyu Yoon,et al.  Fault diagnosis with multivariate statistical models part I: using steady state fault signatures , 2001 .

[11]  Massimo Pacella,et al.  Statistical Process Control for Geometric Specifications: On the Monitoring of Roundness Profiles , 2008 .

[12]  D. A. Evans,et al.  An approach to the probability distribution of cusum run length , 1972 .

[13]  Jeffrey E. Jarrett,et al.  The quality control chart for monitoring multivariate autocorrelated processes , 2007, Comput. Stat. Data Anal..

[14]  Frederick W. Faltin,et al.  Statistical Control by Monitoring and Feedback Adjustment , 1999, Technometrics.

[15]  Ratna Babu Chinnam,et al.  Robust kernel distance multivariate control chart using support vector principles , 2008 .

[16]  Rassoul Noorossana,et al.  Effect of Autocorrelation on Performance of the MCUSUM Control Chart , 2006, Qual. Reliab. Eng. Int..

[17]  Alberto Luceño,et al.  Cuscore Charts to Detect Level Shifts in Autocorrelated Noise , 2004 .

[18]  Harriet Black Nembhard,et al.  Periodic Cuscore charts to detect step shifts in autocorrelated processes , 2008, Qual. Reliab. Eng. Int..

[19]  Charles W. Champ,et al.  A multivariate exponentially weighted moving average control chart , 1992 .

[20]  Fugee Tsung,et al.  An Adaptive T2 Chart for Monitoring Dynamic Systems , 2008 .

[21]  Stelios Psarakis,et al.  Multivariate statistical process control charts: an overview , 2007, Qual. Reliab. Eng. Int..

[22]  William H. Woodall,et al.  A one‐sided MEWMA chart for health surveillance , 2008, Qual. Reliab. Eng. Int..

[23]  Gwilym M. Jenkins,et al.  Time series analysis, forecasting and control , 1972 .

[24]  Chinmo Roan,et al.  CHANGE PATTERNS OF TIME SERIES-BASED CONTROL CHARTS , 1996 .

[25]  Jianjun Shi,et al.  The GLRT for statistical process control of autocorrelated processes , 1999 .

[26]  Harriet Nembhard,et al.  Cuscore Statistics: Directed Process Monitoring for Early Problem Detection , 2006 .

[27]  B. Bakshi.  Multiscale PCA with application to multivariate statistical process monitoring , 1998 .

[28]  Ruey S. Tsay,et al.  Analysis of Financial Time Series: Tsay/Analysis of Financial Time Series , 2005 .

[29]  George E. P. Box,et al.  Sampling and Bayes' inference in scientific modelling and robustness , 1980 .

[30]  Fugee Tsung,et al.  Comparison of the cuscore, GLRT and cusum control charts for detecting a dynamic mean change , 2005 .

[31]  R. Crosier.  Multivariate generalizations of cumulative sum quality-control schemes , 1988 .

[32]  Douglas M. Hawkins,et al.  A Multivariate Change-Point Model for Change in Mean Vector and/or Covariance Structure , 2009 .

[33]  George C. Runger,et al.  Multivariate one-sided control charts , 2006 .

[34]  Fugee Tsung,et al.  A Reference-Free Cuscore Chart for Dynamic Mean Change Detection and a Unified Framework for Charting Performance Comparison , 2006 .

[35]  Jonathan D. Cryer,et al.  Time Series Analysis , 1986 .

[36]  Elisabeth J. Umble,et al.  Cumulative Sum Charts and Charting for Quality Improvement , 2001, Technometrics.

[37]  George C. Runger,et al.  A Markov Chain Model for the Multivariate Exponentially Weighted Moving Averages Control Chart , 1996 .

[38]  Nien Fan Zhang,et al.  Forecasting and time series analysis , 1976 .

[39]  H. Hotelling.  Multivariate Quality Control-illustrated by the air testing of sample bombsights , 1947 .

[40]  Michael Pokojovy,et al.  A Multistep, Cluster-Based Multivariate Chart for Retrospective Monitoring of Individuals , 2009 .