Abstract

[This abstract is based on the authors' abstract.]The popular multivariate exponentially weighted moving average chart (MEWMA) focuses on changes in the mean vector, but changes can occur in either the location or the variability of the correlated multivariate quality characteristic that call for parallel methodologies for detecting changes in the covariance matrix. An exponentially weighted moving covariance matrix is considered for monitoring the stability of the covariance matrix of a process. When used together with the location MEWMA, this chart monitors both mean and variability as required by proper process control. The chart generally outperforms competitive charts for the covariance matrix.