Improved Monitoring of Multivariate Process Variability

Abstract

The standard way of using the determinant of the average of sample covariance matrices for multivariate process variability control charting leads to a biased estimate of the control limits. In this paper, we present an improved control chart having unbiased control limits. It has smaller ARL than the standard chart and, hence, it is more sensitive to shifts in the process variability. An example is given to illustrate these advantages.