Performance comparison of shewhart joint monitoring schemes for mean and variance

Abstract

In quality control, joint monitoring schemes for mean and variance are preferred for situations in which special causes can result in change in both the mean and the variance. Several such joint monitoring schemes are reported in the literature to monitor the mean and variance simultaneously of a normally distributed process. Like in the single monitoring of one variable, in the joint monitoring also, combined Shewhart scheme for mean and variance is preferred by many practitioners because of its simplicity compared to combined cumulative sum (CUSUM) and exponentially weighted moving average (EWMA) schemes. Four such Shewhart combined schemes are proposed by three authors in the literature. In this study, the performances of these four schemes are compared using average run length properties under a common platform. Overall, the Shewhart distance scheme performs best and the poorest performance is observed for Shewhart scheme with rectangular acceptance region.