The effectiveness of MEWMA control chart with mild correlation

Abstract

Since it is undeniable that whenever conventional control charts are used there is the implied presumption that this observations tend to be independently as well as identically distributed with time. However, in reality, this kind of observations generated through continuous as well as discrete production procedures tend to be serially correlated, which violates the independence assumption of conventional control charts in addition to affect the performance of control charts negatively. In this paper, we investigate the performance of MEWMA control chart with autocorrelated data with mild correlation being controlled. The generated data were applied to MEWMA control chart procedure and showed an in-control state, while the generated observations were subjected to normality tests from the assumptions and sensitivities for departure to normality, and turned out to be normal by all standard. Hence, this gives an alternative for the quality practitioners to adopt for the continuous and discrete production processes also the autocorrelation has no effect on the performance of MEWMA control limits when the mild correlation has been controlled.