An improvement for mewma in multivariate process control

Abstract Multivariate exponentially weighted moving average(MEWMA) chart is a control chart under multivariate process based on the idea of EWMA that rapidly detect small dispersion with a trend. Woodalll [1] first studied the MEWMA, however, the problem existing in his approach is that the selection of constant or diagonal matrix for smooth parameter, because of the correlativity among the multivariate in general case. the smooth parameter should be of general pattern. For this reason, this paper will investigate the MEWMA chart with the generalised smooth paramepter matrix. The background and technique of the MEWMA to be studied in this paper will be much better to fit the practical production process than Woodall's.