Track: Quality Engineering, Control and Management

Abstract

The quality of process or product is often characterized by the univariate or multivariate distribution of quality characteristics. The linear profiles are widely used to model the relationship between dependent and independent variables. This paper aims to investigate the statistical performance of the double exponentially weighted moving average control chart (dEWMA) in quickly detecting the existence of assignable causes under different patterns of process shift. The performance of three existing control chart is estimated and compared using the Average Run Length (ARL) as a criterion of comparison.  For all the compared charts, the ARL and standard deviation (SDRL) at several parameters of shift patterns are estimated in order to draw conclusions and provide future work guidelines.