Adaptive Max charts for monitoring process mean and variability

Abstract

Max chart uses a single chart to control the process mean and variance. It performs with similar efficiency to joint and s charts. This study combined adaptive chart and Max chart to increase the detecting ability. A Markov chain approach was used to develop the statistical indicator and statistical design model of adaptive Max charts. This study designed adaptive Max charts to detect the mean shift and variance increase, and measured their performance for comparison with EWMA, CUSUM, and double-sampling charts. A comparison study revealed that the adaptive Max charts with variable parameter exhibited superior performance to detect the small shift than that of EWMA, CUSUM, and double sampling charts.