Novel design of composite generally weighted moving average and cumulative sum charts

摘要

The cumulative sum (CUSUM) and exponentially weighted moving average (EWMA) charts are popular statistical tools to improve the performance of the Shewhart chart in detecting small process shifts. In this study, we propose the mixed generally weighted moving average (GWMA)-CUSUM chart and its reverse-order CUSUM-GWMA chart to enhance detection ability compared with existing counterparts. The simulation revealed that the mixed GWMA-CUSUM and mixed CUSUM-GWMA charts have the sensitivity to detect small process shifts and efficient structures compared with the mixed EWMA-CUSUM and mixed CUSUM-EWMA charts, respectively. Moreover, the mixed GWMA-CUSUM chart with a large design parameter has robust performance, regardless of the high tail t distribution or right skewness gamma distribution.