An enhanced GWMA chart for process mean

Abstract

Memory-type control charts are frequently used to monitor small-to-moderate disturbances in the process parameter(s). Recently, an auxiliary-information-based (AIB) exponentially weighted moving average (EWMA) chart has been proposed for efficiently monitoring the process mean, called the AIB-EWMA chart, that surpasses the classical EWMA chart. On similar lines, in order to further enhance the sensitivity of the AIB-EWMA chart, we propose an AIB generally weighted moving average (GWMA) chart for monitoring the process mean, named the AIB-GWMA chart. The existing AIB-Shewhart and AIB-EWMA charts are special cases of the AIB-GWMA chart. With the help of Monte Carlo simulation, the run length characteristics of the AIB-GWMA chart are computed. It turns out that the AIB-GWMA chart performs uniformly and substantially better than the classical GWMA chart. Moreover, the AIB-GWMA chart also outperforms the AIB-EWMA chart when detecting mean shifts within certain ranges. A similar trend is observed when the proposed and existing control charts are compared with the fast initial response features. An example is used to explain the working and implementation of the GWMA, AIB-EWMA and AIB-GWMA charts.

KEYWORDS:

• Auxiliary information
• Average run length
• EWMA
• Fast initial response
• GWMA
• Statistical process control