Beyond the EWMA chart: Critiquing hybrid approaches

Abstract

AbstractThis study reveals that recent variants of the EWMA chart have problematic weighting functions. These functions assign more weight to past data values than to the most recent ones, a practice that is less advisable. What we observed is that these weighting functions have a detrimental impact on the run length properties of these modified charts. Surprisingly, it turns out that, in practice, using a standard EWMA chart often results in superior overall performance when compared to the recent variants of the EWMA chart.