An acronym for Exponentially Weighted Moving Average Chart, this control chart is used to detect small or gradual shifts in the performance of a continuous process. It was developed by S W Roberts in 1959.
The EWMA chart is essentially a moving average chart with a weight (0 < w < 1) assigned to the current measurement and its complement (1-w) assigned to the sum of all the previous measurements. Thus, all of the collected data are included at each step to help judge the stability of the process.
But the user decides how much influence the preceding measurements can have on the plotted statistic by the choice of the weighting factor. This results in a 'smoothing' effect on the noise variation, allowing faster detection of small shifts in the mean.
EWMA charts model the time series data in order to forecast the next value and detect lack of control in the time series. Because it uses information from all of the data, it is able to detect much smaller process shifts than a regular control chart can. However, due to its smoothing effect, it is sometimes slow to respond to sudden or large changes.
Using Exponentially Weighted Average (EWMA) Charts by Keith M. Bower, Minitab. Originally published in Asia Pacific Process Engineer, October 2000. - http://www.minitab.com/resources/articles/UsingEWMACharts.pdf EWMA Control Charts from NIST: - http://www.itl.nist.gov/div898/handbook/pmc/section3/pmc324.htm