An EWMA Solution to Detect Shifts in a Bernoulli Process in an Out‐of‐control Environment

Abstract

AbstractThe Exponentially Weighted Moving Average (EWMA) control chart has mainly been used to monitor continuous data, usually under the normality assumption. In addition, a number of EWMA control charts have been proposed for Poisson data. Here, however, we suggest applying the EWMA to hypergeometric data originating from a multivariate Bernoulli process. The problem studied in this paper concerns the wear‐out of electronics testers resulting in unnecessary and costly reparations of electronic units. Assuming that the testing process is in statistical control, although the quality of the tested units is not, we can detect the wear‐out of a tester by finding assignable causes of variation in that tester. This reasoning forms the basis of a new EWMA procedure designed to detect shifts in a Bernoulli process in an out‐of‐control environment. Copyright © 2005 John Wiley & Sons, Ltd.