Desensitized control charts with operational importance for autocorrelated processes

Abstract

ABSTRACT Organizations that achieved higher capabilities to meet expectations by successfully implementing quality improvement plans can make adjustments to process monitoring without any losses for customers. This approach may reduce costs due to desensitization in monitoring. In this regard, the acceptance control chart (ACC) under the assumptions of independence and normality has already been introduced. However, in practice, particular correlation patterns can be extracted among samples that violate the assumption of independence. Autocorrelation reduces the performance of traditional control charts. In the present study, three types of ACC are extended for monitoring the mean of the most widely used autocorrelated processes. The performance of the proposed charts is evaluated using statistical measures. It is found that the residual-based exponentially weighted moving average ACC (R-EWMA-ACC) has the best performance. Also, two numerical examples illustrate its application. Moreover, optimizing the economic-statistical model indicates that the R-EWMA-ACC spends less than the ARMA control chart.