Control charts for monitoring mean of generalized exponential distribution with type‐I censoring

Abstract

AbstractMonitoring censored data is a challenging task as the traditional charts reported poor performance in the case of censored data. This article presents exponentially weighted moving average control charts for monitoring type‐I censored data assuming generalized exponential (GE) distribution. In particular, we replace the censored observation with the conditional expected value (CEV), conditional median (CM), and conditional standard deviation (CSD) to assess the performance of the charts. To evaluate the performance of the memory‐type charts, this study considers different shifts, smoothing weights, and censoring rates in order to compare their effectiveness. Besides a comprehensive study of average run length (ARL) using Monte Carlo simulations, special attention is paid to the standard deviation of the run length (SDRL), because it can be substantially larger for skewed distributions. Furthermore, the application of the proposed charts is illustrated by using two real data sets. Different situations are identified where CEV, CM, and CSD are preferred over each other.