On the performance of EWMA and DEWMA control charts for censored data

Syed Muhammad Muslim Raza, Muhammad Riaz*, Sajid Ali

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

25 Scopus citations

Abstract

Exponentially weighted moving average (EWMA) control charts are very widely used for the detection of small shifts. Another similar charting structure is double EWMA (DEWMA) control chart for the improved detection of the shifts. Many interesting features of EWMA and DEWMA have been described in the literature. This study intends to investigate EWMA and DEWMA control charts under Type-I censoring for gamma-distributed lifetimes. The idea of conditional expected values is used to monitor the mean level. The performance evaluations are carried out using average run length as a measure in this study. The optimum sample size comparisons for the specified and unspecified parameter are also part of the study. To assess the overall performance of the control charts, we also used extra quadratic loss and it is found DEWMA is an efficient chart for the detection of shift in scale parameter. Moreover, an illustrative example for practical considerations is included in the study. It is observed that varying censoring rates affect the performance of the chart depending upon the type of chart, the method of estimation, and the amount of shift.

Original languageEnglish
Pages (from-to)714-722
Number of pages9
JournalJournal of the Chinese Institute of Engineers, Transactions of the Chinese Institute of Engineers,Series A/Chung-kuo Kung Ch'eng Hsuch K'an
Volume38
Issue number6
DOIs
StatePublished - 18 Aug 2015

Bibliographical note

Publisher Copyright:
© 2015 The Chinese Institute of Engineers.

Keywords

  • Bayes estimates
  • EWMA
  • MLE
  • average run length
  • conditional expected values
  • control charts
  • extra quadratic loss
  • gamma processes
  • type-I censoring

ASJC Scopus subject areas

  • General Engineering

Fingerprint

Dive into the research topics of 'On the performance of EWMA and DEWMA control charts for censored data'. Together they form a unique fingerprint.

Cite this