A Robust Phase I Exponentially Weighted Moving Average Chart for Dispersion

Publisher: John Wiley & Sons Inc

E-ISSN: 1099-1638|31|6|989-999

ISSN: 0748-8017

Source: QUALITY AND RELIABILITY ENGINEERING INTERNATIONAL, Vol.31, Iss.6, 2015-10, pp. : 989-999

Disclaimer: Any content in publications that violate the sovereignty, the constitution or regulations of the PRC is not accepted or approved by CNPIEC.

Previous Menu Next

Abstract

A Phase I estimator of the dispersion should be efficient under in‐control data and robust against contaminations. Most estimation methods proposed in the literature are either efficient or robust against either sustained shifts or scattered disturbances. In this article, we propose a new estimation method of the dispersion parameter, based on exponentially weighted moving average charting, which is efficient and robust to both types of unacceptable observations in Phase I. We compare the method with various existing estimation methods and show that the proposed method has the best overall performance if it is unknown what type of contaminations are present in Phase I. We also study the effect of the robust estimator from Phase I on the Phase II exponentially weighted moving average control chart performance. Copyright © 2014 John Wiley & Sons, Ltd.