Performance Comparisons of EWMA Control Chart Schemes

作     者：Simoes, Bruno F. T. Epprecht, Eugenio K. Costa, Antonio F. B. 

作者机构：Pontif Catholic Univ Rio de Janeiro PUC Rio Dept Ind Engn Rio De Janeiro Brazil Sao Paulo State Univ UNESP Dept Prod Guaratingueta SP Brazil 

出 版 物：《QUALITY TECHNOLOGY AND QUANTITATIVE MANAGEMENT》 

年 卷 期：2010年第7卷第3期

页      面：249-261页

核心收录：

基　　金：CAPES (Coordenacao de Aperfeicoamento de Pessoal de Nivel Superior) CNPq (Conselho Nacional de Desenvolvimento Cientifico e Tecnologico) 

主　　题：Adaptive control charts ARL profiles exponentially weighted moving average statistical process control variable parameters variable smoothing constant 

摘      要：Combining an EWMA chart with a Shewhart chart is traditionally recommended as a means of providing good protection against both small and large shifts in the process mean. Capizzi and Masarotto have proposed an EWMA chart with a variable smoothing constant (AEWMA) for the same purpose. In this paper, we optimize the designs of the AEWMA and of the combined EWMA-Shewhart schemes with regard to pairs of shifts in the process mean and compare their performances. When the schemes are optimized for the same pair of shifts, their ARL profiles practically coincide. The choice between the AEWMA and the EWMA-Shewhart scheme then becomes a matter of personal preference. We also explore using an AEWMA together with a Shewhart chart, but we find no performance improvement. An additional contribution of this paper is the tables of optimal designs of each scheme for several pairs of shifts.