A Performance Analysis of Multivariate Control Charts

Authors

  • Rodrigo Luiz Pereira Lara
  • José Ivo Ribeiro Júnior
  • Rafael Lédo Rocha de Oliveira

Keywords:

False alarme, Design, ARL IC

Abstract

The improvement and quality monitoring are concerns and needs for many organizations. Among the tools that compose the statistical process control, control charts are simple and efficient monitoring of the mean and variability of several traits evaluated. Particularly, multivariate control charts are valuable tools for the quality control of industries. The objective of this study is to determine which control charts are effective in detecting change in mean vector of a process. Therefore, the following multivariate control charts are presented: Hotelling, MEWMA, MCUSUM and ELR with ARL IC = 200. For each of them, 20,000 bivariate processes were simulated under control based on rational subgroups of 500 individual observations. From the upper limit selected to compose the graphics control, we changed the mean vector so as to observe in which subgroup rational, on average, the graphic process detects the shift. The MEWMA control chart performed better, while the ELR and MCUSUM also showed satisfactory performances.

Published

22-01-2015

How to Cite

Lara, R. L. P., Ribeiro Júnior, J. I., & de Oliveira, R. L. R. (2015). A Performance Analysis of Multivariate Control Charts. Sigmae, 3(2), 55–59. Retrieved from https://publicacoes.unifal-mg.edu.br/revistas/index.php/sigmae/article/view/343