Title: EWMA control charts for multivariate Poisson-distributed data

Authors: Busaba Laungrungrong, Connie M. Borror, Douglas C. Montgomery

Addresses: Department of Industrial Engineering, Arizona State University, P.O. Box 875906, Tempe, AZ 85287-5906, USA. ' Division of Mathematical and Natural Sciences, Arizona State University at the West Campus, P.O. Box 875906, Tempe, AZ 85287-5906, USA. ' Department of Industrial Engineering, Arizona State University, PO Box 875906, Tempe, AZ 85287-5906, USA

Abstract: Much of the research involving simultaneous monitoring of several related quality characteristics that follow a multivariate Poisson distribution has relied on using the normal approximation to the Poisson distribution in order to determine the appropriate control limits. In this paper, evaluation and implementation of MEWMA schemes for count rates using the multivariate Poisson framework itself are presented. We demonstrate that the multivariate EWMA chart-based directly on the multivariate Poisson distribution is superior to one based on normal-theory with respect to the in-control average run length. The proposed scheme performs similarly to one based on normal-theory for detecting an out-of-control process. We also illustrate a step-by-step numerical example on the practical use of the new control chart.

Keywords: MEWMA control charts; multivariate Poisson distribution; EWMA control charts; statistical process control; SPC; multivariate EWMA.

DOI: 10.1504/IJQET.2011.041227

International Journal of Quality Engineering and Technology, 2011 Vol.2 No.3, pp.185 - 211

Published online: 21 Feb 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article