Manuel Roman Piña-Monarrez


In multivariate control process, the T2 Hotelling chart is the most useful, unfortunately, until now there was not an efficient decomposition method when the chart signals. The MTY method (Mason, Tracy and Young) is the most used method, but unfortunately it does not present a not unique decomposition and it is complex when there are many variables. In particular the estimated Mahalanobis distance (MD) and its decomposition for an out of control signal, not hold. In this article, by considering the inverse of the covariance matrix of phase I as a standard one, a practical decomposition method based on the relations of each pair of variables in this inverse matrix, is proposed. Results of its application show that the decomposition holds with the estimated MD, and that, let us to separate the contribution of each individual variable; in contribution due the variance and contribution due the correlation.


Decomposition method, T2 Hotelling chart, Mahalanobis distance, Multivariate control

Full Text:

 Subscribers Only

Online ISSN 1943-670X

The International Journal of Industrial Engineering is a Non-Profit Organization under Section 501(c)(3).

The journal is registered with the Library of Congress and the Copyright Clearance Center. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopy, recording or otherwise, without the prior written permission of the journal.

The International Journal of Industrial Engineering is abstracted by Thomson Reuters Science Citation Index, Engineering Index/Compendex, EIPAGE1, and CURRENT CONTENTS/ENGINEERING, COMPUTING AND TECHNOLOGY.