A PROPOSED ENHANCEMENT FOR THE SHORT RUNS MULTIVARIATE CONTROL CHART FOR THE PROCESS MEAN
Loading...
Date
2006-03
Authors
FOO, NG THEAM
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Multivariate control charts are gaining attention because in practice, most
processes involve the monitoring of several correlated variables. The most well
known multivariate process monitoring method is the Hotelling's T 2 control
chart. The Hotelling's T 2 control chart is used in process monitoring which
involves mass production where the data to estimate the mean vector and
covariance matrix are readily available. Recently, manufacturing industries tend
to produce products in smaller lot sizes or low volume production. Such a
production is known as short runs production or more commonly known as short
runs. Short runs are becoming more important as a result of increased
emphasis on just-in-time techniques, synchronous manufacturing, job-shop
settings and the reduction of in-process inventory and costs. Thus, to enable
the Hotelling's T 2 chart to be used in short runs, modifications of the Hotelling's
T 2 statistics are required. The existing short runs multivariate charts for the
process mean are based on individual measurements. In this thesis, a method
to enhance the performance of the short runs multivariate control chart will be
proposed for the case where the process covariance matrix is unknown and
need to be estimated. The proposed method is based on the use of a robust
estimator of the modified mean square successive differences (MSSD),
replacing the sample covariance matrix estimator. The performance of the
enhanced chart will be compared to that of the existing chart. The simulation
results show that the enhanced chart gives superior results. An example is also
given to show how the enhanced chart is used in a real situation.
Description
Keywords
A PROPOSED ENHANCEMENT FOR THE SHORT RUNS MULTIVARIATE CONTROL