Under the normality assurnption. four univariate exponentially nioving average single control charts are proposed and they are designed to monitor simultaneously both the process mean and the process variability The perforniances of these four charts are evaluated by comparing their average run lengths aniong tliemselves as well as to two other competing combination charts. Based on thcl cornparison of the six ~inivariate charts. a multivariate exponentially nioving average single control chart is deveioped as an extension of one of the best univariatt? charts. This chart perfornis better than the combination of the two widely iisctl riiultivariate chürts when small changes are of interest. In dealing with positively-skewed distributed data. the direct logarit hniic trarisformation may result in a control chart wi th inappropriate control paranieters in the application of quality control. When a specific interval for the lognornial mean is given. a new method is introduced to set up two control charts and these two charts can monitor a process for which the underlying distribution of the quality characteristic is lognormal.
[1]
J. Edward Jackson,et al.
Quality Control Methods for Several Related Variables
,
1959
.
[2]
Smiley W. Cheng,et al.
Semicircle Control Chart for Variables Data
,
1996
.
[3]
Herbert Robbins,et al.
The Distribution of a Definite Quadratic Form
,
1948
.
[4]
Stephen V. Crowder,et al.
An EWMA for Monitoring a Process Standard Deviation
,
1992
.
[5]
Rickie J. Domangue,et al.
Some omnibus exponentially weighted moving average statistical process monitoring schemes
,
1991
.
[6]
S. W. Roberts,et al.
Control Chart Tests Based on Geometric Moving Averages
,
2000,
Technometrics.
[7]
D. Montgomery,et al.
Contributors to a multivariate statistical process control chart signal
,
1996
.
[8]
Joachim I. Weindling,et al.
Mean Action Time of the X-Bar Control Chart with Warning Limits
,
1970
.
[9]
Francis Bernard Alt.
Aspects of multivariate control charts
,
1973
.