1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
|
\name{sewma.q.prerun}
\alias{sewma.q.prerun}
\alias{sewma.q.crit.prerun}
\title{Compute RL quantiles of EWMA (variance charts) control charts under pre-run uncertainty}
\description{Computation of quantiles of the Run Length (RL) for EWMA control charts monitoring
normal variance.}
\usage{sewma.q.prerun(l,cl,cu,sigma,df1,df2,alpha,hs=1,sided="upper",
r=40,qm=30,qm.sigma=30,truncate=1e-10)
sewma.q.crit.prerun(l,L0,alpha,df1,df2,sigma0=1,cl=NULL,cu=NULL,hs=1,
sided="upper",mode="fixed",r=40, qm=30,qm.sigma=30,truncate=1e-10,
tail_approx=TRUE,c.error=1e-10,a.error=1e-9)}
\arguments{
\item{l}{smoothing parameter lambda of the EWMA control chart.}
\item{cl}{deployed for \code{sided}=\code{"Rupper"}, that is, upper variance control chart
with lower reflecting barrier \code{cl}.}
\item{cu}{for two-sided (\code{sided}=\code{"two"}) and fixed upper control limit
(\code{mode}=\code{"fixed"}) a value larger than \code{sigma0}
has to been given, for all other cases \code{cu} is ignored.}
\item{sigma,sigma0}{true and in-control standard deviation, respectively.}
\item{L0}{in-control quantile value.}
\item{alpha}{quantile level.}
\item{df1}{actual degrees of freedom, corresponds to subgroup size (for known mean it is
equal to the subgroup size, for unknown mean it is equal to subgroup size minus one.}
\item{df2}{degrees of freedom of the pre-run variance estimator.}
\item{hs}{so-called headstart (enables fast initial response).}
\item{sided}{distinguishes between one- and two-sided two-sided EWMA-\eqn{S^2}{S^2}
control charts by choosing \code{"upper"} (upper chart
without reflection at \code{cl} -- the actual value of \code{cl} is not used),
\code{"Rupper"} (upper chart with reflection at \code{cl}),
\code{"Rlower"} (lower chart with reflection at \code{cu}),
and \code{"two"} (two-sided chart), respectively.}
\item{mode}{only deployed for \code{sided}=\code{"two"} -- with \code{"fixed"}
an upper control limit (see \code{cu}) is set and only the lower is
calculated to obtain the in-control ARL \code{L0}, while with \code{"unbiased"} a certain unbiasedness
of the ARL function is guaranteed (here, both the
lower and the upper control limit are calculated).}
\item{r}{dimension of the resulting linear equation system (highest order
of the collocation polynomials).}
\item{qm}{number of quadrature nodes for calculating the collocation definite integrals.}
\item{qm.sigma}{number of quadrature nodes for convoluting the standard deviation uncertainty.}
\item{truncate}{size of truncated tail.}
\item{tail_approx}{controls whether the geometric tail approximation is used (is faster) or not.}
\item{c.error}{error bound for two succeeding values of the critical value
during applying the secant rule.}
\item{a.error}{error bound for the quantile level \code{alpha} during applying the secant rule.}
}
\details{
Instead of the popular ARL (Average Run Length) quantiles of the EWMA
stopping time (Run Length) are determined. The algorithm is based on
Waldmann's survival function iteration procedure.
Thereby the ideas presented in Knoth (2007) are used.
\code{sewma.q.crit.prerun} determines the critical values (similar to alarm limits)
for given in-control RL quantile \code{L0} at level \code{alpha} by applying secant
rule and using \code{sewma.sf()}.
In case of \code{sided}=\code{"two"} and \code{mode}=\code{"unbiased"}
a two-dimensional secant rule is applied that also ensures that the
minimum of the cdf for given standard deviation is attained at \code{sigma0}.
}
\value{Returns a single value which resembles the RL quantile of order \code{alpha} and
the lower and upper control limit \code{cl} and \code{cu}, respectively.}
\references{
S. Knoth (2007),
Accurate ARL calculation for EWMA control charts monitoring simultaneously normal mean and variance,
\emph{Sequential Analysis 26}, 251-264.
K.-H. Waldmann (1986),
Bounds for the distribution of the run length of geometric moving
average charts, \emph{Appl. Statist. 35}, 151-158.
}
\author{Sven Knoth}
\seealso{\code{sewma.q} and \code{sewma.q.crit} for the version w/o pre-run uncertainty.}
\examples{
## will follow
}
\keyword{ts}
|
