<meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />안녕, 모두
추가 실행 규칙이 있는 Shewhart 관리도의 실행 길이 분포를 평가하기 위한 기존 JMP 스크립트 또는 프로세스가 있는지 묻는 고객이 있습니까? 동봉된 SAS 스크립트는 이 작업을 수행합니다. JMP에 해당하는 것을 아는 사람이 있습니까?
감사!
실내 변기
업데이트: 기록을 위해 이것은 저에게 일시적인 문제였던 것 같습니다... 이제 파일을 다운로드할 수 있습니다.
이안,
이렇게 해주셔서 대단히 감사합니다!안타깝게도 스크립트에 액세스할 수 없는 것 같습니다... 스크립트를 열거나 컴퓨터에 저장하려고 하면 브라우저가 잠깁니다. 또한 이를 요청한 사용자(및 다른 사용자)가 이를 사용할 수 있도록 제안하는 방법은 무엇입니까? File Exchange에 추가할 의향이 있습니까?
감사!
실내 변기
참고로 어제 JSL로 번역했는데 재미있는 연습이었습니다. UI를 만들지 않은 이유는 (주제의 전문적인 특성을 감안할 때) 필요로 하는 사람은 누구나 코드를 약간만 변경해도 기뻐할 것이라고 생각하기 때문입니다. 사소한 경우를 제외하고는 테스트가 없었습니다.
이는 '평균의 결함'과 ARL만 사용하여 '좋은' 차트를 선택하는 것의 위험성에 대한 또 다른 좋은 예입니다.
건배,
이안
Jerry가 참조하는 SAS Pogram은 다음과 같습니다.
/****************************************************************/
/* S A S S A M P L E L I B R A R Y */
/* */
/* NAME: SHWARL2 */
/* TITLE: ARL With Supplementary Run Rules (SHWARL2) */
/* PRODUCT: QC */
/* SYSTEM: ALL */
/* KEYS: Shewhart Charts, Average Run Lengths, */
/* PROCS: IML GPLOT */
/* DATA: */
/* */
/* NOTES: */
/* MISC: */
/* */
/* REF: SAS/QC Software: Usage and Reference, Version 6, */
/* First Edition, Volume 1 and Volume 2 */
/* */
/* Champ C. W. and Woodall W. H. (1990), "A Program to */
/* Evaluate the Run Length Distribution of a Shewhart */
/* Control Chart with Supplementary Run Rules" */
/* Journal of Quality Technology, Vol. 22, pg. 68-73. */
/* */
/* Champ C. W. and Woodall W. H. (1987), "Exact */
/* Results for Shewhart Control Chart With */
/* Supplementary Runs Rules". Technometrics, Vol. 29, */
/* pg. 393-399. */
/* */
/* Examples in the documentation were created using the */
/* statement: ods listing style=statistical; */
/* */
/****************************************************************/
options ps=60 ls=80 nodate;
title1 'ARL with Supplementary Runs Rules';
* Input the Rules Defining the Chart ;
data rules;
input k m a b;
cards;
1 1 3 9
2 3 2 3
2 3 -3 -2
1 1 -9 -3
;
proc iml;
use rules; read all var {k m a b};
/********************************************************/
/* Input the Standarized Shift, */
/* the Upper Limit for the Calculation of ARL and Std, */
/* and the Upper Limit for the CDF Graph */
/********************************************************/
shf = 0.6;
uppcal = 0.99999;
uppgra = 0.99;
* Find the Regions of the Chart ;
r = -9//a//b//9;
r = unique(r);
nt = nrow(k);
nr = ncol(r)-1;
/***************************************************/
/* Find the Length of the State Vector */
/* and the pointers to the End of the Subvector */
/* Associated with Each Runs Rule */
/***************************************************/
nv = sum(m)-nt;
if(sum(m-k) = 0) then nv = nv+1;
di = j(nt,1,0);
di[1] = m[1]-1;
do i = 2 to nt;
if(m[i] > 1) then di[i] = di[i-1]+m[i]-1;
end;
/**************************************************/
/* Determine the Value of the Present */
/* Indicator Variable by Runs Rule and Region */
/* Combination */
/**************************************************/
a1 = j(nt,1,1);
a1 = a1*r;
x = j(nt,nr,0);
do j = 1 to nr;
x[,j] = (a <= a1[,j])#(b >= a1[,j+1]);
end;
* Determine the State to State Transitions by Regions ;
ps = j(nv,1,0);
s = j(nt,1,0);
a1 = j(1,nr,1);
nx = j(nv,1,0);
qqns = 2**nv-1;
ns = 1;
h = 1;
qq = 0;
q = a1;
do while ( h <= ns );
qh = qq[h];
do L = 1 to nv;
ps[L] = qh-2*int(qh/2);
qh = int(qh/2);
end;
do i = 1 to nt;
s[i] = 0;
if (m[i] > 1) then do
L = di[i]-m[i]+2 to di[i];
s[i] = s[i]+ps[L];
end;
end;
do j = 1 to nr;
sg = 0;
do i = 1 to nt;
if (sg = 0) then do;
if (s[i]+x[i,j] >= k[i]) then sg = 1;
else do;
if (m[i] > 1) then nx[di[i]-m[i]+2] = x[i,j];
if (m[i] > 2) then do
L = di[i]-m[i]+3 to di[i];
nx[L] = ps[L-1];
end;
end;
if (x[i,j] = 0 & m[i] > 1) then do;
tmp = s[i]-ps[di[i]]+1;
L = di[i];
ck = 0;
do while ( ck = 0 & L >= di[i]-m[i]+2 );
if (nx[L] = 1) then do;
ck = 1;
if (tmp < k[i]) then do;
nx[L] = 0;
tmp = tmp-1;
ck = 0;
end;
end;
L = L-1;
tmp = tmp+1;
end;
end;
end;
end;
if (sg = 0) then do;
qh = nx[1];
do L = 2 to nv;
qh = qh+nx[L]*(2**(L-1));
end;
ck = 0;
do L = 1 to ns;
if (ck = 0 & qh = qq[L]) then do;
q[h,j] = qq[L];
ck = 1;
end;
end;
if (ck = 0) then do;
ns = ns+1;
qq = qq//qh;
q[h,j] = qh;
end;
end;
else
q[h,j] = qqns;
end;
h = h+1;
q = q//(qqns*a1);
end;
ns = ns+1;
qq = qq//qqns;
free / q qq ns nr r shf qqns uppcal uppgra;
/******************************************/
/* Sort the States in Ascending Order */
/* of Their Base Two Representations */
/******************************************/
a1 = q;
a1[rank(qq),] = q;
q = a1;
a1 = qq;
a1[rank(qq),] = qq;
qq = a1;
* Remove any Duplicate States ;
ck = 1;
do while (ck = 1);
ck = 0;
i = 1;
do while (i <= ns-1);
j = i+1;
do while (j <= ns);
if (q[i,] = q[j,]) then do;
ck = 1;
q = q[remove(1:ns,j),];
a = (q = (0*q+qq[j]));
q = (q#(1-a))+(qq[i]#a);
qq = qq[remove(1:ns,j),];
j = j-1;
ns = ns-1;
end;
j = j+1;
end;
i = i+1;
end;
end;
* Number the Next-State Transitions ;
a1 = j(1,ns,1);
a2 = (qq*a1)`;
a3 = j(ns,nr,0);
do j = 1 to nr;
a3[,j] = ((q[,j]*a1) = a2)*((1:ns)`);
end;
a1 = (q < (0*q+qqns));
q = (ns#(1-a1))+(a3#a1);
free qq qqns ck a2 a3;
/**********************************************************/
/* Compute the Cumulative Distribution, */
/* Function, the Average Run Length (ARL) and the */
/* Standard Deviation (STD) of the Run Length for a */
/* Given Standarized Shift (SHF) in the Mean. */
/* */
/* The Iterative Process for the Calculation of the */
/* CDF, ARL and STD stops when the CDF is > UPPCAL */
/* */
/* The Number of Recorded Points for the CDF Graph */
/* Stops When CDF is > UPPGRA */
/**********************************************************/
pr = probnorm((r-shf)`);
pr = (pr[2:nr+1,]//0)-pr;
pr = pr[1:nr,];
pt = j(ns,ns,0);
do i = 1 to ns;
do j = 1 to nr;
pt[i,q[i,j]] = pt[i,q[i,j]]+pr[j];
end;
end;
a1 = pt[1:ns-1,ns];
pt = pt[1:ns-1,1:ns-1];
cdf = a1[1];
acu = a1[1];
n = 1;
do while ( acu <= uppcal );
a1 = pt*a1;
acu = acu+a1[1];
if ( acu <= uppgra ) then n=n+1;
cdf = cdf//a1[1];
end;
rl = do(1,nrow(cdf),1);
arl = rl*cdf;
std = sqrt(((rl##2)*cdf)-(arl*arl));
cdf = cusum(cdf);
rl = rl[,1:n]`;
cdf = cdf[1:n,];
create par var{arl std shf}; append; close par;
create cdf var{rl cdf}; append; close cdf;
reset center noname pagesize = 800 linesize = 80;
c1 = {"ARL" "STD"};
arl = arl//std;
print arl[rowname = c1 format = 10.6];
quit;
* Print Every 10th Run Length and associated CDF value ;
title1 'Run Length CDF';
proc print data=cdf noobs label;
label rl='Run Length' cdf='CDF of Run Length';
where mod(rl,10)=0;
run;
/******************************************************/
/* Annotate Dataset that will Superimpose a Table */
/* Containing the Mean and Standard Deviation of the */
/* Run Length */
/******************************************************/
data anno;
set par;
length text function style color $ 8;
shf = put(shf,7.3);
arl = put(arl,7.3);
std = put(std,7.3);
xsys = '1';
ysys = '1';
hsys = '1';
function = 'move'; x = 1; y = 99; output;
style = 'solid';
color = 'blue';
when = 'A';
xsys = 'A';
ysys = 'A';
hsys = '6';
function = 'poly'; x = 0; y = 0; output;
color = 'white';
function = 'polycont'; x = 0; y = -7; output;
function = 'polycont'; x = 17; y = -7; output;
function = 'polycont'; x = 17; y = 0; output;
xsys = '1';
ysys = '1';
hsys = '1';
function = 'move'; x = 1; y = 99; output;
function = 'cntl2txt'; output;
color = 'white';
line = 1;
style = 'none';
size = 1;
position = 'c';
xsys = 'A';
ysys = 'A';
hsys = '6';
x = 1;
function = 'label'; y = -2; text = 'SHIFT'; output;
function = 'cntl2txt'; output;
function = 'label'; y = -4; text = 'ARL'; output;
function = 'cntl2txt'; output;
function = 'label'; y = -6; text = 'STD'; output;
x = 7;
function = 'cntl2txt'; output;
function = 'label'; y = -2; text = '='; output;
function = 'cntl2txt'; output;
function = 'label'; y = -4; text = '='; output;
function = 'cntl2txt'; output;
function = 'label'; y = -6; text = '='; output;
position = 'a';
x = 16;
function = 'cntl2txt'; output;
function = 'label'; y = -2; text = shf; output;
function = 'cntl2txt'; output;
function = 'label'; y = -4; text = arl; output;
function = 'cntl2txt'; output;
function = 'label'; y = -6; text = std; output;
run;
/*************************************************************/
/* Display Graph of the Run Length Cumulative Distribution */
/* Function and the Annotate Dataset */
/*************************************************************/
title1 'Run Length Cumulative Distribution Function';
symbol1 color = white width = 2 interpol = join value = none
line = 1 ;
axis1 width = 1 label = ('Run Length');
axis2 width = 1 order = (0.0 to 1.0 by 0.1)
label = (justify = c a=-90 r=90 'Probability');
proc gplot data = cdf;
plot cdf*rl /
frame
annotate = anno
vminor = 0
hminor = 0
haxis = axis1
vaxis = axis2
vref = 0.2 to 0.8 by 0.2
lvref = 3
cvref = white
cframe = gray
nolegend;
run;
quit;
goptions reset=all;
