" - . & - # / 0 1 - " 2 - , . 34 . / " 5 678 9 - : 34 & ;
<3 31 = 1
> 4 - - & ?1
12
1
٢
) @A !"#
!$% &' (%)*
&+, - . / , %
0" &0"# 1 &#
. 3 #. 4! 5 %6 4+7 &#
80 9 0 : &# : 0 $ 0 ' %; <0 . .
= * / 7 > ?0 %; . 3 @ A # 8$ B C$
<0
% .
&+, . D @ A &$ E 3 F A #
4 # %A
. / % # E!8G H 0 ' %; %I 3 #
& %;
!J0 ! 4 <0 , 4<6 , 4<6 D . &' , K
.
H 0 ' %;
% / 8 H !J0 &# % L% M N <0 # O &+, <!' &)? 0# !A .
P ' 3 #
% =C, O &+, H &+, <!' &)?
. % $ 0 ' . % 7 &# &+, 0 H Q RS
! T <0 .
% , K &+, @ A # !J0 ! & %; / % L # U: H 0 ' %; 4 .
+;7
4 &0 T # % 4 V %0 !M <0 # !A H 0 ' %;
. %; / , 9 0 &" E
%; # !A $ 0 '
& %; / % # $ 0 ' &W %; 0 # +;7 4 , 4# P !
.
- B 4
. E!8G H 0 ' %;
&+, - H H ! 4 V %0 % L% M N
&0 1!;M G5 .
Designing Adaptive Control Charts Based on Variable Sampling Intervals for Monitoring Two Stages Processes With Poisson Response
M. Esmaeeli, A. Amiri
ABSTRACT
Nowadays, most of the products are the result of a series of dependent process stages. Due to cascade property in these processes, using common control charts leads to misleading results. Hence, cause selecting control charts (CSC) are used widely to control these processes. In this paper monitoring a two stages process with Poisson responses is considered and two adaptive control charts based on variable sampling intervals (VSI) with warning limits and double warning limits are designed. In the proposed control charts, first, the quality characteristic of the second stage is transformed to normal variable by the NORTA inverse transformation method. This leads to removing the effect of the quality characteristic in the first stage on the second stage. Then, the standardized normal observations in both stages are used as statistics. Finally, using warning limits and double warning limits separate control charts based on variable sampling intervals for monitoring both stages are designed. Performance of the proposed control charts is evaluated though the adjusted average time to signal (AATS) criterion. Comparing the results of the proposed control charts whit control charts designed based on the fixed sampling intervals (FSI) reveals the better performance of the proposed control charts.
KEYWORDS
Adaptive sampling interval; Cascade processes; NORTA inverse method; Generalized linear models.
;+7 Q ! %X7 0 1
3<+
: 51212065
^ 021
1 ! . 5
_' Q(%)*
0M !B 0M 17
# &+, - . / , .
& %;
&' 4 ` 0 % . . 3
0" a+0? Q MGT . H%)* &# ' %87 M0
/ 8
% . b !!J &' 4 . . 3
M# /, I <!' &)? !!J : &# : 0# /, H%)* <!' &)?
% . . b 3 &#
&+, - #. 4! 5
% . /, 4!<!' O 7 . 3
; !"< A /, . +;7 aMc &# %
# %# +8T /, H%)* >%+G 4!<!' 4 3 ; &' - ' .
&+, - . 0! 5 3! - %6 /!$ #
+;7 d%)5 80 ! < &# : Q % 9 $ 0 ' %; <0 . a+0? /,
% . e 5 &+, f!? O 7
H 0 ' %; V % H 0 ' 3 ; !B . & % 3 @ A # %; 3 <0 gh;7 U! i !+0
%;
VE i !+0 f!? H 0 ' e 5 4$ , %
e 5 &+, 4 % H 0 '
.
&+, - . @ A # 3 0 #
H # 3!$ #. 4! 5 %6 # 1980
H RS &W i ` V %
1984
= * / 7 > ?0 H 0 ' %; % 7 # . )1
(CSC
%; M .
&M % %; 3 EE* %5 U! M#
.
%; '%' L '%$
EWMA CSC @ A ; % o *0 3! ! H 0 ' %; V % ;!T # E &' ' &W
% .
%; i ` H 0 ' , !!J H ;7 # H M 6CSC
' .
4<! a ' Q % H 0 ' %; aMc
!8G H 0 ' %; N 0" &# : D-%' E
. 8$ B &' %; 3
& %; / %; 0 A . p% &# &0"# Q % 0 %; &# 48" 0 # +;7 )
(n
& %; / %
! ) (h
H 0 ' , q c )
(k !!J -
% . 3!!M $ 0 ' %; /8T &+, 1 . 4!MT% &# &0"# 0 A 3
% .
#
. &' H _ H 8 &# . $ ;0, 4<! # % 1 , 3 &# D U $ $ 0 ' , /5
.
. . 30 H 0 ' e 5 4 - &
.
& %; C$
C5 . 0 %' &+ M#
% ) &# &0"#
E!8G .(
&W E!8G H 0 ' %; . , K ; $ # 3!$ # .
' %; . &+ # H 0
!J0 ! & %;
) VSI2 ( 0 I .
0 '
!J0 & %; # H 0 ' %; D )
(VSS . H 0 ' Q <0 & %; &' ' &W <0
%#
. % i ' <0 &+, - . @ A # E!8G H 0 ' %; O% < # 3!$
. & %; & .
%; <0 &+, . D 3! ! @ A # a+0?
. . #
H ! 4 V %0 f5 #
3) (AATS
&W %; &' 4# P & %; # &# %; &# 48" 0 # +;7
.
& %; &+ & <0 # U! % i %; a+0? !
&+, . D 3! ! @ A # VSI
%; , K .
s .
&+, /E0" t < D p%T &+, &' ' %6
. U; . &' $ 0 ' %; D % i 3! u;
& %; &+ & %;
&+, . D @ A # 0" !J0 ! ' &W
.
% o *0 3! ! %; % i U! &
!A ; .
4 ! # C v "$ %
& %; &+ 3! u; & %; q!' a+0?
$ 0 ' %; a+0? !
VSSI4 ' &W .
! & %; &+ & %; . &' %; &W , # E!8G H 0 ' %; ! A
4 <0 Q <0 .
,
! & %; &+ &# w%# & %; &# w%#
# . ! T 4!MT% &# &0"#
<0 4 ! %; a+0? , % .
% <0 . @ A # .
1Cause selecting control chart
2Variable sampling intervals
3Adjusted average time to signal
4 Variable sample size and sampling interval
&)? %# H s t% % O ; A # O u! %# . D< ( vU6 <!'
F . H !B
&+, - , K
.
; !A H % %5 Q x . !"#
%# H s 3 # # '
s . #
; ; &# 0 .
&+, . D @ A &$ E 3 3 F A &'
!B . %; 1! < <0 # 0" H
&6% % E!8G H 0 '
# . s 3! u;
O : O . @ A %# f? 0 A &'
% . %I 3!; &#
& %; / % # E!8G H 0 ' %;
& , , D # !J0 ! <0
% . &' O &+, <!' &)?
4!M8 H !B M % % L% M N <0 # '
/ 8 H 
0 E . /! # % / ,
H
&0 1!;M G5 @ A2
. &# O @?# &' 4 Q % 3 # &$ E 05 b &05 A I &+y" z
%5 !# &+y"
. &W !A %; U! O% @?#
.
&:!0 !A N +;7 # q! &# U! &$ E !
&W 1: A O - @?#
.
2 ! . C D +8
/ 1 . D
&+, 4 ! ' %) &#
. E+ O% < 3 . #. 4! 5 %6 !!J &u - &' '
4 C %5 P U! O &+, <!' &)? # { H &+, <!' &)? % 0 A .
&+, <!' &)? I .
H ) (x
H %
f? H . 0 A &' # .
0 A &' 4 %A O &+, <!' &)? %
4 # H &+, <!' &)? G# # Q %)# &0 1!;M G5 H <0 # . 3! ! .
. q c &' 4 . # s
f? H U! V# # 3 .
E@F
1 . & # / G G 8 4F - 4 -
&)? &' Q M# /T , N /, 3!# %! &G# T # %I &# # H % &+, <!'
1
4 !;K /# T ' .
&G# : # %5 4 %5 ' N 3 F A !J0 %# H !B s # & . H ,
O H /, <!' &)? 3!# %!
H 4" # 4 D;' &0 1!;M G5
. F A !J0 &' # H 3
%
&+;6 %A ; H
# ' ' !A
.
# %)? V# # % F A !J0 % p% &# &0"#
w%# H &+, <!' &)? &# F A !J0 3! ! &' %6 . .
T # # / x &G# &$ E 3 K 8 3! -
4 <0 .
)
1 log = . (
t% &G#
R H &+, <!' &)? &# w%#
%! # D 3!!M q c # 4 GLM
.
3 !
- + 1 9 - : 34
%A % O &+, <!' &)? 'x &' %G ; . @ A 4$% 4 6 %#
1! ' / 8 H &# . 4" # .
<0 # O &+, 3! ! # | 0 E H &+, <!' &)? # %6% Q D # 0# q! 3 #
&G#
) 1 (
!A 1! ' .
& %; RA <!' &)? O &+, !
# &+, 3 % L% M N <0
H & % # &0 1!;M G5
O !!J
# 0 H % &0 O !!J !J0 &' 1!
.
&)? & 3;c O H &+, <!'
U! /E0"
% .
&)? 3 # # <!'
!J0 D H 0 ' %; # @ A . &+, .
1 NORTA inverse
2 Generalized linear models
y x
N &# 48" % L% M N 0 # +;7 O !!J
. !J0 / 8 %I &# N 3 L # M;: % # #y
F (y) 6 !J0 &#
ý
&G# 0 H % # )
2 ( <0
%
) 2 ý = ∅ (F (y)) (
&'
∅ (. ) 4 0 H M;: % # L% M !#
.
&G# <0 # )
2
́ (
%# %5 0 H % .
!J0
H % U! H &+,
# .
# O &+, <!' &)? Q M;: Q( ;0, & &# &6% # &8 * O &+, 3! ! # H &+, <!' &)? P L
O &+, <!' &)? P VE %# H &+, <!' &)? E /E0" . 4 # Q( ;0, E C$
# .
/E0" ' !A H % Q( ;0, 3 L% M E 3 # # H &+,
# .
3 ! 1
! : + & G 2 - , . 34 E 3/ " 9 : 34
O ' &' % &0 I $ 0 ' %; D &+, # /8T 4;"T z &+, . @ A %I &#
%5 &# d%)? H 0 ' , ,
# . Q 8 * %I &#
= =
= =
s
% .
3! ! # H % H &+, <!' &)?
4 . 3! ! 0 A { H &+, 3! ! 0<! &u - &#
+ :# 6 ' . 3!# w 8 !# %I &#
& %; / %
! ! %; $ 0 ' %; # 1 . 4!MT%
# 1!"E &EG - &# O H /, !A
% . / 2 # 1!"E 3 .
I : $− , '
I : (− , − ) ∪ ( , ) I): (− , )
I*: (−∞, − ) ∪ ( , +∞)
&EG O &+, . 1 H &+, . 1 &u - I
&+ M# & %; # . ! T %5 > ?0 )
.
&EG /, . I
&EG &+, . I
. ! T
& %; &+ % M#
U # M# ! %
&EG . &u - I
. ! T t
& %; &+
%# %5 M# ! .
&EG . /T ,
I*
Oh7 H 0 ' e 5 . . ! T
% . )> >
&' E!8G 4$ , ! & %; / % V %0 4 H 0 ' 4* . 4$ , ! & %; &+ # # # 4" #
% E!8G !B .
M : )
3 /( 0| = = 0) (
= 3× 5 ((567 89 57:; ℎ: :; =7; >76 6:?: | = = 0)
&' &+ 3
4 E!8G !B %; ! & %;
. , E . V# =%' !: d %5 <0 #
& % # w
> ?0
& %; 3!# V %0 &' % %; !
& %; &+ # #VSI 4$ , !
FSI2
% ] 7 .[
1 Least square
2 Fixed sampling interval
E@F 2 . 34 5 678 9 : 34 G ( -
) 4
= ∅ @−4B ± √16B − 16GH (
8G J
&'
&G#
) 4 ( E A B &G# C ) 5 (
% &8 * :
) 5 G = ) − 2 + (
B = − )+ 2 ∅ k + − 2t ∅
H = −$ 3 2∅ − 1 − )+ 4 ∅ − 4 ∅ '
3 ! 2
! . , : + & 2 - , . 34 E 3/ " 9 : 34 )1
(DWL
# H 0 ' %; &' %G ;
& %; / % @ A # !J0 , # , D !J0 !
<0 .
% . V# =%' !: d %5 <0 # H %; , )
4 ( ) 5 (
&!8 &+! %# O %;
%5 f?
. / 3 , # $ 0 ' %; D &
.
C f? 1"T &# %; a+0? €K
% . I : $ , − '
I : , ' ∪ − , − ' I): − , − ∪ , I*: − ,
IM: −∞, − ∪ , +∞
&u - Oh7 H 0 ' e 5 . # H ! %; /T ,
% . 3!!M 4 • &# N 3 ! & %; / % :
0 = NO OP OO
Q if T ∈ IV ∩ TX ∈ IY if T ∈ IV ∩ TX∈ IY
) if T ∈ IV)∩ TX∈ IY)
* if T ∈ IV ∩ TX∈ IY ∪ T ∈ IV ∩ TX∈ IY
M if T ∈ IV ∩ TX∈ IY)∪ T ∈ IV)∩ TX∈ IY
Z if T ∈ IV)∩ TX∈ IY ∪ T ∈ IV ∩ TX∈ IY)
[
) 6 > *> > M> Z> ) (
4
! - + 1 9 - : 34 @ 4> : "
&+, . &' 4 . # s I &+y"
4 # D . #. 4! 5 %#
. D /, O '
<!' &)? H % H &+, <!' &)? &' %# <!' &)?
# %A % O &+, .
# <0 #
&G# 'x V#
) 1 ( V8 H &+, <!' &)? &# O &+, 3! !
% .
&!8 # RS &+, <!' &)? E
E@F 3 . , : + & " 9 : 34
4 # t% &G# / , 3! ! # O .
. <0 # / ,
&G#
) 2 ( %; V % / 8 H &#
$ 0 '
%5 @ A !A
.
H &+, <!' &)? &+y" 3 )
(x 4!M8 D !M = * < 3! ! # H % '
. # 0 A E g ;c
V#
\3 q! &#\ 5 ./
1 ./
&0 I .
&!8 %I &#
H % L # H &+, <!' &)? E 0# .
!$% %'C 0 A #
% .
&G# <0 # O &+, <!' &)? # 3! ! E RS )
1 ( &)? E &8 *
# f? E &# O &+, <!' !$% %A % L # ) Q %)# 3! !
% . RS !$%
&G# V % O &+, # )
2 (
%5 / 8 0 H % &#
.
M O &+, # &0 O !!J Q E %# H !;K 4 6 5000
&!8 H % . <0 #
q ! U O 14
/ / , 9 0 &' # 4
% .
%# H !;K H%), RA
% %# " O &+, &0 O !!J H &+, Q # .
4" #
& % # $ 0 ' %; 0 A
& %; &+ c ! &$ 8 H%K V %0 &' % , K 4 V %0 . 8 &# !
& %; &+ &$ 8 H%K V %0 # q! &# E!8G 4$ , H ! 4 V %0 !
% # # E!8G !B 4$ , H ! .
!A %; +;7 # %I &#
48P . # H ! 4 V %0 H ;7 I . 4<! p%
.
, %; 3 +;7 &' % 3!!M & % # !A %; 0 A %I 3 # # { . 0<! ƒ! &' 0$
% " E!8G !B 4$ , .
% # # 1 # E!8G !B E!8G 4$ , &$ 8 H%K V %0 f5 q! 3 # .
M $ 0 ' %; 0 A k
w w1 w2
# # # q! 0#
: 3 81 / 0 4 / 0 5 / 1 .
4
! 1 ! : 1 : L F ]^
; !!J %; 0 A O ' ƒ! &' 0$ , $ 0 ' %; !A H 0 ' %; +;7 &" E &# 4;"T 3 '
0 A a+0? 4<! 4*
\3 &05 A
% . &' %; +;7 Q 8M#
\3 &#
\3+ _ !!J
# % ' T
4 %5
. H 6 9 0 &' %G ; 1
! & %; &+ # $ 0 ' %; # &" E !A H 0 ' %; % &I,h
4<! 4# P
' 0 %' M 0 !# 47 # . {
. %; +;7 &" E 3;c
!A 4 , D # $ 0 ' %; &# 48" , # $ 0 ' %; +;7 # ' ,
. / , 9 0 3 5000
#
# &!8 .
4
! 2 ! : 1 : L F ]`
4 I E!8G !B %; # !A %; +;7 &" E 9 0 4;"T 3 .
u - 0 A 0<! &
. { \
&# %'C 0 A 4 %5 !!J \ +
.
H 6 9 0
2 !A %; +;7 &' 4 q+G 3 !#
E@F 4 . / M / 28 - ) - " 9 <3 (#
0 !# 47 # 4<! , # $ 0 ' %; 3! u; 4 0 # E!8G !B %; # &" E , D # $ 0 ' %; &# 48"
'
. 9 N
1 : L F LO8 -( . 6F P ]^
H ! 4 V %0
4<!
E!8G !B %;
) (FSI E!8G %;
) (VSI , # E!8G %;
&
) (DWL
0 1183
1181 1182
1 / 0 5
/ 867 5
/ 843 835
2 / 0 628
5 / 569 5
/ 560
25 / 0 525
475 1
/ 470
3 / 0 434
379 374
4 / 0 291
5 / 258 5
/ 253
5 / 0 195
5 / 161 151
7 / 0 89
69 66
1 5
/ 32 22
35 / 20
9 N 2 : L F LO8 -( . 6F P ]`
H ! 4 V %0
4<!
E!8G !B %;
) (FSI E!8G %;
) (VSI , # E!8G %;
&
) (DWL
0 1183
1180 5
/ 1181
05 / 0 5
/ 1161 5
/ 1129 1120
1 / 0 5
/ 1156 1103
5 / 1094
2 / 0 884
892 884
25 / 0 760
738 5
/ 720
5 / 0 5
/ 238 5
/ 229 226
75 / 0 100
5 / 93 91
1 57
56 55
5
! .%
- ,
/ % # H 0 ' %; &$ E 3
& %;
&+, . D @ A 4 6 , , D # !J0 ! F A #
&W H !B .
L% M N &0 1!;M G5 % $ q!' / , O &+, <!' &)? @ A 4 6 H 0 ' %; .
%
# .
!A H 0 ' %; +;7 &" E E!8G !B H 0 ' %; # . H ! 4 V %0 f5 V %
# $ 0 ' %; 3! u; %5 # 0 !# 47 . { 4<! !A %; &' . 4 , , D # $ 0 ' %; # &" E , 4 %5 # 0 # +;7
.
N
Costa A.F.B. (1994) “ xb Chart with variable sample size”. Journal of Quality Technology. Vol. 26, No. 3, pp.
155–163.
Faraz A., Parsian A. (2006) “Hotelling's T2 control chart with double warning lines”. Statistical Papers. Vol.
47, No. 4, pp. 569-593.
Lucas. J.M., Saccucci , M. (1990) “Exponentially weighted moving average control schemes: properties and enhancements” Technometrics. Vol. 32, No. 1, pp. 1–29.
Niaki, S.T.A., Abbasi, B. (2009) “Monitoring multi-attribute processes based on NORTA inverse transformed vectors”. Communication in Statistics – Theory and Methods. Vol. 38, No. 7, pp. 964-979.
Noorossana R., Ashkezary M.S. (2011) ”Monitoring Two Dependent Process Steps Using Special Variable Sample Sizes and Sampling Intervals Cause-Selecting Control Charts”. Quality and Reliability Engineering International. Vol. 28, No.4, pp. 437-453.
Reynolds, MR. Jr., Amin, RW., Arnold, JC., Nachlas JA. (1988) “ xc Charts with variable sampling intervals”. Technometric. Vol.30. No. 2, pp.181–192.
Wade, M. , Woodall, W. (1993) “A review and analysis of cause-selecting control charts” Journalof Quality Technology. Vol. 25, No. 3, pp. 161-169.
Yang S.F., Su H.C. (2006) “Controlling-dependent process steps using variable sample size control charts”.
Applied Stochastic Models in Business and Industry. Vol. 22, No. 5-6, pp.503-517.
Yang S.F., Su H.C. (2007) “Adaptive control schemes for two dependent process steps”. Journal of Loss Prevention in the Process Industries. Vol. 20, No. 1, pp. 15-25.
Yang S.F., Su H.C. (2007) “Adaptive sampling interval cause-selecting control charts”. The International Journal of Advanced Manufacturing Technology. Vol. 31, No. 11, pp. 1169-1180.
Yang S.F., Yu Y.N. (2007) “Using VSI EWMA charts to monitor dependent steps with incorrect adjustment”. Expert Systems with Applications. Vol. 36, No. 1, pp. 442–454.
Zhang, G. (1984) “A new type of control charts and theory of diagnosis with control charts” WorldQuality Congress Transactions American Society for Quality Control, pp. 175-185.
! . p "7 .
O
. ) 1391 .(
&+, . @ A
&0 1!;M G5 % $ ! # # 6 5 !J0 #
% L% M N 4!<!' 4 &
+6 2 ; 1 &*<
55
^ 61 .