10.02.04 1
Multivariate Statistical Process Multivariate Statistical Process Control and OptimizationControl and OptimizationAlexey Pomerantsev & Oxana Rodionova
Semenov Institute of Chemical PhysicsRussian Chemometrics Society
© Chris Marks
10.02.04 2
AgendaAgenda
1. Introduction
2. SPC
3. MSPC
4. Passive optimization (E-MSPC)
5. Active optimization (MSPO)
6. Conclusions
10.02.04 3
Statistical Process Control (SPC)Statistical Process Control (SPC)
SPC Objective
To monitor the performance of the process
SPC Method
Conventional statistical methods
SPC Approach
To plot univariate chart in order to monitor key process variables
SPC Concept
To study historical data representing good past process behaviour
10.02.04 4
Historical Process Data (Chemical Historical Process Data (Chemical Reactor)Reactor)
X17
9.74E-031.01E-02
-1.43E-039.07E-035.78E-03
-9.49E-04-6.79E-03-3.42E-03-9.86E-034.18E-03
-4.84E-039.44E-03
-4.99E-03-6.81E-031.23E-039.90E-033.65E-03
-6.78E-03s54 6.61E-02 -5.40E-01 7.19E-03 -2.85E-01 -5.19E-04 -5.78E-01 1.81E-04 -2.67E-04 -6.23E-05
…
…
X1 X2 X3 X4 X5 X6 X7 X8 X9
s1 -1.19E-01 7.28E-01 -2.15E-02 5.22E-01 7.06E-04 7.32E-01 3.10E-04 -6.13E-04 -5.92E-05s2 -1.37E-01 7.28E-01 -2.89E-02 6.08E-01 7.09E-04 7.02E-01 6.58E-04 -1.22E-03 -1.49E-04s3 2.51E-02 -9.15E-02 6.73E-03 -1.13E-01 -9.07E-05 -7.58E-02 -2.29E-04 4.10E-04 5.65E-05s4 -1.14E-01 6.70E-01 -2.18E-02 5.04E-01 6.50E-04 6.65E-01 3.83E-04 -7.34E-04 -7.96E-05s5 -7.93E-02 4.14E-01 -1.69E-02 3.51E-01 4.04E-04 3.98E-01 3.96E-04 -7.35E-04 -9.05E-05s6 1.51E-02 -6.38E-02 3.74E-03 -6.75E-02 -6.28E-05 -5.67E-02 -1.15E-04 2.07E-04 2.78E-05s7 7.44E-02 -5.24E-01 1.11E-02 -3.24E-01 -5.06E-04 -5.45E-01 -1.73E-05 7.92E-05 -1.07E-05s8 3.65E-02 -2.66E-01 5.12E-03 -1.59E-01 -2.56E-04 -2.78E-01 1.43E-05 -3.95E-07 -1.14E-05s9 1.36E-01 -7.06E-01 2.89E-02 -6.01E-01 -6.88E-04 -6.77E-01 -6.83E-04 1.26E-03 1.56E-04s10 -2.74E-02 3.60E-01 1.82E-03 1.12E-01 3.42E-04 4.12E-01 -4.31E-04 7.24E-04 1.22E-04s11 7.47E-02 -3.31E-01 1.80E-02 -3.34E-01 -3.25E-04 -2.99E-01 -5.30E-04 9.62E-04 1.28E-04s12 -1.17E-01 7.02E-01 -2.16E-02 5.13E-01 6.81E-04 7.03E-01 3.40E-04 -6.63E-04 -6.76E-05s13 1.06E-01 -2.82E-01 3.23E-02 -4.82E-01 -2.85E-04 -1.87E-01 -1.25E-03 2.21E-03 3.14E-04s14 7.39E-02 -5.28E-01 1.07E-02 -3.21E-01 -5.09E-04 -5.50E-01 2.49E-06 4.48E-05 -1.59E-05s15 -9.87E-03 1.02E-01 -3.21E-04 4.17E-02 9.75E-05 1.13E-01 -8.29E-05 1.36E-04 2.44E-05s16 -1.06E-01 7.68E-01 -1.52E-02 4.62E-01 7.41E-04 8.03E-01 -2.54E-05 -2.68E-05 2.88E-05s17 -4.76E-02 2.66E-01 -9.52E-03 2.10E-01 2.59E-04 2.61E-01 1.92E-04 -3.61E-04 -4.19E-05
P
rod
uct
ion
cyc
les
s1,
s2,
...
,s5
4
Key process variables (sensors) X1, X2, ... , X17
10.02.04 5
Shewart Charts (1931)Shewart Charts (1931)
X1 Normal
X2 Normal
X1 Control
X1 Control
-0.5
-0.4
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
s1 s3 s5 s7 s9
s11
s13
s15
s17
s19
s21
s23
s25
s27
s29
s31
s33
s35
s37
s39
s41
s43
s46
s49
s51
s54
s56
Cycles (time)
X1
Se
ns
or
X2 Normal
X2 Normal
X2 Control
X2 Control
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
s1 s3 s5 s7 s9
s11
s13
s15
s17
s19
s21
s23
s25
s27
s29
s31
s33
s35
s37
s39
s41
s43
s46
s49
s51
s54
s56
Cycles (time)
X2
Se
ns
or
X1 Normal
X2 Normal
X1 Control
X1 Control
X2 Normal
X2 Normal
X2 Control
X2 Control
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
s1 s3 s5 s7 s9
s11
s13
s15
s17
s19
s21
s23
s25
s27
s29
s31
s33
s35
s37
s39
s41
s43
s46
s49
s51
s54
s56
Cycles (time)
X1
& X
2 S
en
so
rs
Normal
Normal
Control
Control
-1.2
-0.8
-0.4
0.0
0.4
0.8
1.2
s1 s3 s5 s7 s9
s11
s13
s15
s17
s19
s21
s23
s25
s27
s29
s31
s33
s35
s37
s39
s41
s43
s46
s49
s51
s54
s56
Cycles (time)
All
Se
ns
ors
10.02.04 6
Panel Process Control (just a game)Panel Process Control (just a game)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
0.31 -0.4 0.26 -0.3 -0.3 -0.4 -0.1 0.13 0.08 -0.3 -0.3 #### -0.3 -0.3 #### -0.1 -0.4
Time till the end of shift: 7:59:10
0.31
-0.36
0.26
-0.31-0.27-0.37
-0.09
0.13 0.08
-0.35 -0.35
0.01
-0.28-0.35
-0.01
-0.06
-0.38
OnOff Exit
-3
-2
-1
0
1
2
3
-4 -3 -2 -1 0 1 2 3 4
PC1
PC
2
10.02.04 7
Multivariate Statistical Process Control Multivariate Statistical Process Control (MSPC)(MSPC)
MSPC Objective
To monitor the performance of the process
MSPC Method
Projection methods of Multivariate Data Analysis (PCA, PCR, PLS)
MSPC Approach
To plot multivariate score plots to monitor the process behavior
MSPC Concept
To study historical data representing good past process behavior
10.02.04 8
Projection MethodsProjection Methods
Initial Data
Data Plane
Data Center
PCs
Data Projections
10.02.04 9
Low Dimensional PresentationLow Dimensional Presentation
10.02.04 10
Loadings
X1
X2
X3
X4
X5
X6
X7
X8X9
X10X11
X12
X13
X14
X15X16
X17
-0.3
0
0.3
0.6
-0.4 -0.2 0 0.2 0.4
PC1
PC2
MSPC Charts (Chemical Reactor)MSPC Charts (Chemical Reactor)
Scores
s53
s52
s51s50
s49
s48s47
s46s45 s44
s43
s42s41
s40
s39 s38 s37
s36s35s34
s33
s32s31
s30
s29s28
s27 s26
s25
s24
s23
s22
s21s20
s19s18
s17
s16
s15
s14
s13
s12
s11
s10
s9s8
s7
s6 s5
s4
s3 s2
s1
-3
-2
-1
0
1
2
3
-4 -3 -2 -1 0 1 2 3 4
PC1
PC2
Samples Key Variables
10.02.04 11
Panel Process Control (not just a game)Panel Process Control (not just a game)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
0.31 -0.4 0.26 -0.3 -0.3 -0.4 -0.1 0.13 0.08 -0.3 -0.3 #### -0.3 -0.3 #### -0.1 -0.4
Time till the end of shift: 7:59:10
0.31
-0.36
0.26
-0.31-0.27-0.37
-0.09
0.13 0.08
-0.35 -0.35
0.01
-0.28-0.35
-0.01
-0.06
-0.38
OnOff Exit
-3
-2
-1
0
1
2
3
-4 -3 -2 -1 0 1 2 3 4
PC1
PC
2
10.02.04 12
Cruise Ship Control (by Kim Esbensen)Cruise Ship Control (by Kim Esbensen)
10.02.04 13
Key Process VariablesKey Process Variables
10.02.04 14
PLS1 Prediction of Fuel Consumption PLS1 Prediction of Fuel Consumption
Scores
-3
-2
-1
0
1
2
3
-4 -3 -2 -1 0 1 2 3 4
PC1
PC2
Samples Predicted vs. Measured
Slope: 0.95Offset: 0.02Correl: 0.98RMSEP: 0.23SEP: 0.24Bias: -0.005
-3
-2
-1
0
1
2
3
-3 -2 -1 0 1 2 3
Measured Fuel
Pre
dic
ted
Fu
el
Weather conditionsX1, X2, X3, X4
PLS1Fuel Consumption Y
Cap’s setupX5, X6, X7
10.02.04 15
Passive OptimizationPassive Optimization
Weather conditions
Order!!! Prediction !Order!!!X5, X6, X7
Prediction !
Prediction ?
Fuck
Captain StudentComputer
4242
X1, X2, X3, X4X5, X6, X7
censored
10.02.04 16
Active OptimizationActive Optimization
Weather conditions
Advice!!!
Censored
Order?
CaptainStudent Computer
X1, X2, X3, X4
X5 X6, X7
Optimal
X5, X6, X7
42
10.02.04 17
In Hard Thinking about PC and PCsIn Hard Thinking about PC and PCs
Forty twocensored
10.02.04 18
Multivariate Statistical Process Multivariate Statistical Process Optimization (MSPO)Optimization (MSPO)
MSPO Objective
To optimize the performance of the process (product quality)
MSPO Methods
Projection methods and Simple Interval Calculation (SIC) method
MSPO Approach
To plot predicted quality at each process stage
MSPO Concept
To study historical data representing good past process behavior
10.02.04 19
Technological Scheme. Multistage Process Technological Scheme. Multistage Process
S1 S2 S3
M1 M2 M3 CM1 CM2 CM3
W1 W2 W3 CW1 CW2 CW3
WR1 WR2
MR1 MR2
S
W CW
M CM PA1 A2 A3 A4 A5 A6
S1 S2 S3
M1 M2 M3 CM1 CM2 CM3
W1 W2 W3 CW1 CW2 CW3
WR1 WR2
MR1 MR2
S
W CW
M CM PA1 A2 A3 A4 A5 A6
I6
II8
III11
IV14
V16
VI19
VII25
10.02.04 20
Historical Process DataHistorical Process Data
X preprocessing Y preprocessing
S1
S2
S3
W1
W2
W3
WR
1
WR
2
CW
1
CW
2
CW
3
M1
M2
M3
MR
1
MR
2
CM
1
CM
2
CM
3
A1
A2
A3
A4
A5
A6 Y
Tra
inin
g
Se
t (1
02
)
Y
Tes
t S
et
(52
)
Y
XV XVI XVII
XI XII XIII XIV XV XVI XVII
XI XII XIII XIV
10.02.04 21
Quality Data (Standardized Y Set)Quality Data (Standardized Y Set)Training Set Samples
Lowest Quality Y=-1
Highest Quality Y=+1
-1.2
-0.8
-0.4
0.0
0.4
0.8
1.2
1 21 41 61 81 101
Y
Test Set Samples
Lowest Quality Y=-1
Highest Quality Y=+1
-1.2
-0.8
-0.4
0.0
0.4
0.8
1.2
1 11 21 31 41 51
Y
10.02.04 22
General PLS ModelGeneral PLS Model
0
0.1
0.2
0.3
PC_0 PC_2 PC_4 PC_6 PC_8
PCs
RMSE Calbration
Validation
S1
S2
S3
W1
W2
W3
WR
1
WR
2
CW
1
CW
2
CW
3
M1
M2
M3
MR
1
MR
2
CM
1
CM
2
CM
3
A1
A2
A3
A4
A5
A6 Y
Y
YXTEST
XTRAINING
^
^PLS
6 PCs
10.02.04 23
SIC Prediction. All Test SamplesSIC Prediction. All Test SamplesSIC Prediction
1
2
3
4
5
6
7
8
9
1011
12 15
17
18
19
20
21
2223
25
26
27
28
29
30
3132
33
34
35
38
40
41
42
44
45
46
47
48
49
5051
13
16
24
43
36
52
3739
14
-1.0
-0.5
0.0
0.5
1.0
Test Samples
Y SIC PLS1 Test
Status plot
5251
50
49
48
4746
454443
42
41
40
393837
36
35
34
33 32
31 30
29
2827
26
25
24
23
22
21
20
19
18
17
16
1514
13 1211
10
9
8
7
6
5
4
3
21
-1.0
-0.5
0.0
0.5
1.0
SIC-Leverage
SIC
-Re
sid
ua
l
Status plot
12
3
4
5
6
7
8
9
10
111213
1415
16
17
18
19
20
21
22
23
24
25
26
2728
29
3031
3233
3435
36
3738 39
40
41
42
43 4445
4647
48
49
50
5152
-1.0
-0.5
0.0
0.5
1.0
SIC-Leverage
SIC
-Re
sid
ua
l
SIC Prediction
1
2
3
4
5
6
7
8
9
1011
12 15
17
18
19
20
21
2223
25
26
27
28
29
30
3132
33
34
35
38
40
41
42
44
45
46
47
48
49
5051
14
3937
52
36
43
24
16
13
-1.0
-0.5
0.0
0.5
1.0
Test Samples
Y SIC PLS1 Test
10.02.04 24
SIC Prediction. Selected Test SamplesSIC Prediction. Selected Test Samples
Sample No Quality status SIC Status
1 Normal Insider
2 High Outsider
3 Normal Absolute outsider
4 Low Outsider
5 Normal Insider
SIC Prediction
5
4
31
2
-1.0
-0.5
0.0
0.5
1.0
Selected Test Samples
YObject Status plot
1.01
2
34
5
-1.0
-0.5
0.0
0.5
1.0
SIC-Leverage
SIC
-Re
sid
ua
l
Insiders
Outsiders
Outsiders
Ab
s. O
uts
ider
s
10.02.04 25
Passive Optimization in PracticePassive Optimization in Practice
Objective
To predict future process output being in the middle of the process
Method
Simple Interval Prediction
Approach
Expanding Multivariate Statistical Process Control (E-MSPC)
Concept
To study historical data representing good past process behaviour
10.02.04 26
Expanding MSPC, Sample 1Expanding MSPC, Sample 1S
1
S2
S3
W1
W2
W3 Y
Tra
inin
g
Se
t (1
02
)
Y
1 yxI XI
Sample 1, Normal Quality Insider
-1.0
-0.5
0.0
0.5
1.0 SIC PLS1
Y x1
S1
S2
S3
W1
W2
W3
WR
1
WR
2
Y
Tra
inin
g
Se
t (1
02
)
Y
1 yxI xII
XI XII Sample 1, Normal Quality Insider
-1.0
-0.5
0.0
0.5
1.0 SIC PLS1
Y x1
S1
S2
S3
W1
W2
W3
WR
1
WR
2
CW
1
CW
2
CW
3
Y
Tra
inin
g
Se
t (1
02
)
Y
1 y
xI xII xIII
XI XII XIII
Sample 1, Normal Quality Insider
-1.0
-0.5
0.0
0.5
1.0 SIC PLS1
Y x1
S1
S2
S3
W1
W2
W3
WR
1
WR
2
CW
1
CW
2
CW
3
M1
M2
M3 Y
Tra
inin
g
Se
t (1
02
)
Y
1 yxI xII xIII xIV XI XII XIII XIV
Sample 1, Normal Quality Insider
-1.0
-0.5
0.0
0.5
1.0 SIC PLS1
Y x1
S1
S2
S3
W1
W2
W3
WR
1
WR
2
CW
1
CW
2
CW
3
M1
M2
M3
MR
1
MR
2
Y
Tra
inin
g
Se
t (1
02
)
Y
1 yxVxI xII xIII xIV
XVXI XII XIII XIV Sample 1, Normal Quality Insider
-1.0
-0.5
0.0
0.5
1.0 SIC PLS1
Y x1
S1
S2
S3
W1
W2
W3
WR
1
WR
2
CW
1
CW
2
CW
3
M1
M2
M3
MR
1
MR
2
CM
1
CM
2
CM
3
Y
Tra
inin
g
Se
t (1
02
)
Y
1 yxV xVI xI xII xIII xIV
XV XVI XI XII XIII XIV
Sample 1, Normal Quality Insider
-1.0
-0.5
0.0
0.5
1.0 SIC PLS1
Y x1
S1
S2
S3
W1
W2
W3
WR
1
WR
2
CW
1
CW
2
CW
3
M1
M2
M3
MR
1
MR
2
CM
1
CM
2
CM
3
A1
A2
A3
A4
A5
A6 Y
Tra
inin
g
Se
t (1
02
)
Y
1 y
XV XVI XVIIXI XII XIII XIV
xV xVI xVIIxI xII xIII xIV
Sample 1, Normal Quality Insider
-1.0
-0.5
0.0
0.5
1.0 SIC PLS1
Y x1
10.02.04 27
Expanding MSPC , Samples 2 & 3Expanding MSPC , Samples 2 & 3Sample 2, High Quality, Outsider
-1.0
-0.5
0.0
0.5
1.0
x2 SIC
PLS1 Y
Sample 3, Normal Quality, Absolute Outsider
-1.0
-0.5
0.0
0.5
1.0
x3 SIC
PLS1 Y
10.02.04 28
Expanding MSPC , Samples 4 & 5Expanding MSPC , Samples 4 & 5
Sample 4, Low Quality, Outsider
-1.0
-0.5
0.0
0.5
1.0 x4 SIC
PLS1 Y
Sample 5, Normal Quality, Insider
-1.0
-0.5
0.0
0.5
1.0x5 SIC
PLS1 Y
10.02.04 29
Active Optimization in PracticeActive Optimization in Practice
Objective
To find corrections for each process stage that improve the future process output (product quality)
Method
Simple Interval Prediction and Status Classification
Approach
Multivariate Statistical Process Optimization (MSPO)
Concept
Corrections are admissible if they are similar to ones that sometimes happened in the historical data in the similar situation
10.02.04 30
Linear Optimization Linear Optimization
Linear function always reaches extremum at the border.
So, the main problem of linear optimization is not to find a
solution, but to restrict the area, where this solution should
be found.
x
y=a*x
x
y=a*x
x
y=a*x
x
y=a*x
10.02.04 31
Optimization ProblemOptimization Problem
Weather conditionsX1, X2, X3, X4
PLS1Fuel Consumption Y
Cap’s setupX5, X6, X7
Fixed variables Xfix
PLS1 Quality measure Y
Optimized Xopt
Y = X*a = Y0 + Xopt*a2, where Y0 = Xfix*a1 = Const
Model
For given Xfix and a1 to find Xopt that maxi(mini)mizes Y
Task
max (Y) = Y0 + max (Xopt)*a2, as all a > 0 (by factor)
Solution
10.02.04 32
Interval Prediction of Interval Prediction of XXoptopt
Borders
1 2 3 4 5
-1.2
-0.8
-0.4
0.0
0.4
0.8
1.2
Selected Test Samples
M1
S1
S2
S3
W1
W2
W3
WR
1
WR
2
CW
1
CW
2
CW
3
M1
M2
M3
X fix X opt
XIVXI XII XIII PLS2
PLS Prediction
54321
-1.2
-0.8
-0.4
0.0
0.4
0.8
1.2
Selected Test Samples
M1 PLS±2*RMSEP
1 2 3 4 5
-1.2
-0.8
-0.4
0.0
0.4
0.8
1.2
Selected Test Samples
M1 SIC Prediction
54321
-1.2
-0.8
-0.4
0.0
0.4
0.8
1.2
Selected Test Samples
M1 SIC Prediction
5
4
3
21
-0.4
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
Selected Test Samples
M1Xopt
10.02.04 33
Dubious Result of OptimizationDubious Result of Optimization
Optimized
III III IV V VI VII
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
Y x4 Opt
x4 Test
Optimized
VIIVIVIVIIII II
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
SIC PLS
Y x4 Opt
x4 Test
Optimized
III III IV V VI VII
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
SIC PLS
Y x4 Opt
x4 Test Limits
Predicted Xopt variables are out of model!
10.02.04 34
Adjustment with SIC Object Status Adjustment with SIC Object Status Concept
Corrections are admissible if they are similar to ones that sometimes happened in the historical data in the similar situation.
Optimal variables Xopt should be within the model !
Object Status plot
Insiders
1
2
3
4
5
-1.0
-0.5
0.0
0.5
1.0
SIC-Leverage
SIC
-Re
sid
ua
l
SIC Prediction
1 2 43 5
-0.4
-0.2
0.0
0.2
0.4
Selected Test Samples
M1 Object Status plot
Insiders
1
2
3
4
5
-1.0
-0.5
0.0
0.5
1.0
SIC-Leverage
SIC
-Re
sid
ua
l
SIC Prediction
1 2 43 5
-0.4
-0.2
0.0
0.2
0.4
Selected Test Samples
M1 Object Status plot
Insiders
1
2
3
4
5
-1.0
-0.5
0.0
0.5
1.0
SIC-Leverage
SIC
-Re
sid
ua
l
SIC Prediction
1 2 43 5
-0.4
-0.2
0.0
0.2
0.4
Selected Test Samples
M1 Object Status plot
Insiders
1
2
3
4
5
-1.0
-0.5
0.0
0.5
1.0
SIC-Leverage
SIC
-Re
sid
ua
l
SIC Prediction
1 2 43 5
-0.4
-0.2
0.0
0.2
0.4
Selected Test Samples
M1
10.02.04 35
Sample 1 Normal Quality InsiderSample 1 Normal Quality InsiderOptimized
VIIVIVIVIIII II-1.0
-0.5
0.0
0.5
1.0 SIC PLS
Y x1
Test III III IV V VI VII-1.0
-0.5
0.0
0.5
1.0 SIC PLS
Y x1
Object Status plot
1.05
43
2
1
-1.0
-0.5
0.0
0.5
1.0
SIC-Leverage
SIC
-Re
sid
ua
l
10.02.04 36
Sample 2 High Quality OutsiderSample 2 High Quality OutsiderOptimized
III III IV V VI VII-1.0
-0.5
0.0
0.5
1.0
SIC PLS
Y x2
Test VIIVIVIVIIII II-1.0
-0.5
0.0
0.5
1.0
SIC PLS
Y x2
Object Status plot
1.01
2
34
5
-1.0
-0.5
0.0
0.5
1.0
SIC-Leverage
SIC
-Re
sid
ua
l
10.02.04 37
Sample 3 Normal Quality Abs. OutsiderSample 3 Normal Quality Abs. OutsiderOptimized
III III IV V VI VII-1.0
-0.5
0.0
0.5
1.0
SIC PLS
Y x3
Test VIIVIVIVIIII II-1.0
-0.5
0.0
0.5
1.0
SIC PLS
Y x3
Object Status plot
1.01
2
34
5
-1.0
-0.5
0.0
0.5
1.0
SIC-Leverage
SIC
-Re
sid
ua
l
10.02.04 38
Sample 4 Low Quality OutsiderSample 4 Low Quality OutsiderOptimized
III III IV V VI VII
-1.0
-0.5
0.0
0.5
1.0 SIC PLS
Y x4
Test VIIVIVIVIIII II-1.0
-0.5
0.0
0.5
1.0 SIC PLS
Y x4
Object Status plot
1.05
43
2
1
-1.0
-0.5
0.0
0.5
1.0
SIC-Leverage
SIC
-Re
sid
ua
l
10.02.04 39
Sample 5 Normal Quality InsiderSample 5 Normal Quality InsiderOptimized
III III IV V VI VII-1.0
-0.5
0.0
0.5
1.0 SIC PLS
Y x5
Test VIIVIVIVIIII II-1.0
-0.5
0.0
0.5
1.0 SIC PLS
Y x5
Object Status plot
1.01
2
34
5
-1.0
-0.5
0.0
0.5
1.0
SIC-Leverage
SIC
-Re
sid
ua
l
10.02.04 40
Philosophy of MSPO. Food IndustryPhilosophy of MSPO. Food IndustryF
oo
d Q
ual
ity
Production Effectiveness
Restaurant qualityStandard (descriptive) control
Fast Food qualityISO-9000
Home-made qualityMSPO
Home-made qualityIntuitive (expert) control
10.02.04 41
ConclusionsConclusions
Thanks and ...
Bon Appetite!