1 risk assessment via monte carlo simulation: tolerances versus statistics b. ross barmish ece...
TRANSCRIPT
![Page 1: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/1.jpg)
1
Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics
B. Ross BarmishECE Department
University of Wisconsin, MadisonMadison, WI 53706
![Page 2: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/2.jpg)
2
• A. C. Antoniades, UC Berkeley
• A. Ganesan, UC Berkeley
• C. M. Lagoa, Penn State University
• H. Kettani, University of Wisconsin/Alabama
• M. L. Muhler, DLR, Oberfaffenhofen
• B. T. Polyak, Moscow Control Sciences
• P. S. Shcherbakov, Moscow Control Sciences
• S. R. Ross, University of Wisconsin/Berkeley
• R. Tempo, CENS/CNR, Italy
COLLABORATORS
![Page 3: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/3.jpg)
3
Overview
• Motivation
• The New Monte Carlo Method
• Truncation Principle
• Surprising Results
• Conclusion
![Page 4: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/4.jpg)
4
Monte Carlo Simulation
• Used Extensively to Assess System Safety
• Uncertain Parameters with Tolerances
• Generate “Thousands” of Sample Realizations
• Determine Range of Outcomes, Averages, Probabilities etc.
• How to Initialize the Random Number Generator
• High Sensitivity to Choice of Distribution
• Unduly Optimistic Risk Assessment
![Page 5: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/5.jpg)
5
It's Arithmetic Time
• Consider
1 + 2 + 3 + 4 + 5 + + 20 = 210
• Data and Parameter Errors
• Classical Error Accumulation Issues
(1 § 1) + (2 § 1) + (3 § 1) + (20 § 1) = 210 § 20
![Page 6: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/6.jpg)
6
Two Results
Actual Result Alternative Result
190 · SUM · 230
190 195 200 205 210 215 220 225 230
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
x
Student Version of MATLAB
![Page 7: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/7.jpg)
7
Circuit Example
+
_
10
10
10 5
R
R
R
R
2
3
I1I2
I3
I4
x
1
1 VVout+ _
Output Voltage
• Range of Gain Versus Distribution
![Page 8: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/8.jpg)
8
More Generally
k1
b
k2
2b
k3
3b
k4
4
m =12
m =11 m =13
m =14
b1
y
F
• Desired Simulation
![Page 9: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/9.jpg)
9
Two Approaches
• Interval of Gain
• Monte Carlo
0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.90
100
200
300
400
500
600
700
800
900
1000
Student Version of MATLAB
• How to Generate Samples?
Gain Histogram: 20,000 Uniform Samples
![Page 10: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/10.jpg)
10
Key Issue
What probability distribution should be used? Conclusions are often sensitive to choice of distribution.
• Intractability of Trial and Error
• Combinatoric Explosion Issue
![Page 11: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/11.jpg)
11
Key Idea in New Research
Classical Monte Carlo New Monte Carlo
RandomNumber
Generator
F
Samples of q
System Model
Samples of F(q)
Statistics of q
RandomNumber
Generator
F
Samples of q
System Model
Samples of F(q)
Statistics of q
NewTheory
Bounds on q
![Page 12: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/12.jpg)
12
The Central Issue
What probability distribution to use?
![Page 13: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/13.jpg)
13
Manufacturing Motivation
![Page 14: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/14.jpg)
14
Distributions for Capacitor
-20 -15 -10 -5 0 5 10 15 200
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
![Page 15: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/15.jpg)
15
Interpretation
![Page 16: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/16.jpg)
16
Interpretation (cont.)
RandomNumber
Generator
F
Samples of q
System Model
Samples of F(q)
Statistics of q
NewTheory
Bounds on q
![Page 17: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/17.jpg)
17
Truncation Principle
r1 t1 r1t1 t2 t 2 r2 r2x1 x2
f x1 1( ) f x2 2( )
![Page 18: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/18.jpg)
18
Example 1: Ladder Network
...
...
+
-
VoutVin
R1
R3
R2
R4
R6
R5
Rn-2
Rn
Rn-1
• Study Expected Gain
![Page 19: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/19.jpg)
19
Illustration for Ladder Network
![Page 20: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/20.jpg)
20
Example 2: RLC Circuit
Consider the RLC circuit
1
L
I
+
-
CC R V1 2 021R
![Page 21: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/21.jpg)
21
Solution Summary For RLC
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
x 10-7
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
x 10-6
t1
t2
![Page 22: 1 Risk Assessment Via Monte Carlo Simulation: Tolerances Versus Statistics B. Ross Barmish ECE Department University of Wisconsin, Madison Madison, WI](https://reader030.vdocuments.us/reader030/viewer/2022032414/56649ee65503460f94bf6b0c/html5/thumbnails/22.jpg)
22
Current and Further Research
• The Optimal Truncation Problem
• Exploitation of Structure
• Correlated Parameters
• New Application Areas; e.g., Cash Flows