lu decomposition and matrix inversion chapter 10

10
Copyright © 2006 The McGraw-Hill Companies, Inc. Permission required for reproduction or display. 1 LU Decomposition and Matrix Inversion Chapter 10 Credit: Prof. Lale Yurttas, Chemical Eng., Texas A&M University

Upload: wendi

Post on 06-Jan-2016

101 views

Category:

Documents


12 download

DESCRIPTION

LU Decomposition and Matrix Inversion Chapter 10. Credit: Prof. Lale Yurttas, Chemical Eng., Texas A&M University. 0. 0. Solve A . x = b (system of linear equations) Decompose A = L . U * L : Lower Triangular Matrix U : Upper Triangular Matrix. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: LU Decomposition  and  Matrix Inversion Chapter 10

Copyright © 2006 The McGraw-Hill Companies, Inc. Permission required for reproduction or display.

1

LU Decomposition and

Matrix Inversion

Chapter 10

Credit: Prof. Lale Yurttas, Chemical Eng., Texas A&M University

Page 2: LU Decomposition  and  Matrix Inversion Chapter 10

Copyright © 2006 The McGraw-Hill Companies, Inc. Permission required for reproduction or display.

2

Solve A . x = b (system of linear equations)

Decompose A = L . U

*

L : Lower Triangular Matrix U : Upper Triangular Matrix

00

Page 3: LU Decomposition  and  Matrix Inversion Chapter 10

Copyright © 2006 The McGraw-Hill Companies, Inc. Permission required for reproduction or display.

3

To solve [A]{x}={b}

[L][U]=[A] [L][U]{x}={b}

Consider [U]{x}={d}[L]{d}={b}

1. Solve [L]{d}={b} using forward substitution to get {d}

2. Use back substitution to solve [U]{x}={d} to get {x}

Both phases, (1) and (2), take O(n2) steps.

Page 4: LU Decomposition  and  Matrix Inversion Chapter 10

Copyright © 2006 The McGraw-Hill Companies, Inc. Permission required for reproduction or display.

4

333231

232221

131211

aaa

aaa

aaa

A

bxA bxUL

[ U ][ L ]

33

2322

131211

3231

21

00

0

1

01

001

u

uu

uuu

ll

l

Page 5: LU Decomposition  and  Matrix Inversion Chapter 10

Copyright © 2006 The McGraw-Hill Companies, Inc. Permission required for reproduction or display.

''

'

''

''

3

2

1

3

2

1

33

2322

131211

00

0

b

b

b

x

x

x

a

aa

aaa

5

3

2

1

3

2

1

333231

232221

131211

b

b

b

x

x

x

aaa

aaa

aaa

Gauss Elimination

bxA bxUL

[ U ]

1

01

001

3231

21

ll

lLCoefficients used during the elimination step

Page 6: LU Decomposition  and  Matrix Inversion Chapter 10

Copyright © 2006 The McGraw-Hill Companies, Inc. Permission required for reproduction or display.

6

''

''

33

2322

131211

3231

21

333231

232221

131211

00

0

1

01

001

a

aa

aaa

ll

l

aaa

aaa

aaa

A

[ L . U ]

?32

11

3131

11

2121

l

a

al

a

al

Page 7: LU Decomposition  and  Matrix Inversion Chapter 10

Copyright © 2006 The McGraw-Hill Companies, Inc. Permission required for reproduction or display.

7

300

140

55.21

134

012

001

20224

1192

55.21

Example: A = L . U

20224

1192

55.21Coefficients

l21 = -2/-1= 2

l31 = 4/-1= -4

0120

140

55.21

l32 = -12/4= -3

Gauss Elimination

[ U ]

[ L ]

1??4

012

001

L

134

012

001

L

0120

140

55.21

300

140

55.21

Page 8: LU Decomposition  and  Matrix Inversion Chapter 10

Copyright © 2006 The McGraw-Hill Companies, Inc. Permission required for reproduction or display.8

Example: A = L . U

55.21

1192

20224Coefficients

l21 = -2/4= -.5

l31 = -1/4= -.25

10 0

030

20224

120

030

20224Coefficients

l32 = -2/-3

Gauss Elimination with pivoting

[ U ]

pivoting

20224

1192

55.21

120

030

20224

pivoting

1??25.0

01

001

0.5L

1??

0125.0

001

0.5

1

0125.0

001

0.660.5

55.21

1192

20224

030

0

20224

12

Page 9: LU Decomposition  and  Matrix Inversion Chapter 10

Copyright © 2006 The McGraw-Hill Companies, Inc. Permission required for reproduction or display.

9

• Gauss Elimination can be used to decompose [A] into [L] and [U]. Therefore, it requires the same total FLOPs as for Gauss elimination:

In the order of (proportional to) N3 where N is the # of unknowns.

• lij values (the factors generated during the elimination step) can be stored in the lower part of the matrix to save storage. This can be done because these are converted to zeros anyway and unnecessary for the future operations.

• Provides efficient means to compute the matrix inverse

LU decomposition

Page 10: LU Decomposition  and  Matrix Inversion Chapter 10

Copyright © 2006 The McGraw-Hill Companies, Inc. Permission required for reproduction or display.

10

MATRIX INVERSE

A. A-1 = I

100

010

001

333231

232221

131211

333231

232221

131211

xxx

xxx

xxx

aaa

aaa

aaa

0

0

1

31

21

11

333231

232221

131211

x

x

x

aaa

aaa

aaa

Solve in n=3 major phases

1 2 3

0

1

0

32

22

12

333231

232221

131211

x

x

x

aaa

aaa

aaa

1

0

0

33

23

13

333231

232221

131211

x

x

x

aaa

aaa

aaa

Solve each one

using A=L.U method e.g.

0

0

1

31

21

11

x

x

x

LU

Solve Problem 10.6. Solution file is available on the web

Each solution takesO(n2) steps.

Therefore, The Total time = O(n3)