lecture02 - factorization methods

Upload: na2ry

Post on 29-May-2018

226 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/9/2019 Lecture02 - Factorization Methods

    1/23

    Solution of Linear System of Equations

    Lecture 2:

    Factorization Methods

    MTH2212 Computational Methods and Statistics

  • 8/9/2019 Lecture02 - Factorization Methods

    2/23

    Dr. M. HrairiDr. M. Hrairi MTH2212MTH2212 -- Computational Methods and StatisticsComputational Methods and Statistics 22

    Objectives

    Introduction

    LU Decomposition

    Computational complexity

    The Matrix Inverse

    Extending the Gaussian Elimination Process

  • 8/9/2019 Lecture02 - Factorization Methods

    3/23

    Dr. M. HrairiDr. M. Hrairi MTH2212MTH2212 -- Computational Methods and StatisticsComputational Methods and Statistics 33

    Introduction

    Provides an efficient way to compute matrix inverse by

    separating the time consuming elimination of the Matrix [A]

    from manipulations of the right-hand side {B}.

    Gauss elimination, in which the forward elimination

    comprises the bulk of the computational effort, can be

    implemented as an LU decomposition.

  • 8/9/2019 Lecture02 - Factorization Methods

    4/23

    Dr. M. HrairiDr. M. Hrairi MTH2212MTH2212 -- Computational Methods and StatisticsComputational Methods and Statistics 44

    LU Decomposition

    The matrix [A] for the linear system [A]{X}={B} is factorizedinto the product of two matrices [L] and [U] (L- lower triangularmatrix and U- upper triangular matrix)

    [L][U]=[A]

    [L][U]{X}={B}

    Similar to first phase ofGauss elimination, consider

    [U]{X}={D}

    [L]{D}={B}

    The solution can be obtained by1. First solve [L]{D}={B} to generate an intermediate vector{D}by

    forward substitution

    2. Then, solve [U]{X}={D} to get {X}by back substitution.

  • 8/9/2019 Lecture02 - Factorization Methods

    5/23

    Dr. M. HrairiDr. M. Hrairi MTH2212MTH2212 -- Computational Methods and StatisticsComputational Methods and Statistics 55

    LU Decomposition

  • 8/9/2019 Lecture02 - Factorization Methods

    6/23

    Dr. M. HrairiDr. M. Hrairi MTH2212MTH2212 -- Computational Methods and StatisticsComputational Methods and Statistics 66

    LU Decomposition

    In matrix form, this is written as

    How to obtain the triangular factorization?

    Use Gauss elimination and store the multipliers mij as the

    subdiagonal entries in [L]

    !

    -

    3

    2

    1

    3

    2

    1

    333231

    232221

    131211

    b

    b

    b

    x

    x

    x

    aaa

    aaa

    aaa

    -

    -

    !

    333231

    23

    2221

    131211

    100

    010

    001

    aaa

    aaa

    aaa

    A

  • 8/9/2019 Lecture02 - Factorization Methods

    7/23

    Dr. M. HrairiDr. M. Hrairi MTH2212MTH2212 -- Computational Methods and StatisticsComputational Methods and Statistics 77

    LU Decomposition

    The multipliers are

    The triangular factorization of matrix [A]

    A = [L] [U]

    -

    -

    !33

    "

    3''

    3

    33 0000

    00

    a

    aa

    aaa

    mm

    mA

    11

    21

    21

    a

    am !

    11

    1

    1

    a

    am !

    22'

    32'

    32a

    a!

  • 8/9/2019 Lecture02 - Factorization Methods

    8/23

    Dr. M. HrairiDr. M. Hrairi MTH2212MTH2212 -- Computational Methods and StatisticsComputational Methods and Statistics 88

    Example 1

    Use LU decomposition to solve:

    3x1 0.1x2 0.2x3 = 7.85

    0.1x1 + 7x2 0.3x3 = -19.3

    0.3x1 0.2x2 + 10x3 = 71.4

    use 6 significant figures in your computation.

  • 8/9/2019 Lecture02 - Factorization Methods

    9/23

    Dr. M. HrairiDr. M. Hrairi MTHMTH22122212 -- Computational Methods and StatisticsComputational Methods and Statistics 99

    Example 1 - Solution

    In matrix form

    The multipliers are

    -

    4.71

    3.19

    85.7

    10.03.0

    3.071.0

    .01.03

    3

    1

    x

    x

    x

    0333333.03

    1.021 !!m

    100000.03

    3.0

    31!!

    m

    0271300.000333.7

    19.0

    32

    m

  • 8/9/2019 Lecture02 - Factorization Methods

    10/23

    Dr. M. HrairiDr. M. Hrairi MTHMTH22122212 -- Computational Methods and StatisticsComputational Methods and Statistics 1010

    Example 1 - Solution

    The LU decomposition is

    The solution can be obtained by

    1. First solve [L]{D}={B} for{D}by forward substitution

    -

    -

    !0120.1000

    293333.000333

    .70

    2.01.03

    10271 00.0100000.0

    01

    0333333.0

    001

    A

    !

    -

    4.71

    3.19

    85.7

    10271300.0100000.0

    010333333.0

    001

    3

    2

    1

    d

    d

    d

  • 8/9/2019 Lecture02 - Factorization Methods

    11/23

    Dr. M. HrairiDr. M. Hrairi MTHMTH22122212 -- Computational Methods and StatisticsComputational Methods and Statistics 1111

    Example 1 - Solution

    Then, solve [U]{X}={D} to get {X}by back substitution.

    84.56.9.85..4.

    56.985...9

    85.

    !!

    !!

    !

    d

    d

    d

    !

    -

    0843.70

    5617.19

    85.7

    01 0.1000

    93333.000333.70

    .01.03

    3

    1

    x

    x

    x

    33/))00003.7(.0)5.(1.085.7(

    5.00333.7/))00003.7(93333.05617.19(

    00003.701 0.10/0843.70

    1

    3

    !!

    !!

    !!

    x

    x

    x

  • 8/9/2019 Lecture02 - Factorization Methods

    12/23

    Dr. M. HrairiDr. M. Hrairi MTHMTH22122212 -- Computational Methods and StatisticsComputational Methods and Statistics 1212

    The triangular factorization portion of[A]=[L][U] requires

    (N3-N)/3 multiplications and divisions

    (2N3-3N2+N)/6subtractions

    Finding the solution to [L][U]{X}={B} requires

    N2

    multiplications and divisions N2-Nsubtractions

    The bulk of the calculation lies in the triangularization portion.

    LU decomposition is usually chosen over Gauss elimination when thelinear system is to be solved many times, with the same [A] but withdifferent {B}.

    Saves computing time by separating time-consuming elimination step fromthe manipulations of the right hand side.

    Provides efficient means to compute the matrix inverse which provides ameans to test whether systems are ill-conditioned

    ComputationalComplexity

  • 8/9/2019 Lecture02 - Factorization Methods

    13/23

    Dr. M. HrairiDr. M. Hrairi MTHMTH22122212 -- Computational Methods and StatisticsComputational Methods and Statistics 1313

    Find matrix [A]-1, the inverse of[A], for which

    [A][A]-1 = [A]-1 [A]=[I]

    The inverse can be computed in a column-by-column

    fashion by generating solutions with unit vectors {B}constants.

    The solution of[L][U]{X}={B} with will be the firstcolumn of[A]-1

    The solution of[L][U]{X}={B} with will be the second

    column of[A]-1

    The solution of[L][U]{X}={B} with will be the thirdcolumn of[A]-1

    The MatrixInverse

    _ a

    !

    0

    0

    1

    B

    _ a

    !0

    1

    0

    B

    _ a

    !

    1

    0

    0

    B

  • 8/9/2019 Lecture02 - Factorization Methods

    14/23

    Dr. M. HrairiDr. M. Hrairi MTHMTH22122212 -- Computational Methods and StatisticsComputational Methods and Statistics 1414

    Example 2

    Use LU decomposition to determine the matrix inverse for

    the following system and use it to find the solution:

    3x1 0.1x2 0.2x3 = 7.85

    0.1x1 + 7x2 0.3x3 = -19.3

    0.3x1 0.2x2 + 10x3 = 71.4

    use 6 significant figures in your computation.

  • 8/9/2019 Lecture02 - Factorization Methods

    15/23

    Dr. M. HrairiDr. M. Hrairi MTHMTH22122212 -- Computational Methods and StatisticsComputational Methods and Statistics 1515

    Example 2- Solution

    In matrix form

    The triangular factorization of[A]

    ? A

    -

    !

    102.03.0

    3.071.0

    2.01.03

    A

    -

    !

    10271300.0100000.0010333333.0

    001

    L

    -

    01 0.100093333.000333.70

    .01.03

    U

  • 8/9/2019 Lecture02 - Factorization Methods

    16/23

    Dr. M. HrairiDr. M. Hrairi MTHMTH22122212 -- Computational Methods and StatisticsComputational Methods and Statistics 1616

    Example 2- Solution

    The first column of[A]-1

    _ a

    !

    !

    -

    1009.0

    03333.0

    1

    0

    0

    1

    10271300.0100000.0

    010333333.0

    001

    3

    2

    1

    d

    d

    d

    _ a

    -

    01008.0

    00518.0

    33249.0

    1009.0

    03333.0

    1

    0120.1000

    293333.000333.70

    2.01.03

    3

    2

    1

    X

  • 8/9/2019 Lecture02 - Factorization Methods

    17/23

    Dr. M. HrairiDr. M. Hrairi MTHMTH22122212 -- Computational Methods and StatisticsComputational Methods and Statistics 1717

    Example 2- Solution

    The second column of[A]-1

    _ a

    !

    !

    -

    02713.0

    1

    0

    0

    1

    0

    10271300.0100000.0

    010333333.0

    001

    3

    2

    1

    d

    d

    d

    _ a

    !

    !

    -

    00271.0

    142903.0

    004944.0

    02713.0

    1

    0

    0120.1000

    293333.000333.70

    2.01.03

    3

    2

    1

    X

  • 8/9/2019 Lecture02 - Factorization Methods

    18/23

    Dr. M. HrairiDr. M. Hrairi MTHMTH22122212 -- Computational Methods and StatisticsComputational Methods and Statistics 1818

    Example 2- Solution

    The third column of[A]-1

    _ a

    !

    !

    -

    1

    0

    0

    1

    0

    0

    10271300.0100000.0

    010333333.0

    001

    3

    2

    1

    d

    d

    d

    _ a

    !

    !

    -

    09988.0

    004183.0

    006798.0

    1

    0

    0

    0120.1000

    293333.000333.70

    2.01.03

    3

    2

    1

    X

    x

    x

    x

  • 8/9/2019 Lecture02 - Factorization Methods

    19/23

    Dr. M. HrairiDr. M. Hrairi MTHMTH22122212 -- Computational Methods and StatisticsComputational Methods and Statistics 1919

    Example 2- Solution

    The matrix inverse [A]-1 is:

    Check your result by verifying that [A][A]-1 =[I]

    The final solution is

    -

    !

    09988.000271.001008.0

    004183.0142903.000518.0

    00 798.0004944.033249.01

    A

    _ a ? A _ a

    !

    -

    !!

    7

    50002.23

    4.71

    3.1985.7

    09988.000271.001008.0

    004183.0142903.000518.0006

    798

    .0004944

    .033249

    .01

    BAX

  • 8/9/2019 Lecture02 - Factorization Methods

    20/23

    Dr. M. HrairiDr. M. Hrairi MTHMTH22122212 -- Computational Methods and StatisticsComputational Methods and Statistics 2020

    Extending the Gaussian Elimination Process

    If pivoting is required to solve [A]{X}={B}, then there exists

    a permutation matrix [P] so that:

    [P][A ]=[L][U]

    The solution {X} is found in four steps:

    1. Construct the matrices [L], [U] and [P].

    2. Compute the column vector[P]{B}.

    . Solve [L]{D}=[P]{B} for{D} using forward substitution.

    . Solve [U]{X}={D} for{X} using back substitution.

  • 8/9/2019 Lecture02 - Factorization Methods

    21/23

    Dr. M. HrairiDr. M. Hrairi MTHMTH22122212 -- Computational Methods and StatisticsComputational Methods and Statistics 2121

    Example 3

    Use LU decomposition with permutation to solve the

    following system of equations

    0.0003 x1 + 3.0000 x2 = 2.0001

    1.0000 x1 + 1.0000 x2 = 1.0000

  • 8/9/2019 Lecture02 - Factorization Methods

    22/23

    Dr. M. HrairiDr. M. Hrairi MTHMTH22122212 -- Computational Methods and StatisticsComputational Methods and Statistics 2222

    Example 3 - Solution

    In matrix form [A ]{X}={B}

    We saw previously that pivoting is required to solve this

    system of equations, hence [P][A ]=[L][U]

    The solution {X} is found in four steps:

    1. Construct the matrices [L], [U] and [P].

    !

    -

    1

    0001.2

    11

    30003.0

    2

    1

    x

    x

    ? A

    -

    !

    01

    10P ? A

    -

    10003.0

    01L ? A

    -

    !

    9997.20

    11U

  • 8/9/2019 Lecture02 - Factorization Methods

    23/23

    Dr. M. HrairiDr. M. Hrairi MTHMTH22122212 -- Computational Methods and StatisticsComputational Methods and Statistics 2323

    Example 3 - Solution

    2. Compute the column vector[P]{B}.

    . Solve [L]{D}=[P]{B} for{D} using forward substitution.

    . Solve [U]{X}={D} for{X} using back substitution.

    !

    -

    0001.2

    1

    1

    0001.2

    01

    10

    _ a

    !

    !

    -

    8.1

    1

    0001.2

    1

    1000.0

    01

    2

    1

    _ a

    66667.0

    33333.0

    9998.1

    1

    9997.20

    11

    2

    1X

    x

    x