Romberg Method

Embed Size (px)

Citation preview

  • 8/7/2019 Romberg Method

    1/27

    1/10/2010 http://numericalmethods.eng.usf.edu 1

    Romberg Rule of Integration

    Major: All Engineering Majors

    Authors: Autar Kaw, Charlie Barker

    http://numericalmethods.eng.usf.eduTransforming Numerical Methods Education for STEM

    Undergraduates

    http://numericalmethods.eng.usf.edu/http://numericalmethods.eng.usf.edu/
  • 8/7/2019 Romberg Method

    2/27

    Romberg Rule ofIntegration

    http://numericalmethods.eng.usf.edu

    http://numericalmethods.eng.usf.edu/http://numericalmethods.eng.usf.edu/
  • 8/7/2019 Romberg Method

    3/27

    http://numericalmethods.eng.usf.edu3

    Basis of Romberg RuleIntegration

    =b

    a

    dx)x(fI

    The process of measuringthe area under a curve.

    Where:

    f(x)is the integrand

    a= lower limit of integration

    b= upper limit of integration

    f(x)

    a b

    y

    x

    b

    a

    dx)x(f

  • 8/7/2019 Romberg Method

    4/27

    http://numericalmethods.eng.usf.edu4

    What is The Romberg Rule?

    Romberg Integration is an extrapolation formula of

    the Trapezoidal Rule for integration. It provides abetter approximation of the integral by reducing theTrue Error.

  • 8/7/2019 Romberg Method

    5/27

    http://numericalmethods.eng.usf.edu5

    Error in Multiple Segment

    Trapezoidal RuleThe true error in a multiple segment Trapezoidal

    Rule with n segments for an integral

    Is given by

    =b

    a

    dx)x(fI

    ( )

    ( )

    n

    f

    nabE

    n

    i

    i

    t

    =

    = 12

    3

    12

    where for each i, is a point somewhere in thedomain , .

    i( )[ ]iha,hia ++ 1

  • 8/7/2019 Romberg Method

    6/27

    http://numericalmethods.eng.usf.edu6

    Error in Multiple Segment

    Trapezoidal RuleThe term can be viewed as an( )

    n

    fn

    ii

    =

    1

    approximate average value of in .( )xf [ ]b,a

    This leads us to say that the true error, Et

    previously defined can be approximated as

    2

    1

    nEt

  • 8/7/2019 Romberg Method

    7/27

  • 8/7/2019 Romberg Method

    8/27http://numericalmethods.eng.usf.edu8

    Error in Multiple Segment

    Trapezoidal Rule

    The true error gets approximately quartered asthe number of segments is doubled. Thisinformation is used to get a better approximationof the integral, and is the basis of Richardsons

    extrapolation.

  • 8/7/2019 Romberg Method

    9/27http://numericalmethods.eng.usf.edu9

    Richardsons Extrapolation for

    Trapezoidal Rule

    The true error, in the n-segment Trapezoidal ruleis estimated as

    tE

    2nCEt

    where Cis an approximate constantofproportionality. Since

    nt ITVE =

    Where TV = true value and = approx. valuenI

  • 8/7/2019 Romberg Method

    10/27http://numericalmethods.eng.usf.edu10

    Richardsons Extrapolation for

    Trapezoidal Rule

    From the previous development, it can be shownthat

    ( ) nITVnC

    222

    when the segment size is doubled and that

    32

    2nn

    n

    IIITV

    +

    which is Richardsons Extrapolation.

  • 8/7/2019 Romberg Method

    11/27http://numericalmethods.eng.usf.edu11

    Example 1The vertical distance covered by a rocket from 8 to 30seconds is given by

    =

    30

    8

    892100140000

    1400002000 dtt.

    tlnx

    a) Use Richardsons rule to find the distance covered.Use the 2-segment and 4-segment Trapezoidalrule results given in Table 1.

    b) Find the true error, Et for part (a).c) Find the absolute relative true error, for part (a).a

  • 8/7/2019 Romberg Method

    12/27http://numericalmethods.eng.usf.edu12

    Solutiona) mI 112662 = mI 111134 =

    Using Richardsons extrapolation formulafor Trapezoidal rule

    3

    2

    2

    nnn

    IIITV

    +

    and choosing n=2,

    3

    24

    4

    IIITV +

    3

    112661111311113 +=

    m11062=

  • 8/7/2019 Romberg Method

    13/27http://numericalmethods.eng.usf.edu13

    Solution (cont.)b) The exact value of the above integral is

    =

    30

    8 892100140000

    140000

    2000 dtt.tlnx

    m11061=

    Hence

    ValueeApproximatValueTrueEt =

    1106211061=

    m1=

  • 8/7/2019 Romberg Method

    14/27http://numericalmethods.eng.usf.edu14

    Solution (cont.)c) The absolute relative true error t would then be

    10011061

    1106211061

    =t

    %.009040=

    Table 2 shows the Richardsons extrapolationresults using 1, 2, 4, 8 segments. Results arecompared with those of Trapezoidal rule.

  • 8/7/2019 Romberg Method

    15/27http://numericalmethods.eng.usf.edu15

    Solution (cont.)Table 2: The values obtained using Richardsonsextrapolation formula for Trapezoidal rule for

    =30

    8

    892100140000

    1400002000 dtt.t

    lnx

    n TrapezoidalRule

    for TrapezoidalRule

    RichardsonsExtrapolation

    for RichardsonsExtrapolation

    1248

    11868112661111311074

    7.2961.8540.46550.1165

    --110651106211061

    --0.036160.0090410.0000

    Table 2: Richardsons Extrapola tion Values

    t t

  • 8/7/2019 Romberg Method

    16/27http://numericalmethods.eng.usf.edu16

    Romberg IntegrationRomberg integration is same as Richardsonsextrapolation formula as given previously. However,Romberg used a recursive algorithm for theextrapolation. Recall

    3

    2

    2

    nnn

    IIITV

    +

    This can alternately be written as

    ( )3

    2

    22

    nnnRn

    IIII

    +=

    1412

    2

    2

    +=

    nnn

    III

  • 8/7/2019 Romberg Method

    17/27http://numericalmethods.eng.usf.edu17

    Note that the variable TVis replaced by as the

    value obtained using Richardsons extrapolation formula.Note also that the sign is replaced by = sign.

    ( )Rn

    I2

    Romberg Integration

    Hence the estimate of the true value now is

    ( ) 42

    ChITVRn

    +

    Where Ch4 is an approximation of the true error.

  • 8/7/2019 Romberg Method

    18/27http://numericalmethods.eng.usf.edu18

    Romberg IntegrationDetermine another integral value with further halvingthe step size (doubling the number of segments),

    ( ) 324

    44

    nnnRn

    II

    II

    +=

    It follows from the two previous expressionsthat the true value TV can be written as

    ( ) ( ) ( )15

    24

    4

    RnRn

    Rn

    IIITV +

    ( ) ( )

    1413

    24

    4

    +=

    RnRnn

    III

  • 8/7/2019 Romberg Method

    19/27http://numericalmethods.eng.usf.edu19

    Romberg Integration

    214

    1

    11111

    +=

    +

    + k,IIII kj,kj,k

    j,kj,k

    The index krepresents the order of extrapolation.

    k=1 represents the values obtained from the regularTrapezoidal rule, k=2 represents values obtained using thetrue estimate as O(h2). The indexjrepresents the more andless accurate estimate of the integral.

    A general expression for Romberg integration can bewritten as

  • 8/7/2019 Romberg Method

    20/27http://numericalmethods.eng.usf.edu20

    Example 2

    The vertical distance covered by a rocket from

    8=t to 30=t seconds is given by

    =

    30

    8

    892100140000

    1400002000 dtt.

    tlnx

    Use Rombergs rule to find the distance covered. Usethe 1, 2, 4, and 8-segment Trapezoidal rule results asgiven in the Table 1.

  • 8/7/2019 Romberg Method

    21/27http://numericalmethods.eng.usf.edu21

    SolutionFrom Table 1, the needed values from originalTrapezoidal rule are

    1186811 =,I 1126621 =,I

    1111331 =,I 1107441 =,I

    where the above four values correspond to using 1, 2,4 and 8 segment Trapezoidal rule, respectively.

  • 8/7/2019 Romberg Method

    22/27

    http://numericalmethods.eng.usf.edu22

    Solution (cont.)To get the first order extrapolation values,

    11065

    3118681126611266

    3

    1,12,1

    2,11,2

    =

    +=

    +=

    IIII

    Similarly,

    11062

    3

    112661111311113

    3

    2,13,1

    3,12,2

    =

    +=

    += IIII

    11061

    3

    111131107411074

    3

    3,14,1

    4,13,2

    =

    +=

    += IIII

  • 8/7/2019 Romberg Method

    23/27

    http://numericalmethods.eng.usf.edu23

    Solution (cont.)For the second order extrapolation values,

    11062

    15

    110651106211062

    15

    1,22,2

    2,21,3

    =

    +=

    +=

    IIII

    Similarly,

    11061

    15

    110621106111061

    15

    2,23,2

    3,22,3

    =

    +=

    +=

    IIII

  • 8/7/2019 Romberg Method

    24/27

    http://numericalmethods.eng.usf.edu24

    Solution (cont.)For the third order extrapolation values,

    63

    1323

    2314

    ,,

    ,,

    IIII

    +=

    63

    110621106111061

    +=

    m11061

    =Table 3 shows these increased correct values in a treegraph.

  • 8/7/2019 Romberg Method

    25/27

    http://numericalmethods.eng.usf.edu25

    Solution (cont.)

    11868

    1126

    11113

    11074

    11065

    11062

    11061

    11062

    11061

    11061

    1-segment

    2-segment

    4-segment

    8-segment

    First Order Second Order Third Order

    Table 3: Improved estimates of the integral value using Romberg Integration

  • 8/7/2019 Romberg Method

    26/27

    Additional ResourcesFor all resources on this topic such as digital audiovisuallectures, primers, textbook chapters, multiple-choicetests, worksheets in MATLAB, MATHEMATICA, MathCad

    and MAPLE, blogs, related physical problems, pleasevisit

    http://numericalmethods.eng.usf.edu/topics/romberg_

    method.html

    http://numericalmethods.eng.usf.edu/topics/romberg_method.htmlhttp://numericalmethods.eng.usf.edu/topics/romberg_method.htmlhttp://numericalmethods.eng.usf.edu/topics/romberg_method.htmlhttp://numericalmethods.eng.usf.edu/topics/romberg_method.html
  • 8/7/2019 Romberg Method

    27/27

    THE END

    http :/ / numericalm etho ds.eng.us f.edu

    http://numericalmethods.eng.usf.edu/http://numericalmethods.eng.usf.edu/