W12_Interpolation0

download W12_Interpolation0

of 27

Transcript of W12_Interpolation0

  • 8/2/2019 W12_Interpolation0

    1/27

    Centre for Computer Technology

    ICT114Mathematics for

    Computing

    Week 12

    Interpolation

  • 8/2/2019 W12_Interpolation0

    2/27

    March 20, 2012 Copyright Box Hill Institute

    Curve Fitting

    A relation between two variables isexpressed mathematically by an equation

    connecting both. Finding equations approximating curves

    that fit the given set of data is called curve

    fitting.

  • 8/2/2019 W12_Interpolation0

    3/27

    March 20, 2012 Copyright Box Hill Institute

    Objectives

    Review week 11

    Problem of Interpolation

    Finite Differences Diagonal differencetable

    Newtons Forward Interpolation

    Newtons Backward Interpolation

    Lagranges Interpolation

  • 8/2/2019 W12_Interpolation0

    4/27

    March 20, 2012 Copyright Box Hill Institute

    Method of Least Squares (1)

    For the data points (x1, y1), (x2,y2),.(xn,yn)

    Let y = a+bx is the least squares line a and b are determined by solving the

    (normal) equations

    y = an + b x xy = a x + b x2

  • 8/2/2019 W12_Interpolation0

    5/27

    March 20, 2012 Copyright Box Hill Institute

    Method of Least Squares (2)

    ( y) (x2)( x) (xy)

    a = ---------------------------------------------

    n x2(x)2

    n xy ( x)( y)

    b = ----------------------------------------------n x2(x)2

  • 8/2/2019 W12_Interpolation0

    6/27

    Centre for Computer Technology

    NewtonsForward/BackwardInterpolation

  • 8/2/2019 W12_Interpolation0

    7/27

    March 20, 2012 Copyright Box Hill Institute

    Problem of Interpolation

    The following table givesthe value of logx for certain

    values of x To determine the value of

    logx for x = 96.45 (which is

    not given in the table), weuse the method ofInterpolation.

    x y=logx

    9596

    97

    9899

    1.97771.9823

    1.9868

    1.99121.9956

  • 8/2/2019 W12_Interpolation0

    8/27

    March 20, 2012 Copyright Box Hill Institute

    Finite Differences Diagonal

    Difference TableIn the table in the next slide

    a any argument (for x)h any positive constant

    f(x) = f(x+h) f(x) = first difference

    f(a) is called the leading termf(a),2f(a),3f(a) are called the leadingdifferences

  • 8/2/2019 W12_Interpolation0

    9/27

    March 20, 2012 Copyright Box Hill Institute

    Argument

    x

    Entry

    f(x)

    1st

    Differencef(x)

    2nd

    Difference2f(x)

    3rd

    Difference3f(x)

    a f(a)

    f(a)

    a+h f(a+h) 2f(a)

    f(a+h) 3f(a)

    a+2h f(a+2h) 2f(a+h)

    f(a+2h)

    a+3h f(a+3h)

  • 8/2/2019 W12_Interpolation0

    10/27

    March 20, 2012 Copyright Box Hill Institute

    Newton's Forward Interpolation

    u = (x x0)/h

    (x)= (x0+uh)= y0 + u.y0 + u(u-1)/2! .

    2y0+ ..

    .+ u(u-1)(u-2).(u-n+1)/n! .ny0

    y0 = f(x0)

  • 8/2/2019 W12_Interpolation0

    11/27

    March 20, 2012 Copyright Box Hill Institute

    Newtons Backward Interpolation

    u = (x xn)/h

    (x)= (xn+uh)= yn + u.yn-1 + u(u+1)/2! .

    2yn-2+ ..

    .+ u(u+1)(u+2).(u+n-1)/n! .ny0

    yn = f(xn)

  • 8/2/2019 W12_Interpolation0

    12/27

    March 20, 2012 Copyright Box Hill Institute

    Example

    From the followingtable determine thevalue of e0.1245 and

    e0.1895

    The first step is toconstruct the diagonal

    difference table

    x y=ex

    0.12

    0.13

    0.140.15

    0.16

    0.17

    0.18

    0.19

    1.127497

    1.138828

    1.1502741.161834

    1.173511

    1.185305

    1.197217

    1.209250

  • 8/2/2019 W12_Interpolation0

    13/27

    X y=ex y 2y

    0.12

    0.13

    0.14

    0.15

    0.16

    0.17

    0.18

    0.19

    1.127497

    1.138828

    1.150274

    1.161834

    1.173511

    1.185305

    1.197217

    1.209250

    0.011331

    0.011446

    0.011560

    0.011677

    0.011794

    0.011912

    0.012033

    0.000115

    0.000114

    0.000117

    0.000117

    0.000118

    0.000121

  • 8/2/2019 W12_Interpolation0

    14/27

    March 20, 2012 Copyright Box Hill Institute

    To determine e0.1245 we have to interpolate at thebeginning of the table.

    h = 0.01, x0 = 0.12, x = 0.1245

    so, u = (0.1245 0.12)/0.01 = 0.45

    therefore,f(0.1245) =y0 + u.y0 + u(u-1)/2! .

    2y0

    = 1.127497 + 0.45 x 0.011331

    + 0.45 x (0.45-1) /2! x 0.000115

    = 1.13258172

    Actual value = 1.13258202

  • 8/2/2019 W12_Interpolation0

    15/27

    March 20, 2012Copyright Box Hill Institute

    To determine e0.1895 we have to interpolate at theend of the table.

    h = 0.01, xn = 0.19, x = 0.1895

    so, u = (0.1895 0.19)/0.01 = -0.05

    therefore,

    (0.1895) =yn + u.yn-1+ u(u+1)/2! .2yn-2= 1.209250 + (-0.05) x 0.012033

    +(-0.05)(-0.05+1) / 2! x 0.000121

    = 1.20864548

    Actual value = 1.208645124

  • 8/2/2019 W12_Interpolation0

    16/27

    March 20, 2012Copyright Box Hill Institute

    Question

    Find the value of

    log(96.45) and

    log(98.75)using Newton's interpolation formulae fromthe table given in the second slide.

    Compare them with the actual value.

  • 8/2/2019 W12_Interpolation0

    17/27

    Centre for Computer Technology

    Lagranges Interpolation

  • 8/2/2019 W12_Interpolation0

    18/27

    March 20, 2012Copyright Box Hill Institute

    Lagranges Interpolation

    Newtons formulae can only be used for

    equidistant values of the argument

    But sometimes it is difficult to obtaintabulated values of a function forequidistant values of the argument.

    To deal with such a situation we useLagranges interpolation formula

  • 8/2/2019 W12_Interpolation0

    19/27

    March 20, 2012Copyright Box Hill Institute

    The Lagranges Interpolation formula is given below

    (x-x1)(x-x2)..(x-xn)(x) = --------------------------------- y0

    (x0-x1)(x0-x2)(x0-xn)

    (x-x0)(x-x2)..(x-xn)+ ---------------------------------- y1 +

    (x1-x0)(x1-x2)(x1-xn)

    (x-x0)(x-x2)..(x-xn-1)+ --------------------------------- yn

    (xn-x0)(xn-x2)(xn-xn-1)

  • 8/2/2019 W12_Interpolation0

    20/27

    March 20, 2012Copyright Box Hill Institute

    Example

    From the given table find the value of y for x = 3.743

    x y

    3.7428037

    3.7428822

    3.7429607

    3.74303923.7431176

    5531

    5532

    5533

    55345535

    x0 = 3.7428037, y0 = 5531

    x1 = 3.7428822, y1 = 5532x2 = 3.7429607, y2 = 5533

    x3 = 3.7430392, y3 = 5534

    x4 = 3.7431176, y4 = 5535

    x = 3.743

  • 8/2/2019 W12_Interpolation0

    21/27

    March 20, 2012Copyright Box Hill Institute

    (x-x1)(x-x2)(x-x3) (x-x4)(x) = ---------------------------------- y0 (0.0234250 x 5531)

    (x0-x1)(x0-x2)(x0-x3)(x0-x4)

    (x-x0)(x-x2)(x-x3)(x-x4)+ ------------------------------------ y1 + (-0.156157 x 5532)

    (x1-x0)(x1-x2)(x1-x3)(x1-x4)

    (x-x0

    )(x-x1

    )(x-x3

    )(x-x4

    )+ ------------------------------------ y2 + (0.702259 x 5533)

    (x2-x0)(x2-x1)(x2-x3)(x2-x4)

    (x-x0)(x-x1)(x-x2)(x-x4)+ ------------------------------------ y3 + (0.469666 x 5534)

    (x3-x0)(x3-x1)(x3-x2)(x3-x4)

    (x-x0)(x-x1)(x-x2)(x-x3)+ ------------------------------------ y4 (-0.0391929 x 5535)

    (x4-x0)(x4-x1)(x4-x2)(x4-x3)

  • 8/2/2019 W12_Interpolation0

    22/27

    March 20, 2012Copyright Box Hill Institute

    y = (3.743)

    = (0.0234250 x 5531)

    + (-0.156157 x 5532)

    + (0.702259 x 5533)+ (0.469666 x 5534)+ (-0.0391929 x 5535)

    = 5533.501

  • 8/2/2019 W12_Interpolation0

    23/27

    March 20, 2012Copyright Box Hill Institute

    Question

    From the given table find the value of y for x = 0.135

    x y

    0.12

    0.13

    0.14

    0.15

    1.127497

    1.138828

    1.150274

    1.161834

  • 8/2/2019 W12_Interpolation0

    24/27

    March 20, 2012Copyright Box Hill Institute

    Summary

    Newtons Forward Interpolation formula

    u = (x x0)/h

    (x) = (x0+uh)

    = y0

    + u.y0

    + u(u-1)/2! .2y0

    + ..

    .+ u(u-1)(u-2).(u-n+1)/n! .ny0

    y0 = f(x0)

  • 8/2/2019 W12_Interpolation0

    25/27

    March 20, 2012Copyright Box Hill Institute

    Summary

    Newtons Backward Interpolation formula

    u = (x xn)/h

    (x)= (xn+uh)= yn + u.yn-1+ u(u+1)/2! .

    2yn-2+ ..

    .+ u(u+1)(u+2).(u+n-1)/n! .ny0

    yn = f(xn)

  • 8/2/2019 W12_Interpolation0

    26/27

    March 20, 2012

    Copyright Box Hill Institute

    SummaryThe Lagranges Interpolation formula is given below

    (x-x1)(x-x2)..(x-xn)(x) = --------------------------------- y0

    (x0-x1)(x0-x2)(x0-xn)

    (x-x0)(x-x2)..(x-xn)+ ---------------------------------- y1 +

    (x1-x0)(x1-x2)(x1-xn)

    (x-x0)(x-x2)..(x-xn-1)+ --------------------------------- yn

    (xn-x0)(xn-x2)(xn-xn-1)

  • 8/2/2019 W12_Interpolation0

    27/27

    March 20, 2012

    Copyright Box Hill Institute

    References

    H L Verma and C W Gross : Introduction toQuantitative Methods,John Wiley

    JB Scarborough : Numerical Mathematical

    Analysis, Jon Hopkins Hall, New Jersey Gerald W. Recktenwald, Numerical Methods

    with MATLAB, Implementation and Application,Prentice Hall

    Murray Spiegel, John Schiller, Alu Srinivasan,Probability and Statistics, Schaums easyOutlines

    http://mathworld.wolfram.com