L-6 de Series Solution

download L-6 de Series Solution

of 88

Transcript of L-6 de Series Solution

  • 8/4/2019 L-6 de Series Solution

    1/88

    Solution in Series

  • 8/4/2019 L-6 de Series Solution

    2/88

    A linear second order homogeneous differential equation is of the

    form

    where are the functions of x (may be complexvalued) real or complex.

    10210 yxyxyx x,x,x 210

    The function is not identically zero, but may vanish at

    certain points.

    x0

    The points where are called Ordinary Points 00 x

    0 yxQyxPy

    Dividing the equation (1) by we get x0

    The points where are called Singular Points. 00 x

    A singular pointx = a is called regular singular point or

    regular singularity if xQaxxPax 2and have derivatives of all orders in the neighbourhood ofa.

  • 8/4/2019 L-6 de Series Solution

    3/88

    We note the following theorems for the solutions of the equation:

    10210 yxyxyx Theorem 1. Whenx = a is an ordinary point of (1), its every

    solution can be expressed in the form

    22210 .......axaaxaay

    Theorem 2. Whenx = a is a regularsingular point of (1), at

    least one solution can be expressed in the form

    32210 ... ........axaaxaaaxym

    Theorem 3. The series (2) and (3) are uniformly convergent at

    every point within the circle of convergence ata.

    This series is known as Frobenius series.

  • 8/4/2019 L-6 de Series Solution

    4/88

    Solve the equation

    10 yyxy

    in the neighbourhood ofx = 0.Sincex = 0 is an ordinary point, we take

    ......xaxaay 2210

    Differentiating (2), we get.......xaxaay 2321 32 31

    0

    1

    n

    nn xan

    20

    n

    nnxa

    and ......2432 34232 xaxaay

    4120

    2

    n

    nn xann

    Substitute (2), (3) and (4) in (1)

  • 8/4/2019 L-6 de Series Solution

    5/88

    or

    011200

    1

    0

    2

    n

    nn

    n

    nn

    n

    nn xaxanxxann

    011200

    110

    2

    nnn

    nnn

    nnn xaxanxann

    or 012010

    2

    n

    nn

    n

    nn

    n

    nn xaxnaxann

    or 011221

    202

    n

    n

    nn xanannaa

    This holds for all values of x if, and only if

    ,........,,,nanannaa

    nn 4321011202

    2

    02

    or ,. .......,,,nan

    aaa nn 4321for2

    1and

    2

    1202

  • 8/4/2019 L-6 de Series Solution

    6/88

    06

    04

    02

    246

    1

    24

    1

    2

    1

    aa

    aa

    aa

    17

    15

    13

    357

    1

    35

    1

    3

    1

    aa

    aa

    aa

    and so on.

    Using these values in (2), we get

    ........xaxaxaxaxaay

    514

    03

    12

    01035

    1

    24

    1

    3

    1

    2

    1

    ..... ....xxxxa

    ..... ....xxxay

    5531

    6420

    357

    1

    35

    1

    3

    1

    246

    1

    24

    1

    2

    11

    This the general solution of (1) with a0 and a1 as arbitrary

    constants.

  • 8/4/2019 L-6 de Series Solution

    7/88

    To solve the differential equation

    whenx = a is a regular singularity, that is

    and

    have derivatives of all orders, then at least one solution can

    be in the form of the Frobenius series

    0210 yxyxyx

    00 a

    xxax

    x

    xax,

    0

    22

    0

    1

    002

    210 aaxaaxaaaxym

    ,...........

    To understand steps involved in this method (Frobenius

    Method) we consider the Bessel differential equation oforder , where is any constant,

    02222

    2

    yxxxdx

    dy

    dx

    yd

  • 8/4/2019 L-6 de Series Solution

    8/88

    Comparing with we find that

    102222

    2

    yxxxdx

    dy

    dx

    yd We note thatx = 0 is a singular point and dividing byx2 we

    get 0

    2

    22

    2

    21

    y

    x

    x

    dx

    dy

    xdx

    yd 0 yxQyxPy

    222

    and1

    xxQxxxPthereforex = 0 is a regular singularity of the DE and we take

    ...........xaxaaxy m 2210 20

    n

    nmnxa

    The first and second derivatives of y are

    ,xanmyn

    nmn

    0

    1

    0

    231

    n

    nmnxanmnmy

    Substituting for in (1) we getyy,y and

  • 8/4/2019 L-6 de Series Solution

    9/88

    0

    1

    0

    22

    0

    1

    0

    22

    n

    nmn

    n

    nmn

    n

    nmn

    xaxxanmx

    xanmnmx

    01

    111

    22

    2

    11

    20

    2

    n

    nmnn

    mm

    xaanmnmnm

    xammmxammm

    or

    010

    2

    0

    2

    n

    nmn

    n

    nmn xaxanmnmnm or

    012

    2

    0

    2

    n

    nmn

    n

    nmn xaxanmnmnm or

  • 8/4/2019 L-6 de Series Solution

    10/88

    Equating the coefficients of powers of x, stating with the

    least power, we get

    401 02

    ammm

    5011 12 ammm

    .................,,,, 5432

    601 22

    n

    aanmnmnm nn

    Sincea0 0, therefore from (4)

    701 2 mmmThis equation is known as indicial equation. The indicial

    equation for a second order DE is a quadratic and gives twovalues of index m. We expect a solution of DE

    corresponding to each value of m. From (7) we have

    022 m 21 and mm

  • 8/4/2019 L-6 de Series Solution

    11/88

    From (5)

    011 12 amm 01 122 am or

    Since first factor is not zero for any value of m , therefore

    From (6), forn= 2, 3, 4, 5, ..

    2

    2

    1

    nmnmnm

    aa nn

    22

    nmnm

    an

    )(82

    nmnm

    an

    ,01

    aSince form (8) we find that

    ............... 7530 aaa

    nmnm

    aa nn

    2forn= 2, 4, 6,

    .a 01

  • 8/4/2019 L-6 de Series Solution

    12/88

    nmnm

    aa

    nn

    22

    122

    )(or forn= 1, 2, 3,

    From this relation we can find in terms of.............,,, 642 aaa .0a

    It is expected that each value of m will provide one set of

    values of these coefficients and hence two solutions of DE.

    ,forNow 1 mm 22212

    2

    nn

    aa nn)(

    ,forand 2 mm

    222

    122

    nn

    aa

    nn

    )(

    nna n2

    12

    2)(

    nn

    a n2

    12

    2

    )(

    Clearly if = 0 we will get identical values and if is an

    integer some coefficients will become infinite after somen.

  • 8/4/2019 L-6 de Series Solution

    13/88

    Case I: When neither = 0 nor is an integer.

    In this case we have two distinct solutions of indicial equation

    which do not differ by an integer. Two linearly independent

    solutions shall be obtained in this case.

    1For mm

    nn

    aa

    nn 2

    122

    2

    )(

    1220

    2 aa

    22222

    4 aa

    121224

    0a

    33224

    6a

    a

    12312326

    0a

    1212

    12

    02

    ......! nnn

    aa

    n

    nn

    and so on.

    For general value of n, (n = 1, 2, 3, ), we have

  • 8/4/2019 L-6 de Series Solution

    14/88

    2forsimilarlyProceeding mm

    solutionget theweforTherefore 1 mm

    .......!!

    123321222121 6

    6

    4

    4

    2

    2

    01

    xxx

    xay

    .......!!

    123321222121 6

    6

    4

    4

    2

    2

    02

    xxx

    xay

    2211issolutiongeneralThe ycycy

  • 8/4/2019 L-6 de Series Solution

    15/88

    Gamma Function

    The Gamma function is defined as

    .)( 0for10

    dxxe xThe integral is known as Eulers integral of second kind.

    dxe x0

    1)(

    0

    xe 1

    dxxe x 0

    1)(

    dxxeexxx 1

    00

    )( )()( 112 1

    )()( 223 !212 )()( 334 !! 323 !)( nn 1 dxxe x 2

    1

    021

  • 8/4/2019 L-6 de Series Solution

    16/88

    )( 121

    0 a

    In

    .......

    !!

    123321222121

    6

    6

    4

    4

    2

    2

    01xxx

    xay

    ..........)(!

    )(!)()(

    112332

    1122211211

    2

    6

    6

    4

    4

    2

    2

    1

    x

    xxxy

    0

    2

    121

    1

    k

    kkx

    kky

    )(!

    0

    2

    211

    1

    k

    kkx

    kk

    )()(

    take

  • 8/4/2019 L-6 de Series Solution

    17/88

    0

    2

    211

    1

    k

    kkx

    kkxJ

    )()()(

    Definition:

    is known as Bessel function of the first kind of order .

    Corresponding to the other solution y2 , we have

    0

    2

    211

    1

    k

    kkx

    kkxJ

    )()()(

    This is known as Bessel function of the first kind of order -.

    Hence the complete solution ofBessel equation can be

    expressed as.)()( xJbxJay

  • 8/4/2019 L-6 de Series Solution

    18/88

  • 8/4/2019 L-6 de Series Solution

    19/88

  • 8/4/2019 L-6 de Series Solution

    20/88

    This gives only one solution instead of two. The second

    solution is given by when = 0.m

    y

    1

    ....

    log

    4

    2

    2

    2

    422

    2

    222

    4

    2

    2

    0

    11

    mmmm

    x

    mm

    xxa

    xymy

    m

    Therefore for m = 0

    ....log211

    422 224

    22

    010

    12

    xxaxymyy

    m

    ....

    222

    6

    22

    4

    2

    2

    01642422

    1xxx

    ay

    The general solution is then 2211 ycycy

  • 8/4/2019 L-6 de Series Solution

    21/88

    ....

    222

    6

    22

    4

    2

    2

    01642422

    1xxx

    ay

    Consider the solution

    ....

    )()(26

    6

    24

    4

    2

    2

    032122122

    1xxx

    a

    ....)!()!( 26

    6

    24

    4

    2

    2

    0322221

    xxx

    a

    From the definition of Bessel function

    0

    2

    0

    211

    1

    k

    kkx

    kk

    xJ

    )()(

    )(

    0

    2

    2

    2

    1

    k

    kkx

    k )!(10 aThus with is Bessel function of first kind of order zero.1y

    The other solution with is Bessel function of second

    kind or Neumann function of order zero and is denoted by .

    10 a2y)(xY0

  • 8/4/2019 L-6 de Series Solution

    22/88

    Case III When is an integer, we have seen some of the

    coefficients become infinite.Now suppose that = n a positive integer, then we obtain the

    solution

    0

    2

    211

    1

    k

    knk

    nx

    nkkxJ

    )()()(

    To find the other solution since one solution is known we take

    the general solution as y = u(x) y1(x)

    Submit the answer to the above question on the day of

    Next Tutorial

    Show that

    and the complete solution is

    constantsareandwhere

    1

    2badx

    xJxbaxu

    n

    )(

    )(

    .

    )(

    )()( dx

    xJx

    xJbxJay

    n

    nn 2

    1

  • 8/4/2019 L-6 de Series Solution

    23/88

    In general if the roots of indicial equation differ by an integer, we

    can obtain only one solution. The second solution is given by

    1

    1mmm

    y

    The function is the Bessel

    function of second kind of order n or Neumann function.

    dx

    xJxxJxY

    n

    nn 2

    1

    )()()(

  • 8/4/2019 L-6 de Series Solution

    24/88

    Recurrence relations forBessel functions

    1

    1

    1 1

    1 1

    1 1

    1

    (1) [ ( )] ( )

    (2) [ ( )] ( )

    (3) 2 ( ) [ ( ) ( )]

    (4) 2 ( ) ( ) ( )

    (5) ( ) 2 ( ) ( )

    (6) ( ) ( ) ( )

    n n

    n n

    n n

    n n

    n n n

    n n n

    n n n

    n n n

    dx J x x J x

    dx

    d

    x J x x J xdx

    nJ x x J x J x

    J x J x J x

    xJ x nJ x xJ x

    xJ x nJ x xJ x

  • 8/4/2019 L-6 de Series Solution

    25/88

    Show that )()]([ xJxxJxdx

    dn

    nn

    n1

    We havekn

    k

    kn

    n

    n x

    knkxxJx

    2

    0 21

    1

    )(!

    )()(

    kn

    kkn

    k

    xknk

    22

    02

    12

    1

    )(!

    )(

    122

    02

    12

    122

    kn

    kkn

    k

    nn xknk

    knxJxdx

    d

    )(!

    ))(()]([

    122

    012

    2

    1

    kn

    kkn

    k

    xknk )(!

    )(

    122

    02

    12

    12

    kn

    kkn

    k

    xknk

    kn

    )(!

    ))((

    12

    012

    2

    1

    kn

    kkn

    kn x

    knkx

    )(!

    )(

    )(xJx nn 1

  • 8/4/2019 L-6 de Series Solution

    26/88

    Show that )()]([ xJxxJxdx

    dn

    nn

    n1

    We have )(xJxn

    n

    k

    kkn

    k

    xknk

    2

    02

    12

    1

    )(!

    )(

    kn

    k

    kn x

    knkx

    2

    0 21

    1

    )(!)(

    )]([ xJxdx

    dnn 12

    12

    12

    12

    k

    kkn

    k

    xknk

    k

    )(!

    )(

    12

    112

    112

    1

    k

    kkn

    k

    xknk )()!(

    )(

    112

    112

    1

    11112

    1

    )(

    )( )()!(

    )( kkn

    kx

    knk

    0k

    12

    012

    112

    1

    k

    kkn

    k

    xknk )()!(

    )(

  • 8/4/2019 L-6 de Series Solution

    27/88

    12

    012

    112

    1

    nk

    kkn

    kn x

    knkx

    )()!(

    )(

    )(xJx

    n

    n

    1

    )]([ xJxdx

    d

    n

    n

    Show that

    )()()(

    )]()([)(

    xJxJxJ

    xJxJxxnJ

    nnn

    nnn

    11

    11

    2

    2

    We have

    )()]([

    )()]([

    xJxxJxdx

    d

    xJxxJxdx

    d

    n

    n

    n

    n

    nn

    nn

    1

    1

  • 8/4/2019 L-6 de Series Solution

    28/88

    )()]([)(

    )()]([)(

    xJxxJ

    dx

    dxxJnx

    xJxxJdx

    dxxJnx

    nn

    nn

    nn

    nn

    nn

    nn

    11

    11

    Therefore

    or

    )()()]([)(

    )()()]([)(

    2

    1

    1

    1

    xxJxJdx

    dxxnJ

    xxJxJdx

    dxxnJ

    nnn

    nnn

    Subtracting (2) from (1) )()]()([)( 32 11 xJxJxxnJ nnn

    Adding (1) and (2)1 12 ( ) ( ) ( ) (4)n n nJ x J x J x

    1( ) ( ) ( ) (6)n n nxJ x nJ x xJ x From (2)

    From (3) 1 1( ) 2 ( ) ( ) (5)n n nxJ x nJ x xJ x

  • 8/4/2019 L-6 de Series Solution

    29/88

    All the Bessel functionsJn(x) and their derivatives can be

    expressed in terms ofJ0(x) andJ1(x) by using recurrence

    relations.

    Example: ExpressJ5(x) in terms ofJ0(x) andJ1(x).

    )]()([)( xJxJxxnJ nnn 112 We have

    )(xJn 1Or

    )()()( xJxJxxJ 123

    4

    Therefore )()()( xJxJx

    xJ 0122

    )()()( xJxJxJxx 101

    24

    )()( xJx

    xJx

    012

    41

    8

    )()( xJxJx

    n

    nn 1

    2

  • 8/4/2019 L-6 de Series Solution

    30/88

    )()()( xJxJx

    xJ 2346

    )()()()( xJxJx

    xJx

    xJxx

    01012 24186

    )()( xJ

    x

    xJ

    xx

    0213

    241

    848

    )()()( xJxJx

    xJ 3458

    )()()()( xJx

    xJx

    xJx

    xJxxx 0120213

    41

    8241

    8488

    )()( xJxx

    xJxx

    03124

    192121

    72384

  • 8/4/2019 L-6 de Series Solution

    31/88

    Example: Express in terms ofJ0(x) andJ1(x).)(xJ1

    )()()( xJxJxJ nnn 112

    We have

    )()()( xJxJxJ nnn 112 Differentiating we get

    )()()()()( xJxJxJxJxJ nnnnn 222

    1

    2

    12

    )()()()( xJxJxJxJ nnnn 224

    1

    2

    1

    4

    1

    )()(2)(4

    122 xJxJxJ nnn

    )()(2)(4

    1)( 3111 xJxJxJxJ Putting n =1

  • 8/4/2019 L-6 de Series Solution

    32/88

    )]()([)(2 11 xJxJxxnJ nnn From the relationFor n = 0 )]()([0 11 xJxJx )]()( 11 xJxJ

    )(4

    )(18

    )( 0123 xJxxJ

    xxJ

    Also from the previous example

    )()(2)(41)( 3111 xJxJxJxJ Therefore substituting in

    )(4)(18)(2)(4

    1)( 012111 xJx

    xJx

    xJxJxJ

    )(4)(484

    1)( 0121 xJx

    xJx

    xJ )(1

    )(12

    012xJ

    xxJ

    x

  • 8/4/2019 L-6 de Series Solution

    33/88

    We havekn

    k

    k

    nx

    knk

    xJ

    2

    0 21

    1

    )(!

    )()(

    nrkeirkn ..Putting

    rn

    nr

    nr

    nx

    rnrxJ

    2

    2)1()!()1()(

    R l ti b t B l d

  • 8/4/2019 L-6 de Series Solution

    34/88

    Relation between Bessel andtrigonometric functions

    ,cos2

    )(21

    xx

    xJ

    Show that

    0

    2

    21

    1

    k

    kkx

    kkxJ

    )(!)( We have

    0

    2

    21

    21

    21 21

    1

    k

    kkx

    kkxJ

    )(!)(

    0

    2

    21 21

    12

    k

    kkx

    kkx )(!

    .........

    )(!)(!)()(

    6

    27

    4

    25

    2

    23

    21 23

    1

    22

    1

    2

    112 xxx

    x

  • 8/4/2019 L-6 de Series Solution

    35/88

    ..........)(!

    )(!)()()(

    6

    21

    21

    23

    25

    4

    21

    21

    23

    2

    21

    21

    21

    23

    1

    22

    1

    2

    112

    21

    x

    xx

    xxJ

    ..........

    )(

    6

    2

    1

    2

    2

    2

    3

    2

    4

    2

    5

    2

    6

    4

    21

    22

    23

    24

    2

    2

    21

    21

    2

    1

    2

    1

    212

    1

    1

    12

    x

    xx

    x

    ..........

    !!!

    642

    6

    1

    4

    1

    2

    11

    2xxx

    x xx cos2

  • 8/4/2019 L-6 de Series Solution

    36/88

    Bessel function is the coefficients of tn in the

    expansion of)(xJn

    )(t

    txe

    121

    This function is known as generating function of Bessel

    function of first kind.

    nn

    ntx xJte t )()( 1

    21

  • 8/4/2019 L-6 de Series Solution

    37/88

    Reduce the equation

    to Bessel form

    )(1022222

    2

    ynxxxdx

    dy

    dx

    yd

    02222

    2

    yxxxdx

    dy

    dx

    yd Write down the solution of (1) in terms of Bessel function.

  • 8/4/2019 L-6 de Series Solution

    38/88

    Orthogonality of Bessel Functions

    where , are roots of Jn (x) = 0

    ,)]([

    ,)()( 2

    12

    1

    1

    0

    0

    n

    nn

    J

    dxxJxxJ

    The solutions of the equations

    and

    are respectively.)()( xJvxJu nn and

    2

    2

    2 2 2 2 0 (2)d v dvdxdx

    x x x n v

    2

    2

    2 2 2 2

    0 (1)

    d u du

    dxdxx x x n u

    Multiply (1) with v, (2) with u and subtract

    022222

    2

    2

    2

    uvxuvxuvxdxdv

    dxdu

    dx

    vd

    dx

    ud )()(

  • 8/4/2019 L-6 de Series Solution

    39/88

    2 2

    2 2

    2 2or ( ) 0d u d v du dvdx dxdx dx

    x v u v u xuv

    dxdv

    dxdu

    dx

    vd

    dx

    ud uvuvxxuv )(2

    2

    2

    222or

    Integrate with respect to x from 0 to 1

    10

    1

    0

    22)(

    dxdv

    dxdu uvxxuvdx )()]([ 31 xdx

    dvdxdu uvx

    Since ( )nu J x

    [ ( ) ]du d ndx dxJ x ( )( )

    [ ( ) ] d xd nd x dxJ x

    ( )nJ x

    Similarly as )( xJv n ])( xJndxdv Substituting in (3)

    )]([dxdv

    dxdu

    dxd uvx

    dxdv

    dxdu

    dxdv

    dxdu

    dxd uvuvxxuv )(22or

  • 8/4/2019 L-6 de Series Solution

    40/88

    22

    1

    0or

    )()()()(

    )()(nnnn

    nn

    JJJJ

    dxxJxxJ

    )()()()()()( nnnnnn JJJJdxxJxxJ 1

    0

    22

    Since and are the roots of , therefore if 0)(xJn

    0

    1

    0 dxxJxxJ nn )()(

    If = , then

    1

    0

    2dxxJx n )]([ 22

    )()()()(lim nnnn

    JJJJ

    22

    )()(lim nn

    JJ])([ 0nJ

  • 8/4/2019 L-6 de Series Solution

    41/88

    12

    0

    [ ( )]nx J x dx ( ) ( )

    lim2

    n nJ J

    1( ) ( ) ( )n n nxJ x nJ x xJ x

    21[ ( )]2

    nJ

    From the relation

    1( ) ( )n nJ J

    12

    0

    [ ( )]nx J x dx 2

    11

    [ ( )]2

    nJ

    Therefore

    3

  • 8/4/2019 L-6 de Series Solution

    42/88

    3

    1

    n 1

    Expand ( ) in the interval 0 3 in terms of ( ),

    where are given by (3 ) 0.

    n

    n

    f x x x J x

    J

    3 11

    n n

    n

    x c J x

    Let

    Multiply both side by and integrate with respect to

    x from 0 to 3.

    1 mxJ x

    3 3

    4

    1 1 1

    10 0

    (1)m n n mn

    x J x dx c J x x J x dx

    3

    1 1

    1 0

    n n m

    n

    c xJ x J x dx

    3

    1 1

    10

    Right hand side of (1) n n mn

    c J x x J x dx

    1

    1 1

    1 0

    R.H.S. 3 3 3 (3 )n n m

    n

    c tJ t J t d t

    Putting x = 3t

  • 8/4/2019 L-6 de Series Solution

    43/88

    1

    1 1

    1 0

    R.H.S. of (1) 3 3 3 3n n mn

    c tJ t J t dt

    1

    2

    1 11 0

    3 3 3n n mn c tJ t J t dt

    22

    2

    3

    3 22 [ ( )] ( )m mc J

    3 4 10

    L.H.S. of (1) = mx J x dx

    1We have [ ( )] ( )n n

    n n

    dx J x x J xdx

    1

    1

    ( ) ( ) (3) n nn nx J x dx x J x1

    ( ) ( ) ( ) ( ) ( )n nn nx J x x J x d x 1( ) ( ) n nn nx J x x J x dx

    3 2 2 10

    = [ ]m

    x x J x dx

    3 4 10

    L.H.S. of (1) = mx J x dx

    Integrate by parts using (3)

    32 23

  • 8/4/2019 L-6 de Series Solution

    44/88

    32 232 2200

    L.H.S. of (1) = 2

    m mm m

    x J x x J xx x dx

    4 3

    2 3

    2

    0

    3 3 2=

    m mm mJ x J x dx 34 32 30

    3 3 2=

    m m

    m m m

    J x J x[By using (3)]

    4 32 32

    3 3 2 3 3= 4( )

    m mm m

    J J x

    22

    2

    33

    2

    [ ( )]m m

    c J 4 32 3

    2

    3 3 2 3 3=

    m mm m

    J J x

    2 32 22

    63 3 2 3

    3[ ]

    [ ( )]m

    m m m

    m

    c J JJ

  • 8/4/2019 L-6 de Series Solution

    45/88

    Hence

    3 2 3 12 21 2

    63 3 2 3

    3[ ]

    [ ( )]m m m

    m m m

    x J J J xJ

  • 8/4/2019 L-6 de Series Solution

    46/88

    2 4 6 8 10x

    -0.4

    -0.2

    0.2

    0.4

    0.6

    0.8

    1

    Jn x

    J1

    J0

    Bessel Fuctions

  • 8/4/2019 L-6 de Series Solution

    47/88

    2 4 6 8 10

    x

    -0.5

    0.5

    1

    1.5

    2

    Jn x

    J 0.5

    J0.5

    Bessel Fns

  • 8/4/2019 L-6 de Series Solution

    48/88

    Solution in Series

    Legendre polynomials

  • 8/4/2019 L-6 de Series Solution

    49/88

    The Legendre equation

    arises in the problems with spherical symmetry. 21 2 1 0 1( ) ( )x y xy n n y

    This equation has two regular singular pointsx = -1 and 1.

    0

    mmm

    y a x

    Substituting in (1), simplifying and equating the coefficients

    of powers of x to zero we get

    2 0

    3 1

    2

    2 1 06 2 1 0

    2 1 1 2 1 0

    2 3

    ( )[ ( )]

    ( )( ) [ ( ) ( )]

    , , ............

    m m

    a n n aa n n a

    m m a m m m n n a

    m

    x = 0 is an ordinary point. Therefore in the neighbourhood of

    0 we can have a series solution of the form.

  • 8/4/2019 L-6 de Series Solution

    50/88

    2 0 3 12 1 0 6 2 1 0( ) [ ( )]a n n a a n n a 2

    1for 2 3

    2 1

    ( )( ), , ............

    ( )( )

    m mn m n m

    a a m

    m m

    This is the recurrence relation or recursion formula.

    Two independent solutions are

    2 4

    1

    1 2 1 3

    1 2 4

    ( ) ( ) ( )( )

    ..................! !

    n n n n n n

    y x x

    3 52

    1 2 3 1 2 4

    3 5

    ( )( ) ( )( )( )( ).........

    ! !

    n n n n n ny x x x

    Both the series converge for |x| < 1.

    If n is a non-negative integer, one of them terminates and one

    solution then is a polynomial.

    The general solution isy(x) = a0 y1 + a1y2

  • 8/4/2019 L-6 de Series Solution

    51/88

    We rewrite the recursion formula

    21

    for 2 32 1

    ( )( ), , ............

    ( )( )m m

    n m n ma a m

    m m

    2

    2 1as where 2

    1

    ( )( )

    ( )( )m m

    m ma a m n

    n m n m

    2na 12 2 1

    ( )( )

    nn n an

    Then using this relation we can express all the coefficients

    in terms of2 4 6, , ..................n n na a a .na

    2

    2

    2

    ( )!

    ( !)n n

    na

    nWe take

    21 2

    2 2 1 2( ) ( )!

    ( ) ( !)n

    n n nn n

    2

    1 2 2 1 2 2

    2 2 1 2

    ( ) ( )( )!

    ( ) ( !)n

    n n n n n

    n n

    2 2

    2 1 2

    ( )!

    ( )!( )!n

    n

    n n

    The value of is so selected that resulting polynomial takes

    the value 1 atx = 1.na

    2 3( )( ) 2 3 2 2( )( ) ( )!

  • 8/4/2019 L-6 de Series Solution

    52/88

    4na 22 34 2 3

    ( )( )

    ( )n

    n na

    n

    2 3 2 2

    4 2 3 2 1 2

    ( )( ) ( )!

    ( ) ( )!( )!n

    n n n

    n n n

    2 3 2 2 2 3 2 4

    4 2 3 2 1 2

    ( )( ) ( )( ) ( )!( ) ( )!( )!n

    n n n n nn n n

    1 2 4

    2 2 2 4

    ( )!

    ( )!( )!n

    n

    n n

    2 4

    2 2 2 4

    ( )!

    ! ( )!( )!n

    n

    n n

    6na 2 63 2 3 6

    ( )!

    ! ( )!( )!n

    n

    n n

    2n ka 2 212 2

    ( )!( )

    ! ( )!( )!

    k

    n

    n k

    k n k n k

    In general when 2 0,n k

  • 8/4/2019 L-6 de Series Solution

    53/88

    The resulting polynomial solution of Legendre DE is

    2 4

    2

    6

    2 2 2 2 4

    2 2 1 2 2 2 2 42 6

    3 2 3 6

    ( )! ( )! ( )!

    ( !) ( )!( )! ! ( )!( )!( )!

    .....................! ( )!( )!

    n n n

    n n n

    n

    n

    n n nx x x

    n n n n nn

    xn n

    20

    12 2

    2 2

    1 12 2

    where or whichever is an integer.

    ( )!

    ( )!( )!( )!

    Mm n m

    nm

    n n

    n m

    xm n m n m

    M

    The polynomial (1) is known as Legendre Polynomial of

    degree n andis denoted by ( ).nP x

  • 8/4/2019 L-6 de Series Solution

    54/88

    20

    12 2

    2 21

    2 2where or whichever is an integer.

    ( )!( )

    !( )!( )!

    Mm n m

    n n

    mn n

    n mP x x

    m n m n mM

    Since

    Therefore

    0 1( )P x

    212

    3 1( )x 2( )P x

    4 218

    35 30 3( )x x 4( )P x

    x1( )P x

    31

    2

    5 3( )x x3( )P x

    5 318

    63 70 15( )x x x5( )P x Observe that 1 1( )nP

    h f d l i l 1

  • 8/4/2019 L-6 de Series Solution

    55/88

    Graph of Legendre PolynomialP0 = 1.

    -1 -0.5 0.5 1

    x

    0.5

    1

    1.5

    2

    P0x

    G h f L d P l i l

  • 8/4/2019 L-6 de Series Solution

    56/88

    Graphs of Legendre Polynomials

    0 11( ) ( )P x P x x

    -1 -0.5 0.5 1

    x

    -1

    -0.5

    0.5

    1

    Pn x

    P1

    P0

    G h f L d P l i l

  • 8/4/2019 L-6 de Series Solution

    57/88

    -1 -0.5 0.5 1

    x

    -1

    -0.5

    0.5

    1

    Pn x

    P2

    P1

    P0

    Graphs of Legendre Polynomials

    210 1 2 2

    1 3 1( ) ( ) ( )P x P x x P x

    G h f L d P l i l

  • 8/4/2019 L-6 de Series Solution

    58/88

    Graphs of Legendre Polynomials

    210 1 2 2

    313 2

    1 3 1

    5 3

    ( ) ( ) ( )

    ( )

    P x P x x P x

    P x x

    -1 -0.5 0.5 1

    x

    -1

    -0.5

    0.5

    1

    Pn x

    P3

    P2

    P1

    P0

  • 8/4/2019 L-6 de Series Solution

    59/88

    -1 -0.5 0.5 1

    x

    -1

    -0.5

    0.5

    1

    Pn x

    P5

    P4

    P3

    P2

    P1

    P0

    Graphs of Legendre Polynomials

    0 1( )P x 21

    23 1( )x 2( )P x

    4 218

    35 30 3( )x x 4( )P x

    x1( )P x 31

    25 3( )x x3( )P x

    5 318

    63 70 15( )x x x5( )P x

    3

  • 8/4/2019 L-6 de Series Solution

    60/88

    Question: Obtain the 7th derivative of

    1 2

    1 2

    If and be two function of possesing derivatives

    of nth order, then

    ( ) ( ) ( ) ( )

    ( ) ( ) ( )

    ( ) ( )

    ( ) ...................

    n n n n n n

    n n r r nr

    u x v x x

    uv u v C u v C u vC u v uv

    Leibnitzs Theorem

    3sinxe x

    Question: Obtain the 7th derivative of2 2 31( ) xx e

  • 8/4/2019 L-6 de Series Solution

    61/88

    Rodrigues Formula

    21 1

    2

    ( ) [( ) ]

    !

    nn

    n n n

    dP x x

    n dx

    2 1( )nv x Let Then 2 12 1( )ndv nx x

    dx

    2 21 2 1( ) ( )

    ndvx nx x

    dx

    2 1 2 0( )

    dvx nxv

    dx

    Differentiate n+1 timesusing Leibnitzs theorem

    2 2 1

    1

    11 1 2 2

    2

    2 1 0

    ( ) ( ) ( )

    ( ) ( )

    ( )( ) ( )

    !

    [ ( ) ]

    n n n

    n n

    n nx v n x v v

    n xv n v

  • 8/4/2019 L-6 de Series Solution

    62/88

    Express in terms of Legendre polynomials3 24 2 3 8

  • 8/4/2019 L-6 de Series Solution

    63/88

    Express in terms of Legendre polynomials.3 24 2 3 8x x x

    Therefore3 24 2 3 8x x x

    0 1( )P x 21

    23 1( )x 2( )P x

    x1( )P x 31

    25 3( )x x3( )P x

    31 2 35

    [ ( ) ]P x x3x 3 11 2 35

    [ ( ) ( )]P x P x2

    12 1

    3[ ( ) ]P x 2x

    3 1 2 14 2

    2 3 2 1 3 85 3

    [ ( ) ( )] [ ( ) ] ( )P x P x P x P x 3 2 1 0

    8 4 3 22

    5 3 5 3( ) ( ) ( ) ( )P x P x P x P x

    Since is a cubic polynomial, it can be

    expressed in terms of

    3 24 2 3 8x x x

    E i Sh h122 n

  • 8/4/2019 L-6 de Series Solution

    64/88

    1 12 221 1

    22 2

    1 2 1 2

    1

    11 2 22 2

    ( )

    ( )

    ( ) [ ( )] ....!

    xt t t x t

    t x t t x t

    Exercise: Show that 220

    1 2 ( )n

    nn

    xt t t P x

    Hint: Using Binomial theorem

    Express right hand side in powers of t to obtain the result.

    Generating function of Legendre Polynomials

    1221 2xt t

  • 8/4/2019 L-6 de Series Solution

    65/88

    Recurrence formulae for Legendre polynomials

    1 11 1 2 1. ( ) ( ) ( ) ( ) ( )n n nn P x n xP x n P x 12. ( ) ( ) ( )n n nnP x x P x P x

    1 13 2 1. ( ) ( ) ( ) ( )n n nn P x P x P x 1 14. ( ) ( ) ( )n n nP x x P x n P x

    215 1. ( ) ( ) [ ( ) ( )]n n nx P x n P x x P x

    All these formulas can be proved using generating function for ( ).nP x

    We have 1220

    1 2 ( ) ( )n

    nn

    xt t t P x I

    Differentiating with respect tot, we get 322 112

    0

    1 2 2 2( ) ( )n

    nn

    xt t x t n t P x

    3

  • 8/4/2019 L-6 de Series Solution

    66/88

    122 2 101 2 1 2( ) ( )n nnx t xt t xt t n t P x 2 1

    0 0

    1 2( ) ( ) ( )n nn nn n

    x t t P x xt t n t P x

    Equating the coefficients oftn on both sides, we get

    1 1 11 2 1( ) ( ) ( ) ( ) ( ) ( ) ( )n n n n nxP x P x n P x nxP x n P x 1 11 2 1 1( ) ( ) ( ) ( ) ( ) ( )n n nn P x n x P x n P x

    322 10

    1 2( ) ( ) ( )n

    nn

    x t xt t n t P x II

  • 8/4/2019 L-6 de Series Solution

    67/88

    Differentiating (I) with respect tox, we get

    322120

    1 2 2( ) ( )n

    n

    n

    xt t t t P x

    322

    0

    1 2 ( ) ( )n

    nn

    t xt t t P x III

    322 10

    1 2( ) ( ) ( )n

    nn

    x t xt t n t P x II

    Also , we haveFrom (II) and (III)

    1

    0 0

    ( ) ( )n n

    n nn n

    n t P x t P x

    x t t

    1

  • 8/4/2019 L-6 de Series Solution

    68/88

    1

    0 0

    ( ) ( ) ( )n n

    n nn n

    t n t P x x t t P x

    Equating the coefficients oftn on both sides, we get

    1 2( ) ( ) ( ) ( )n n nn P x xP x P x

    1 11 2 1 1( ) ( ) ( ) ( ) ( ) ( )n n nn P x n x P x n P x Differentiating the relation

    with respect tox, we get

    1 11 2 1 2 1( ) ( ) ( ) ( ) ( ) ( ) ( )n n n nn P x n P x n xP x n P x Substituting for x from (2)( )nP x

    1

    1 1

    12 1 2 1

    ( ) ( )( ) ( ) ( )[ ( ) ( )] ( )

    n

    n n n n

    n P xn P x n nP x P x n P x

    1 12 1 3( ) ( ) ( ) ( ) ( )n n nn P x P x P x

    O th lit f L d l i l

  • 8/4/2019 L-6 de Series Solution

    69/88

    Orthogonality of Legendre polynomials

    1

    1

    0 for

    2 for2 1

    ( ) ( )n m

    m n

    P x P x dx m nn

    Proof:

    We know that are solution of the

    Legendre equations of ordern andm respectively.

    and( ) ( )n mP x P x

    2

    2

    1 2 1 0 1

    and 1 2 1 0 2

    ( ) ( )

    ( ) ( )

    n n n

    m m m

    x P xP n n P

    x P xP m m P

    Therefore

    Multiplying (1) by , (2) by and subtracting, we get( )mP x ( )nP x

    21 2

    1 1 0

    ( )( ) ( )

    [ ( ) ( )]

    n m n m n m n m

    n m

    x P P P P x P P P P

    n n m m P P

  • 8/4/2019 L-6 de Series Solution

    70/88

    2or 1 1 0[( )( )] ( )( )n m n m n md

    x P P P P m n m n P Pdx

    Integrating with respect to x from -1 to 1, we get

    112

    11

    1 1 0( )( ) ( )( )n m n m n mx P P P P m n m n P P dx

    1

    11 0 3( )( ) ( )n mm n m n P P dx

    11

    Therefore 0 when( ) ( )n mP x P x dx m n

    When m = n, the relation (3) does not provide any information.

    When m= n, it is proved by using the

  • 8/4/2019 L-6 de Series Solution

    71/88

    generating function

    Squaring both sides and integrating w.r.t. xover [-1, 1],we obtain

    From the left hand side, we get

    n

    n

    ntxptxt )()21(

    0

    2

    1

    2

    .])([

    )21(

    21

    1 0

    1

    1

    2dxtxp

    txt

    dx n

    n

    n

    )]21ln()21[ln(2

    1

    ]

    2

    )21ln([

    2122

    1

    1

    21

    1

    2

    ttttt

    t

    txt

    txt

    dx

    ])1ln()1[ln(22

    tt

  • 8/4/2019 L-6 de Series Solution

    72/88

    )1.....](12

    .......53

    1[2

    ....)]2

    (......)2

    [(1

    )]1ln()1[ln(1

    ])1ln()1[ln(2

    242

    22

    n

    ttt

    tt

    tt

    t

    ttt

    ttt

    n

    From the right hand side , we have11

  • 8/4/2019 L-6 de Series Solution

    73/88

    (2)

    Comparing the coefficients of in equation (1) and (2),we get

    Hence we have the result.

    dxxpt

    dxtxptxp

    nn

    n

    n

    nn

    n

    nn

    )(

    )(])([

    1

    1

    2

    0

    2

    2

    0

    1

    1

    221

    1 0

    nt2

    12

    2)(

    1

    1

    2

    ndxxpn

    Ch b h P l i l

  • 8/4/2019 L-6 de Series Solution

    74/88

    Chebyshev Polynomials

    ( )nT xChebyshev Polynomials denoted by , are the solutions

    of the differential equation2 21 0 1( ) ( )x y x y n y 1x Singularities of this equation are

    0

    ( ) mmm

    y x a x

    0x We find a series solution about of the form

    This series solution is convergent for Substituting in

    (1), we get

    1| | .x

    2 2 1 2

    2 1 0

    1 1 0( ) ( ) m m mm m mm m m

    x m m a x x ma x n a x

    Solution

    21 1( ) ( )m m

    http://de%20series%20solution%20i.ppt/http://de%20series%20solution%20i.ppt/
  • 8/4/2019 L-6 de Series Solution

    75/88

    2

    2 2

    2

    1 0

    1 1

    0

    ( ) ( )m mm mm m

    m mm m

    m m

    m m a x m m a x

    ma x n a x

    20 2

    2

    1 0

    2 1 1

    0

    ( )( ) ( )m m

    m mm m

    m mm m

    m m

    m m a x m m a x

    ma x n a x

    22 3 1 0 1

    22

    2

    2 6

    2 1 1 0

    ( )

    [( )( ) ( ) ]m

    m m m mm

    a a x a x n a a x

    m m a m m a ma n a x

    Equating the coefficients of various powers ofx to zero, we get

    2 22 0 3 1 12 0 6 0a n a a a n a

  • 8/4/2019 L-6 de Series Solution

    76/88

    2 0 3 1 1

    22

    2 0 6 0

    2 1 1 0 for 2( )( ) ( )m m m m

    a n a a a n a

    m m a m m a ma n a

    2

    2 02

    na a 23 116

    na a 2 2

    2 0 for 22 1( )( )

    m m

    n m

    a a

    m m

    2 2

    4 22

    4 3

    na a

    2 2 2

    02

    4 3 2

    n na 2 2 2 0

    12

    4( )

    !n n a

    2 2

    5 33

    5 4

    na a

    2 2 2

    13 1

    5 4 6

    n na

    2 2 2

    11 3

    5

    ( )( )

    !

    n na

    and so on.

    Substituting in the assumed power series solution, we get

    2

    11

    3!n a

    2 2 2 2 2 40

    1 11 2( ) ( )y x a n x n n x

  • 8/4/2019 L-6 de Series Solution

    77/88

    0

    2 2 2 2 2 23 5

    1

    1 22 4

    1 1 3

    3 5

    ( ) ( ) .............! !

    ( )( )..........

    ! !

    y x a n x n n x

    n n na x x x

    The two series

    2 2 2 2 2 41

    1 11 2

    2 4( ) ( ) ............

    ! !y x n x n n x

    2 2 2 2 2 23 5

    21 1 3

    and3 5

    ( )( )( ) ..........

    ! !

    n n ny x x x x

    are two linearly independent solution of Chebyshev differential

    equation and converge for |x | 1.

    Forn= 0, 2, 4, ..,y1(x) reduces to the polynomials

    1 1( ) ,y x 21 1 2( ) ,y x x 2 41 1 8 8( ) , ..............y x x x

    Forn= 1, 3, 5, ..,y2(x) reduces to the polynomials

  • 8/4/2019 L-6 de Series Solution

    78/88

    , , , , y2( ) p y

    2 ( ) ,y x x

    32

    13 4

    3( ) ( ),y x x x

    3 52

    15 20 16

    5

    ( ) ( )y x x x x and so on.

    These polynomials give rise to important polynomials called

    Chebyshev Polynomials.

    To define Chebyshev Polynomials in different form

  • 8/4/2019 L-6 de Series Solution

    79/88

    we again consider the Chebyshev differential equation

    2 21 0( )x y x y n y cos ,x Put then dyy

    dx dy d

    d dx 1sin dyd

    2

    2

    d yy

    dx 1

    sin

    d dy d

    d d dx

    2

    2 21 1 cos

    sin sin sin

    d y dy

    dd

    2

    2 2 31 cos

    sin sin

    d y dydd

    Substituting in the differential equation

    22 2

    2 2 31 1 0cossin cos

    sinsin sin

    d y dy dy n yd dd

    2

    2

    20

    d yn y

    d

    The general solution of this differential equation is

  • 8/4/2019 L-6 de Series Solution

    80/88

    The general solution of this differential equation is

    ( ) cos siny A n B n 1 1

    ( ) cos( cos ) sin( cos )y x A n x B n x Therefore the general solution of Chebyshev differential equation is1 1Hence and are linerly independent

    soltuions.

    cos( cos ) sin( cos )n x n x

    When n is an integer, we define

    1 1(x)=( ) cos( cos ) sin( cos )n nT x n x U n x

    ( )nT x is Chebyshev polynomial of first kind of degreen.

    ( )nT x 1.

    1 1Show that 2 0( ) ( ) ( )n n nT x xT x T x

  • 8/4/2019 L-6 de Series Solution

    81/88

    1 where( ) cos( cos ) cos cos .nT x n x n x 1 1S ow t at 0( ) ( ) ( )n n nx x x x

    We have

    11

    11

    1 1

    1 1

    ( ) cos[( )cos ] cos( )

    ( ) cos[( )cos ] cos( )

    n

    n

    T x n x n

    T x n x n

    1

    1

    or ( ) cos( )cos sin( )sin

    ( ) cos( )cos sin( )sin

    n

    n

    T x n n

    T x n n

    Adding

    1 1 2( ) ( ) cos( )cosn nT x T x n 1 1 2( ) ( ) ( )n n nT x T x xT x 1 1or 2 0( ) ( ) ( )n n nT x xT x T x

    Using the recurrence relation

  • 8/4/2019 L-6 de Series Solution

    82/88

    g

    and the definition

    we obtain

    1 12 0( ) ( ) ( ) ,n n nT x xT x T x

    1 where( ) cos( cos ) cos cos .nT x n x n x 0 ( )T x 1 1( )T x x

    2 1 02( ) ( ) ( )T x xT x T x

    1 1or 2( ) ( ) ( )n n nT x xT x T x

    22 1x 3 2 12( ) ( ) ( )T x xT x T x 34 3x x4 3 22( ) ( ) ( )T x xT x T x 4 28 8 1x x 5 4 32( ) ( ) ( )T x xT x T x 5 316 20 5x x x

    Graphs of Chebyshev Polynomials

  • 8/4/2019 L-6 de Series Solution

    83/88

    -1 -0.5 0.5 1

    x

    -1

    -0.5

    0.5

    1

    Tn x

    T2

    T1

    T0

    G ap s o C ebys ev o y o a s

    20 1 21 2 1( ) ( ) ( )T x T x x T x x

    Graphs of Chebyshev Polynomials

  • 8/4/2019 L-6 de Series Solution

    84/88

    -1 -0.5 0.5 1

    x

    -1

    -0.5

    0.5

    1

    Tn x

    T5

    T4

    T3

    T2

    T1

    T0

    p y y

    0 1

    2 3

    2 34 2 5 3

    4 5

    1

    2 1 4 38 8 1 16 20 5

    ( ) ( )

    ( ) ( )( ) ( )

    T x T x x

    T x x T x x xT x x x T x x x x

    Problem:

  • 8/4/2019 L-6 de Series Solution

    85/88

    Express the polynomial in terms of Chebyshev

    polynomials of first kind

    4 35 3x x

    Solution:

    We have 0 12 3

    2 3

    4 2 5 3

    4 5

    1

    2 1 4 3

    8 8 1 16 20 5

    ( ) ( )

    ( ) ( )

    ( ) ( )

    T x T x x

    T x x T x x x

    T x x x T x x x x

    Therefore1( )x T x

    4 34 3 2 1 0

    15 3 [5 2 20 18 15 ]

    8x x T T T T T

    22 2 0

    1 11

    2 2[ ( ) ] [ ( ) ( )]x T x T x T x

    33 3 1

    1 13 3

    4 4

    [ ( ) ] [ ( ) ( )]x T x x T x T x 4 2

    4 4 2 0

    1 18 1 4 5

    8 8[ ( ) ] [ ( ) ( ) ( )]x T x x T x T x T x

  • 8/4/2019 L-6 de Series Solution

    86/88

    Exercise: Show that 120

    1 1 2( ) ( )n

    nn

    xt xt t t T x

    12Generating function of Chebyshev Polynomials

    1 1 2( )xt xt t

    Orthogonality of Chebyshev polynomials

  • 8/4/2019 L-6 de Series Solution

    87/88

    g y y p y

    1

    21

    if 0

    1

    if 02

    0 if

    ( ) ( ),

    ,

    ,

    n mT x T xdx m n

    x

    m n

    m n

    Proof:Case I: m = n = 0

    1 1

    2 21 1

    1

    1 1

    ( ) ( )n mT x T x dx dx

    x x

  • 8/4/2019 L-6 de Series Solution

    88/88

    Case II: m = n 0

    1 1 12 2 1

    2 2 21 1 11 1 1

    ( ) ( ) [ ( )] cos ( cos )n m nT x T x T x n x

    dx dx dxx x x

    Substitute , then obtain1cos x t 1 22

    12

    1

    [ ( )]nT x dx

    x

    Case III: m n

    1 1 1 1

    2 21 11 1

    ( ) ( ) cos( cos )cos( cos )n mT x T x n x m xdx dx

    x x

    Substitute , then obtain1cos x t 12

    0( ) ( )n mT x T x dx