fe1007 lecture 2c

Upload: jean-chan

Post on 04-Jun-2018

219 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/13/2019 Fe1007 Lecture 2c

    1/22

    Part III

    -

    123

    Power Series Method

    Methods for solving some special higher-order equations

    D-operator method (the most importantmethod we would discuss in this part)

    124

    Reduction of order by substitution

    Power series method

  • 8/13/2019 Fe1007 Lecture 2c

    2/22

    F(D) e kx = F(k) e kx

    Theorem 1:

    If k is a constant,

    To prove this, we note that

    Theorems Involving D-Operator :

    125

    kx kxd e kedx

    =

    D e kx = k e kx

    .

    .

    D2e kx = D(De kx) = D(ke kx) = k2e kx

    Dne kx = kne kx

    Consequently:

    F(D) e k = (a 0Dn + a 1Dn1 + + a n1 D + a n) e kx

    = (a 0kn + a 1kn1 + + a n1 k + a n) e kx

    = F(k) e kx

    126

    (1) (6D2 + 4 D 2) e 3x = (6 32 + 4 3 2) e 3x

    = 64 e 3x

    (2) (D2 + 1) e 2x = (22 + 1) e 2x

    = 5 e 2x

    Examples :

  • 8/13/2019 Fe1007 Lecture 2c

    3/22

    Theorem 2 ( Shifting Theorem) :

    If k is a constant and g(x) is a function of x, then :

    F(D) [e kx g(x)] = e kx F (D + k) g(x)

    To prove this, we note that :

    D [e kx g(x)] = e kx D g(x) + ke kx g(x)

    127

    = e kx (D + k) g(x)

    D2 [ekx g(x)] = D [ e kx (D + k) g(x) ]

    = e kx D(D + k) g(x) + ke kx (D + k) g(x)

    = e kx (D + k) (D + k) g(x)= e kx (D + k) 2 g(x)

    n n n 1 n 1

    n

    n nD (uv) (D u)v (D u)Dv ... (Du)D v

    1 n 1

    n

    n n! uD v

    n k k!(n k)!

    = + + +

    +

    =

    Dn [ekx g(x)] = ? Leibniz's Theorem states that :

    128

    n n 1 n 1

    n (D u)v n (D u)Dv ... n (Du)D v

    uD v

    = + + +

    +

    3 3 2 2 3D (uv) (D u)v 3 (D u)Dv 3 (Du)D v uD v= + + +

  • 8/13/2019 Fe1007 Lecture 2c

    4/22

    Dn [g(x) e kx]

    = e kx Dn g(x) + n D (e kx) Dn1 g(x) +

    + n D n1 (e kx) D g(x) + (D n e kx) g(x)

    Replacing V by e kx and U by g(x) , we have :

    129

    = e kx Dn g(x) + n ke kx Dn1 g(x) +

    + n k n1 e kx D g(x) + k n e kx g(x)

    = e kx ( Dn + nk D n1 + + nk n1 D + k n ) g(x)

    = e kx (D + k) n g(x)

    2

    ( 6D 2 + 4D 2) [e3x x2] = e 3x [ 6 (D + 3) 2 + 4(D + 3) 2] x2

    = e 3x ( 6D 2 + 40D + 64 ) x2

    Example: F(D) [e kx g(x)] = e kx F(D + k) g(x)

    130

    x2e 6 40 64 xdxdx

    = + +

    = e 3x (6 2 + 40 2x + 64 x 2)

    = e 3x (12 + 80x + 64 x 2)

  • 8/13/2019 Fe1007 Lecture 2c

    5/22

    Consider the following ODE, (find y as a function of x) :

    Solving ODE Using D-Operator :

    (D2 + 1) y = e 2x (1)

    131

    1LHS : F(D) y = y ;

    F(D)

    =

    F(D) 0

    Define the inverse D-operator, such that :1

    F(D)

    ,

    1RHS : g(x)

    F(D)

    2x1 eF(D)

    F(D) e kx = F(k) e kx (2)

    To solve (1), we recall Theorem 1 :

    Apply the inverse operator on both sides of eqn (2)and we get :

    =kx kx1 = kx1

    F k e

    132

    = kx kx ; F(k)1 1

    e eF

    0(D)

    F(k)

    (3)

    which gives :

    F(D)

    F(D)

  • 8/13/2019 Fe1007 Lecture 2c

    6/22

  • 8/13/2019 Fe1007 Lecture 2c

    7/22

    x2x22p e)D(F

    1e)3D4D(

    1y =++=

    (D2 + 4D + 3) y = e 2x

    The particular solution is given by :

    Example:

    135

    x2x22p

    e151

    e)3242(

    1y =

    ++=

    3x x 2x1 2

    1y c e c e e

    15

    = + +

    ,

    General solution is

    2x 2xp 3 2

    1 1y = e = eF(D)(D +4D +3D)

    (D3 + 4D 2 + 3D) y = e 2x

    The particular solution is given by :

    Another Example:

    136

    2x 2x

    p 3 21 1

    y = e = e30(2 +4 2 +3 2)

    = + + +3x x 2x0 1 21

    y c c e c e e30

    ,

    General solution is

  • 8/13/2019 Fe1007 Lecture 2c

    8/22

    Theorem 3:

    If k is a constant,

    F(D 2) sin (kx) = F( k2) sin (kx) and

    F(D 2) cos (kx) = F( k2) cos (kx)

    137

    where F(D 2) is a polynomial operator with D replaced byD2.

    Example : F(D 2 ) = (D 4+ 3D 2 1)

    D2 sin(3x) = D (3cos(3x)) = 3 2sin(3x)

    D2 cos(3x) = D (3sin(3x)) = 3 2cos(3x)

    (1) (D4 + 3D 2 1) sin (2x) = ((D 2)2 + 3D 2 1) sin (2x)

    = (22)2 + 3( 22) 1) sin (2x)

    = 3 sin (2x)

    Apply to the LHS, we get sin(2x).1

    138

    )8()x2sin(31

    )x2sin()1D3D(

    124 =+

    Therefore:

    +

    Apply it to the RHS , We get+ 4 2

    1 3sin(2x).

    (D 3D 1)

  • 8/13/2019 Fe1007 Lecture 2c

    9/22

    (2) (D 4 2D 2) cos (2x)

    = ((22)2 2(22)) cos (2x)

    = 24 cos (2x)

    139

    Therefore:

    )x2cos(241)x2cos(

    )D2D(

    124

    =

    xhy Ae

    =

    (3) (D + 1)y = 10 sin (2x)

    + = =+ +p1 1

    (D 1)y y 10 sin(2x)D 1 D 1

    140

    2(D 1)

    10sin(2x)D 1

    =

    210(D 1)

    sin(2x)D 1

    =

  • 8/13/2019 Fe1007 Lecture 2c

    10/22

  • 8/13/2019 Fe1007 Lecture 2c

    11/22

    21

    sin(4x)(D 5D 6) +

    1 sin(4x)

    ( 10 5D)=

    Therefore, y p is given by :

    p 100y sin(4x)( 10 5D)=

    20 sin(4x)

    2 D=+

    143

    220 (D 2)

    sin(4x)D 4

    =

    20 (D 2)sin(4x)

    20=

    (D 2)sin(4x)= Dsin(4x) 2sin(4x)=

    4cos(4x) 2sin(4x)=

    2(D 2D)y 24x ;+ =(5)

    x)2D(D

    24yp += x

    )2D(21

    D21 24

    +=

    2xh 1 2y ec c

    = +

    21 2r 2r 0 ; r 0 & r 2+ = = =

    144

    x 2D 1D1 12 +=

    )x(fxD1 Let = x Df(x) =

    dor x f(x)

    dx=

    2xIntegrating both sides gives : f(x) xdx

    2= =

  • 8/13/2019 Fe1007 Lecture 2c

    12/22

    =

    + +

    1 1 1x x

    DD 2 2 12

    x ...8

    D4

    D2D1

    21 32

    ++=

    1 1 x 0....2 2 = +

    2 31 D D D x x x x ...

    2 2 4 8

    = + +

    145

    3x6x6 2 +=

    ph yyy +=

    2p

    x x 1 y 12

    2 2 4

    = +

    43 xy)1D(D =(6)

    43p

    x)1D(D

    1y

    =

    43

    2x

    1D1

    D

    1DD

    +=

    146

    4432

    x D11x

    D

    1

    D

    1D1

    +=

    210x

    30x

    5x 765 = 4432 x ]DDDD1[ +++++

  • 8/13/2019 Fe1007 Lecture 2c

    13/22

    5 6 7

    px x x5 30 210

    y = 24x24x12x4x 234

    r3 (r 1) = 0 r = 0, 0, 0, 1

    43 xy)1D(D =

    147

    yh = (c 1 + c 2x + c 3x2) e 0 + c 4e x

    = (c 1 + c 2x + c 3x2) + c 4e x

    h py y y= +

    0qppr2r 222 =+++

    + + + =2 2 2 ax(D 2pD p q )y e(7)

    where p, q and a are non-zero constants.

    q4p4p4p2r222 +=

    148

    qip +=

    ( )pxh 1 2y e cos(qx) sin(qx)c c= +

    qipr2 =

  • 8/13/2019 Fe1007 Lecture 2c

    14/22

    Particular solution can be obtained by the undeterminedcoefficient method.

    Trial solution : yp = C eax

    D-operator Method :

    ax1

    149

    p 2 2 2D 2pD p q=

    + + +

    ax222

    eqpap2a

    1

    +++=

    ph yyy +=

    SummaryD -Operator method

    F(D)e kx=F(k)e kx

    150

    F(D) [e kxg(x)] = e kxF(D+k) g(x)

    F(D 2) sinkx = F(-k 2) sinkx

    F(D 2) coskx = F(-k 2) coskx

  • 8/13/2019 Fe1007 Lecture 2c

    15/22

    Methods for solving some special higher-order equations

    D-operator method (the most importantmethod we would discuss in this part)

    151

    Reduction of order by substitution

    Power series method

    Reduction of Order

    dxdPy =

    22

    dP dP P or dx

    dx P = =

    2y (y ) =(1)

    =Let y P

    152

    1 x CP

    = + 1Px C

    = +

    1y ln(x C) C= + +

    dy 1 1 ; dy dx

    dx x C x C= =

    + +

  • 8/13/2019 Fe1007 Lecture 2c

    16/22

    P2dxdPx =

    xdx2

    PdP =

    2 ln P ln x C = + 21or P C x=

    y2yx =(2) = =dP

    , yy Pdx

    Let :

    153

    212

    d yi.e., C x

    dx= 2

    31 CxC3

    1dxdy +=

    324

    1 CxCxC121y ++=

    324

    4 CxCxC ++=

    Power Series Method :Solve the differential equation:

    y' = x + ysubject to the condition y(1) = 0

    The Taylor series for y(x), in powers of x 1, is :

    y"(1) = 1 + y'(1) = 2

    y(1) = 0 , y' = x + y , y(1)' = 1 + y(1) = 1

    (n)(1)y ny(1) + y'(1)(x - 1) + ... + + ...(x-1)n!

  • 8/13/2019 Fe1007 Lecture 2c

    17/22

    y" = 1 + y' y'" = y"

    y'"(1) = y"(1) = 2

    (n+1) (n)y (x) = y (x) = 2 for n 2 Recurrence Eqn

    155

    )1x(!3

    12)1x(!2

    12)1x(10)x(y 32 ++++=

    =

    +

    2n

    n

    !n)1(x2)1x(=

    xn

    n 2

    1 (x 1)2 2 1 (x 1)n

    e!

    =

    = + +

    n

    nxn 2

    xSince : 2e 2 1 xn!

    =

    = + +

    156

    n

    2

    x

    n

    1 (x 1)(x 12e 2 (x ) 2n!

    1)

    =

    = +

    x 1 y(x) 2e 2 (x 1) =

    n 2

    n!x

    =+ = +

  • 8/13/2019 Fe1007 Lecture 2c

    18/22

    Example: Solve the initial value problem

    Assume solution of the form :

    y = a 0 + a 1x + a 2x2 + ... + a n1 xn1 + a nxn + ... (1)

    y' y = x , y(0) = 1

    157

    y' = a 1 + 2a 2x + 3a 3x2 + ... + na nxn1 + ... (2)

    y' y = (a 1 a 0) + (2a 2 a 1)x + (3a 3 a 2) x2

    + ... + (na n a n1)xn1 + ... (3)

    Comparing (3) with y' y = x , we get :

    2a 2 a 1 = 1

    3a 3 a 2 = 0

    na n a n1 = 0

    a 1 a 0 = 0 a 1 = a 0

    From (1) : y(0) = a 0 + 0 + ... = 1

    158

    a 0 = 1 , a 1 = 1

    12a1a 12 =

    +=

    !32

    232

    31

    3aa 23 ==== 4 n

    2 2, a , a

    4! n!= =

  • 8/13/2019 Fe1007 Lecture 2c

    19/22

    2 3 nx x xy 1 x 2 2 2

    2! 3! n! = + + + + + +

    !nx !3x!2x2x1 n32

    ++++++=

    159

    )x1e(2x1 x ++=

    x1e2y x =

    Summary Reduction of order :

    By substitution, the order of the equation may belowered (mostly applies to equations without zero-order derivative item, such asy""-y"= g(x) )

    160

    Power series method Mostly by using the Taylor power series and building

    up a relationship between lower-order derivatives andhigher-order derivative. For example, from y"-y=2, wehave y"'-y'=0,y""-y"=0, etc.)

  • 8/13/2019 Fe1007 Lecture 2c

    20/22

    Final Review: first-order equations (1)

    Variable separable equation (the basic case)

    M(x)dx + N(y)dy = 0

    Methods for transforming other equations into variable

    161

    dy = F(ax + by)dx

    separable equations:

    Let v ax by= +

    dy yF

    dx x

    =

    yLet v

    x=

    Exact Differential Equations:

    Final Review: first-order equations (2)

    Linear Equations: dy Py Qdx

    + =1PdxLet = e , y = Qdx

    Bernoullis Equations: ndy Py Qydx

    + = 1 nLet z y =

    162

    M(x,y)dx N(x,y)dy 0, where y x+ = =

    Solution: F(x,y)=c

    where F/ x=M(x,y), and F/ y=N(x,y)

    Applications of ODE (Based on the meaning of Derivate:Derivate measures the rate of change)

  • 8/13/2019 Fe1007 Lecture 2c

    21/22

    Final Review: second-order equations (1)

    Linear second-order homogenous ODE (the basic case)a. Two equal/unequal real number roots

    1 2r x r xh 1 2 1 2y c e c e , when r r= +

    1r xh 1 2 1 2y (c c x)e , when r r= + =

    163

    . x

    h 1 2y e (c cos x c sin x)= +

    Linear second-order non-homogenous equations withconstant coefficients

    General solution=general solution of homogenous equation+ particular solution

    Variation of parameters:

    Particular Solution: p 1 1 2 2y u v u v= +where

    Final Review: second-order equations (2)Methods for finding particular solutions :

    164

    2

    11 2 1 2v ' u u ' u 'u

    =

    12

    1 2 1 2

    u Fv '

    u u ' u 'u=

  • 8/13/2019 Fe1007 Lecture 2c

    22/22

    Undetermined coefficients (3 special F(x) functions):

    Three different cases, depending on whether 0 is

    a root of the characteristic equation.

    2F(x) ax bx c := + +

    rxF(x) e :=

    Final Review: second-order equations (3)

    165

    F(x) sinkx, coskx :=

    ree eren cases e , xe , x e , epen ngon whether r is a root of the characteristic equation.

    Two different cases as below, depending on whetherki is a root of the characteristic equation.

    Bcoskx Csinkx; Bxcoskx Cxsinkx+ +

    D-Operator method

    F(D) e kx = F(k)e kx

    Shifting method: F(D) [e kxg(x)] = e kxF(D+k) g(x)F(D 2) sinkx = F(-k 2) sinkx;

    2 = - 2

    Final Review: second- and higher-order equations

    166

    Reduction of order by substitution

    Power series method