numerical methods and statics

Upload: dch-narrasimhan

Post on 04-Jun-2018

218 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/13/2019 Numerical Methods and statics

    1/27

    MA2264 /MA1251 - NUMERICAL METHODS

    UNIT-I

    SOLUTION OF EQUATIONS AND EIGENVALUE PROBLEMS

    1. If a function f(x) = 0 is continuous in the interval (a, b) and if f (a) and f (b) are ofopposite signs. Then one of the root lies between a and b.

    2. xa!ple of Alge!"#$ e%&"'#()" (i) x#$ 2x % & = 0' (ii) 2x#$ #x $ = 0.

    #. xa!ple of T!")*$e)+e)'"l e%&"'#()" (i) x $ cosx = 0' (ii) xex2 = 0' (iii)012log10 =xx .

    *. Reg&l" F"l*# Me',(+")()(

    )()(

    afbf

    abfbafx

    =

    (+irst iteration of egula +alsi -ethod).

    &. I'e!"'#e Me',(+" )(1 nn xx =+ .

    . onvergence condition of iterative !ethod is 1)(

  • 8/13/2019 Numerical Methods and statics

    2/27

    1&. D#"g()"ll D(#)")'" 8n nn !atrix 8 is said to be diagonall do!inant if theabsolute value of each leading diagonal ele!ent is greater than or e3ual to the su! of the

    absolute values of the re!aining ele!ents in that row.

    4iven sste! of e3uations is 1111 dzcybxa =++ ' 2222 dzcybxa =++ '#### dzcybxa =++ is a diagonal sste! is if 111 cba +

    222 cab + ### bac +

    1. G"&** "$(# Me',(+" If the given sste! of e3uation is diagonall do!inant then

    ( ) ( )( )nnn zcybda

    x111

    1

    )1( 1 =+

    ( ) ( )( )nnn zcxad

    by 222

    2

    )1( 1 =+

    ( ) ( )( )nnn ybxad

    cz

    ###

    #

    )1( 1 =+

    1/. G"&** Se#+el Me',(+" If the given sste! of e3uation is diagonall do!inant then

    ( ) ( )( )nnn zcybda

    x 1111

    )1( 1 =+

    ( ) ( )( )nnn zcxad

    by 2

    1

    22

    2

    )1( 1 = ++

    ( ) ( )( )1#

    1##

    #

    )1( 1 +++ = nnn ybxadc

    z

    1. 6ufficient condition for iterative !ethods (4auss 6eidel -ethod > 4auss 5acobi -ethod)

    to convergence is the coefficient !atrix should be diagonall do!inant.

    1. The iteration !ethod is a self correcting !ethod since the round off error is s!aller.

    20. ?h 4auss 6eidel iteration is a !ethod of successive corrections@

    8ns" 9ecause we replace approxi!ations b corresponding new ones as soon the latter

    have been co!puted.

    21. o!pare G"&** El##)"'#() Me',(+and G"&** (!+") Me',(+

  • 8/13/2019 Numerical Methods and statics

    3/27

    22. I)e!*e ( " M"'!#7" Aet 8 be an nn nonsingular !atrix. If B is the inverse of the!atrix 8 then 8B = I (i.e.) B = I 81. ?e start with aug!ented !atrix of 8 with identit!atrix I of the sa!e order and convert 8 into the re3uired for! (i.e.) identit then the

    inverse is for!ed. 78 C I : 7I C 81:.

    2#. o!pare G"&** El##)"'#() Me',(+and G"&** Se#+el Me',(+.

    4auss li!ination -ethod 4auss 5ordan -ethod

    1. Direct -ethod

    2. oefficient !atrix is

    transfor!ed into uppertriangular !atrix.

    #. ?e obtain the solution

    b bac; substitution!ethod.

    1. Direct -ethod

    2. oefficient !atrix is transfor!ed

    into diagonal !atrix.#. Eo need of bac; substitution

    !ethod. 6ince finall this sste!

    of e3uation each has onl oneun;nown.

    4auss li!ination -ethod 4auss 6eidel -ethod

    1. Direct -ethod2. It has the advantage that it is finite

    and wor;s in theor for an non

    singular set of e3uation.#. ?e obtain exact value.

    1. Indirect -ethod2. It converges onl for

    diagonall

    do!inant.

    #. 8pproxi!ate value

    which is self correct!ethod.

  • 8/13/2019 Numerical Methods and statics

    4/27

    2*. o!pare G"&** "$(#and G"&** Se#+el Me',(+*.

    2&. ?h G"&** Se#+el e',(+is better !ethod than "$(#0* e',(+@8ns" 6ince the current value of the un;nowns at each stage of iteration are used in

    proceeding to the next stage of iteration, the convergence in 4auss 6eidel !ethod will be

    !ore rapid than in 4auss 5acobi !ethod.

    UNIT-II

    INTERPOLATION AND APPRO8IMATION

    2. xplain briefl I)'e!(l"'#(). 8ns" Interpolation is the process of co!puting the values of a function for an value of

    the independent variable within an interval for which so!e values are given.

    2/. Definition of I)'e!(l"'#()and e7'!"(l"'#()9

    4auss 5acobi -ethod 4auss 6eidel -ethod

    1. Indirect -ethod

    2. onvergence rate is slow.#. ondition for convergence is

    diagonall do!inant.

    1. Indirect -ethod

    2. The rate of convergence of this !ethod isroughl twice that of 5acobi.

    #. ondition for convergence is diagonall

    do!inant.

  • 8/13/2019 Numerical Methods and statics

    5/27

    8ns" I)'e!(l"'#()" It is the process of finding the inter!ediate values of a function fro!

    a set of its values specific points given in a tabulated for!. The process of

    co!puting corresponding to x 1,...2,1,0,1 =< 0 then the process is called extrapolation.

    2. 6tate Ne.'()0* F(!."!+ #)'e!(l"'#()for!ula.

    h

    xxuwherey

    n

    nuuuu

    yuuu

    yuu

    yu

    yxyuh

    n 00

    0

    #

    0

    2

    000

    F

    ))1()...(2)(1(...

    ....F#

    )2)(1(

    F2

    )1(

    F1)()=(x"8ns

    =

    +

    +

    +

    ++==+

    2. 6tate Ne.'()0* B"$;."!+ #)'e!(l"'#()for!ula.

    h

    xxpwherey

    n

    npppp

    yppp

    ypp

    yp

    yxyph

    nn

    n

    nnnn

    =

    ++++

    +++

    ++

    ++==+

    F

    ))1()...(2)(1(...

    ....F#

    )2)(1(

    F2

    )1(

    F1)()=(x"8ns #2n

    #0. rror in Ne.'()0* (!."!+"

    n

    n xxwherefn

    nuuuuAns

  • 8/13/2019 Numerical Methods and statics

    6/27

    =01

    01

    01

    01

    x

    ):g(x)7g(x

    x

    ):f(x)7f(x

    xx

    = f(x) g(x).

    #&. Divided difference table"

    B G G 2 G #G

    B0

    B1

    B2

    B#

    G0

    G1

    G2

    G#

    ),(

    )()(

    10

    01

    01

    xxf

    xx

    xfxf

    =

    ),(

    )()(

    21

    12

    12

    xxf

    xx

    xfxf

    =

    ),(

    )()(

    #2

    2#

    2#

    xxfxx

    xfxf

    =

    ),,(

    ),(),(

    210

    02

    1021

    xxxf

    xx

    xxfxxf

    =

    ),,(

    ,(),(

    #21

    1#

    21#2

    xxxf

    xx

    xxfxxf

    =

    ),,,(

    ,,(),,(

    #210

    0#

    210#21

    xxxxf

    xx

    xxxfxxxf

    =

    #. ?rite L"g!")g#")0* (l)(#"l (!&l".

    n

    nnnnnn

    n

    n

    n

    n

    n

    yxxxxxxxxxx

    xxxxxxxxxx

    yxxxxxxxx

    xxxxxxxxy

    xxxxxxxx

    xxxxxxxxyAns

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

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

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

    ))...()()((

    ))...()()((

    ))...()()(("

    1#210

    1#210

    1

    1#12101

    #200

    0#02010

    #21

    +

    +

    +

    =

    #&. ?hat is the assu!ption we !a;e when Aagrange

  • 8/13/2019 Numerical Methods and statics

    7/27

    8ns" Aet ( ))(, ii xfx , i = 0, 1, 2... n be the given (n %1) pairs of a data. The third ordercurves e!ploed to connect each pair of data points are called cubic splines. () 8

    s!ooth polno!ial curve is ;nown as cubic spline.8 cubic spline function f(x) w.r.t. the points x0, x1, .....xn is a polno!ial of

    degree three in each interval (xi1,xi) i = 1, 2, ...n such that )(xf , )(xf and )(xf are continuous.

    #. ?rite down the for!ula of C$ Sl#)e.

    ( )[ ]

    iiiii

    iiiiiii

    xxxMh

    yxxh

    Mh

    yxxh

    MxxMxxh

    xyAns

    +

    ++=

    1

    2

    1

    1

    2

    1

    #

    11

    #

    '

    )(1

    )(

    1)(

    1)("

    and [ ] )1....(#,2,12

    * 11211 =+=++ ++ niforyyyhMMM iiiiii

    where - = y ()

    =+++

    +

    ++++

    i

    ii

    i

    iiiiiiiii

    h

    ff

    h

    ffahahhah 1

    1

    11111 )(2 where 1= iii xxh

    and ii yf = ' i = 1,2,#,....

    UNIT < III

    NUMERICAL DIFFERENTIATION AND INTEGRATION

    #. Derivatives of G based on Ne.'()0* (!."!+ #)'e!(l"'#() (!&l""

    ( ) ( ) ( )

    ( )

    ++++

    +++++

    =

    ........2*10010&*0&

    F&

    1

    221*F*

    12#

    F#

    112

    F2

    1

    1

    0

    &2#*

    0

    *2#

    0

    #2

    0

    2

    0

    yuuuu

    yuuuyuuyuy

    hdx

    dy

  • 8/13/2019 Numerical Methods and statics

    8/27

    ( ) ( )

    +++++++

    = ........102112212

    1

    11112

    1

    )1(1 0&2#

    0

    *2

    0

    #

    0

    2

    22

    2

    yuuuyuuyuyhdx

    yd

    ( ) ( )

    +++++= ......../2*

    1

    #22

    1

    #1 0

    &2

    0

    *

    0

    #

    #

    #

    yuuyuyhdx

    yd whereh

    xxu 0

    =

    If 0xx = then u = 0

    +++=

    =

    .....&

    1

    *

    1

    #

    1

    2

    11 0

    &

    0

    *

    0

    #

    0

    2

    0

    0

    yyyyy

    hdx

    dy

    xx

  • 8/13/2019 Numerical Methods and statics

    9/27

    ++=

    =

    ........

    (

    &

    12

    111 0

    &

    0

    *

    0

    #

    0

    2

    22

    2

    0

    yyyy

    hdx

    yd

    xx

    ++=

    =

    ........*

    /

    2

    #

    1 0

    &

    0

    *

    0

    #

    ##

    #

    0

    yyyhdx

    ydxx

    *0. Derivatives of G based on Ne.'()0* "$;."!+ #)'e!(l"'#() (!&l""

    ( ) ( ) ( )

    ( )

    ++++++

    +++++++++

    =

    ........2*10010&*0&F&

    1

    221*F*

    12#

    F#

    112

    F2

    1

    1

    &2#*

    *2##22

    n

    nnnn

    ypppp

    ypppyppypy

    hdx

    dy

  • 8/13/2019 Numerical Methods and statics

    10/27

    ( ) ( )

    +++++++++++

    = ........102112212

    1

    11112

    1

    )1(1

    &2#*2#2

    22

    2

    nnnn yuupyppypyhdx

    yd

    ( ) ( )

    +++++++= ......../2*

    1

    #22

    1

    #

    1

    &2*#

    #

    #

    nnn yppypyhdx

    yd whereh

    xxp n

    =

    If nxx= then p = 0

    +++++=

    =

    .....&

    1

    *

    1

    #

    1

    2

    11

    &*#2

    nnnnn

    xx

    yyyyy

    hdx

    dy

    n

  • 8/13/2019 Numerical Methods and statics

    11/27

    ++++=

    =

    ........

    (

    &

    12

    111

    &*#2

    22

    2

    nnnn

    xx

    yyyy

    hdx

    yd

    n

    +++=

    =

    ........*

    /

    2

    #

    #1

    &*#

    #

    #

    nnn

    xx

    yyyhdx

    yd

    n

    *1. ?hat are the two tpes of errors involving in the nu!erical co!putation ofderivatives@

    8ns" (i) Truncation error' (ii) ounding error (To produce exact result is rounded tothe nu!ber of digits).

    *2. Define the error of approxi!ation.

    8ns" The approxi!ation !a further deteriorate as the order of derivative increases.The 3uantit (r)= f(r)(x) $ H( r )n(x) is called the error of approxi!ate in the r

    thorder

    derivative. ?here f(x) is the given e3uation and H(x) is approxi!ate values of f(x).

    *#. To find M"7#" and M#)#" of a tabulated function"

    Aet = f(x)

    +inddx

    dyand e3uate to ero. 8nd solving for x.

    +ind2

    2

    dx

    yd' If

    2

    2

    dx

    ydat x is $ve has !axi!u! at that point x.

  • 8/13/2019 Numerical Methods and statics

    12/27

    If2

    2

    dx

    ydat x is %ve has !ini!u! at that point x.

    ?e

  • 8/13/2019 Numerical Methods and statics

    13/27

    *. S#*()0* ()e-',#!+ !&le"

    rule

    #

    1

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

    +

    +=

    +++++++=

    ordinatesevenofSum

    ordinatesoddremainingofSumordinateslastandtheofSumh

    yyyyyyh

    dxxf

    st

    n

    x

    x

    n

    *

    21

    #

    ....*...2#)(#1*20

    0

    *. S#*()0* ',!ee e#g,', !&le"

    rule

    #

    ( ) ( ) ( )[ ]....*...#

    #)( #*210

    0

    ++++++++++= yyyyyyyyyh

    dxxfn

    x

    x

    n

    &0. rror in S#*()0* ()e-',#!+ !&le"

    { },.....,!ax> LLL2LLL0 yyMn

    abhwhere =

    =

    &1. The error in S#*()0* O)e ',#!+ !&leis of (!+e! *h .

    &2. ?hen does S#*()0* !&legive exact result@8ns" Holno!ials of degree 2 .

    . ?hen is T!"e=(#+"l !&leapplicable@8ns" +or an intervals.

    &*. ?hen is S#*()0*#

    1!&leapplicable@

    Mhab

    E10

    )( *

  • 8/13/2019 Numerical Methods and statics

    14/27

    8ns" ?hen there even no. of intervals.

    &&. ?hen is S#*()0*

    #!&leapplicable@

    8ns" ?hen the intervals are in !ultiples of three.

    &9 R(e!g0* Me',(+" =b

    a

    dxxfI )(

    +=#

    122

    IIII

    &/. 6tate R(e!g0* e',(+ #)'eg!"'#() (!&l"to find the value of =b

    a

    dxxfI )(

    using2

    handh .

    &. 6tate T.( P(#)' G"&**#") Q&"+!"'&!e (!&l":

    8ns"

    +

    ==

    #

    1

    #

    1)(

    1

    1

    ffdxxfI

    &. 6tate T,!ee P(#)' G"&**#") Q&"+!"'&!e (!&l":

    8ns"

    ( )

    ++

    ==

    &

    #

    &0

    &

    #

    &)(

    1

    1

    fffdxxfI

    0. I) G"&**#")0* Q&"+!"'&!e" If the li!it is fro! a to b then we shall appl a suitable

    change of variable to bring the integration fro! 1 to 1

    ( ) ( )

    dtab

    dxabtab

    xPut2

    '2

    =++= 9

    =

    +=

    hh

    hh

    h II

    II

    Ih

    hIAns2

    2

    2

    *#

    1

    #2,"

  • 8/13/2019 Numerical Methods and statics

    15/27

    1. 6tate T!"e=(+#"l (!&l" (! D(&le I)'eg!"l*:

    ( )dxdyyxfId

    c

    b

    a= ,

    ( )

    ( )

    ( )

    +

    +=

    nodeseriortheatfofvaluestheofSum

    boundarytheonnodesremainingtheatfofvaluestheofSum

    cornersfourtheatfofvaluesofSumhk

    I

    int*

    2*

    2. 6tate S#*()0* !&le (! D(&le I)'eg!"l*:

    ( )dxdyyxfId

    c

    b

    a

    = ,

    ( )( )( )( )

    ( )( )( )

    ++

    +

    +

    +

    +

    =

    matrixtheofroweventheonpositionseventheatfofvaluestheofSum

    positionsoddtheatfofvaluestheofSum

    matrixtheofrowoddtheon

    positionseventheatfofvaluestheofSum

    positionsoddtheatfofvaluestheofSum

    boundarytheonpositionseventheatfofvaluestheofSum

    boundarytheonpositionsoddtheatfofvaluestheofSum

    cornersfourtheatfofvaluesofSum

    hkI

    1(

    *

    *

    2

  • 8/13/2019 Numerical Methods and statics

    16/27

    #. ?h is Trapeoidal rule so called@

    8ns" 9ecause it approxi!ates the integral b the su! of the areas of n trapeoids.

    *. o!pare Trapeoidal rule and 6i!pson

  • 8/13/2019 Numerical Methods and statics

    17/27

    8 !ethod that uses values fro! !ore than one preceding step is called

    !ultistep !ethod.

    /. 6tate T"l(! *e!#e* (!&l""

    0L

    0

    *LLL

    0

    #LL0

    2L01 .......

    F*F#F2F1)( xxhwherey

    hy

    hy

    hy

    hyxy v =++++==

    . ?hat are the !erits and de!erits of Talor

  • 8/13/2019 Numerical Methods and statics

    18/27

    ( )

    ( )#*

    2#

    12

    1

    ,2

    ,

    2

    2,

    2

    ,

    kyhxhfk

    ky

    hxhfk

    ky

    hxhfk

    yxhfk

    nn

    nn

    nn

    nn

    ++=

    ++=

    ++=

    =

    /2. 6tate F(&!', O!+e! R&)ge-@&''" Me',(+ (!&l""(for +irst order 6i!ultaneous

    differential e3uations)

    Aet ( ) ( )zyxfdx

    dzandzyxf

    dx

    dy,,,, 21 ==

    yyy nn +=+1 > zzz nn +=+1

    ?here ( )*#21 22

    1kkkky +++= and ?here ( )*#21 22

    1llllz +++=

    ( )

    ( )#1*

    21#

    112

    11

    ,

    2,2

    2,

    2

    ,

    kyhxhfk

    ky

    hxhfk

    ky

    hxhfk

    yxhfk

    nn

    nn

    nn

    nn

    ++=

    ++=

    ++=

    =

    ( )

    ( )#2*

    22#

    122

    21

    ,

    2,2

    2,

    2

    ,

    kyhxhfl

    ky

    hxhfl

    ky

    hxhfl

    yxhfl

    nn

    nn

    nn

    nn

    ++=

    ++=

    ++=

    =

    /#. 6tate F(&!', O!+e! R&)ge-@&''" Me',(+ (!&l""(for second order differential

    e3uations)

    Aet ( ) ( )zyxfdx

    dz

    dx

    ydandzzyxf

    dx

    dy,,,, 22

    2

    1 ====

    yyy nn +=+1

    ?here ( )*#21 22

    1kkkky +++= and ?here ( )*#21 22

    1llllz +++=

  • 8/13/2019 Numerical Methods and statics

    19/27

    ( )

    ( )#1*

    21#

    112

    11

    ,2

    ,

    2

    2,

    2

    ,

    kyhxhfk

    ky

    hxhfk

    ky

    hxhfk

    yxhfk

    nn

    nn

    nn

    nn

    ++=

    ++=

    ++=

    =

    ( )

    ( )#2*

    22#

    122

    21

    ,2

    ,

    2

    2,

    2

    ,

    kyhxhfl

    ky

    hxhfl

    ky

    hxhfl

    yxhfl

    nn

    nn

    nn

    nn

    ++=

    ++=

    ++=

    =

    /*. uler

  • 8/13/2019 Numerical Methods and statics

    20/27

    /. 8da!

  • 8/13/2019 Numerical Methods and statics

    21/27

    &. ?h ungePutta for!ula is called fourth order@

    8ns" The fourth order ungePutta !ethod agree with Talor series solution up to

    the ter!s of h*. Nence it is called fourth order .P. !ethod.

    . E!!(!" 7M#l)e0* P!e+#$'(!

    )(*&

    1* )&(1

    & = nyhE where 1# + nn xx

    M#l)e0* C(!!e$'(!

    )(01 )&(1&

    = nyhE where 11 + nn xx

    /. R(&)+ ( e!!(!: ?hen we are wor;ing with deci!al nu!bers. ?e approxi!ate thedeci!als to the re3uired degree of accurac. The error due to

    these approxi!ations is called round off error.

    T!&)$"'#() e!!(!: The error caused b using approxi!ate for!ula in co!putationsis ;nown as truncation error.

    . E!!(!: A+"0* P!e+#$'(!

    ( ) 1#)&(&

    /20

    2&1+ = nn xxwhereyhE

    E!!(!: A+"0* C(!!e$'(!

    ( ) 12)&(&

    /20

    1+

    = nn xxwhereyhE

    . o!pare the -ilne

  • 8/13/2019 Numerical Methods and statics

    22/27

    2. It does not have the sa!e

    instabilit proble!s asthe -ilne !ethod.

    2. 9ut is about as efficient.

    #. 8 !odification of8da! AND PARTIAL

    DIFFERENTIAL EQUATIONS

    0. Define B(&)+"! "l&e !(le. 8ns" ?hen the differential e3uation is to be solved satisfing the conditions

    specified at the end points of an interval the proble! is called boundar valueproble!.

    1. Define D#e!e)$e Q&('#e)'*

    8ns" 8 difference 3uotient is the 3uotient obtained b dividing the differencebetween two values of a function b the difference between two corresponding

    values of the independent variable. ( )11L

    2

    1+ = iii yy

    hy

    2. F#)#'e D#e!e)$e Me',(+*"

    ( )11L

    2

    1+ = iii yy

    hy

    ( )112LL

    21

    + += iiii yyyh

    y

    #. lassification of Hartial Differential 3uations of the 6econd rder

    0,,,,2

    22

    2

    2

    =

    +

    +

    +

    y

    u

    x

    uuyxf

    y

    uC

    yx

    u

    x

    uA

    (i) If 0*2 AC then the e3uation is Nperbolic.

  • 8/13/2019 Numerical Methods and statics

    23/27

    (iii) If 0*2 = AC then the e3uation is Harabolic.

    *. Be)+e!-S$,#+'0* D#e!e)$e E%&"'#()" (xplicit -ethod)2

    22

    x

    u

    t

    u

    =

    1, +!iu

    !iu ,1 !iu , !iu ,1+

    ()

    !iu ,1 !iu , !iu ,1+

    1, +!iu

    2,1,,11, )21(ahk"hereuuuu !i!i!i!i =++= ++ and 21

    =a

    Be)+e!-S$,#+'0* D#e!e)$e E%&"'#()

    If then2

    1=

    ( )!i!i!i uuu ,1,11, 21 ++ +=

    This is valid onl if2

    2h

    ak=

  • 8/13/2019 Numerical Methods and statics

    24/27

    &. C!");-N#$,(l*()0* D#e!e)$e E%&"'#()" 7I!plicit -ethod:2

    22

    x

    u

    t

    u

    =

    ( ) ( ) !i!i!i!i!i!i

    uuuuuu,1,1,1,1,11,1

    1212 +++++ +=++

    ?here 2ah

    k= and 2

    1

    =a

    C!");-N#$,(l*()0* D#e!e)$e E%&"'#()?hen 1= (i.e.) 2ahk= the ran;Eicholson

  • 8/13/2019 Numerical Methods and statics

    25/27

    100. 6tate the explicit sche!e for!ula for the solution of the wave e3uation.

    8ns" The for!ula to solve nu!ericall the wave e3uation ttxx uua =2

    is

    ( ) ( ) ,,1,122

    ,

    22

    1, 12 ++ ++= !i!i!i!i!i uuuauau

    ?hena

    hkei

    h

    k

    a=== ).(

    12

    2

    2

    2

    The wave e3uation is

    1, !iu

    and2

    0)0,( 0,10,11, + +== iiit uuuxu !iu ,1 !iu , !iu ,1+

    1, +!iu

    101. Define the local truncation error.

    8ns" =E( )

    +

    + ++i

    ii

    ii

    iii yh

    yyfy

    h

    yyy

    2

    2 112

    11

    102. 6tate finite difference approxi!ation for2

    2

    dxyd and state the order of truncation

    error.

    8ns" =iy ( )

    2

    11 2

    h

    yyy iii + + and the order of truncation error is (h2).

    10#. +orward +inite difference for!ula"

    h

    yxuyhxuyxux

    ),(),()( 00000,0

    +=

    k

    yxukyxuyxuy

    ),(),()( 00000,0

    +=

    Truncation error is ( ) hxxwhereyuh

    xx + 000,2

    ,,1,11, ++ +=

    !i!i!i!i uuuu

  • 8/13/2019 Numerical Methods and statics

    26/27

    10*. 9ac;ward +inite difference for!ula"h

    yhxuyxuyxux

    ),(),()( 00000,0

    =

    k

    kyxuyxuyxuy

    ),(),()( 00000,0

    =

    Truncation error is ( ) hxxwhereyuh

    xx + 000,2

    10&. 6econd order finite difference for!ulae"

    ( ) ( ) ( )2

    0000000000

    0,0

    ,,2,),(),()(

    h

    yhxuyxuyhxu

    h

    yxuyhxuyxu xxxx

    ++=

    +=

    ( ) ( ) ( )2

    0000000000

    0,0

    ,,2,),(),()(

    k

    kyxuyxukyxu

    k

    yxukyxuyxu

    yy

    yy

    ++=

    +=

    Truncation error is ( ) hxhxwhereyuh

    xxxx + 0002

    ,12

    10. Ea!e at least two nu!erical !ethods that are used to solve one di!ensional

    diffusion e3uation.

    8ns' (i) 9ender6ch!idt -ethod (ii) ran;Eicholson -ethod.

    10/. 6tate standard five point for!ula for solving uxx% u = 0.

    8ns"

    10. 6tate diagonal five point for!ula for solving uxx% u = 0.

    8ns"10. ?rite down one di!ensional wave e3uation and its boundar conditions.

    8ns"2

    22

    2

    2

    x

    u

    t

    u

    =

    9oundar conditions are

    (i) u(0,t) = 0(ii) u(l, t) = 0 0t

    (iii) u(x,0) = f(x), 0QxQl

    (iv) ut(x,0) = 0, 0QxQl

    110. 6tate the explicit for!ula for the one di!ensional wave e3uation with

    m

    #aand

    h

    kwherea === 222 01 .

    8ns"

    !i!i!i!i!i uuuuu ,1,1,,1,1 *=+++ ++

    !i!i!i!i uuuu 1,11,11,11,1 *=+++

    ++++

    ,,1,11, ++ +=

    !i!i!i!i uuuu

  • 8/13/2019 Numerical Methods and statics

    27/27

    111. ?rite down the finite difference for! of the e3uation ),(2 yxfu= 8ns"

    112. ?rite Aaplace