Kap2: EULERS FORMULA: e i = cos +i sin .

. Or n writting in complex numbers: ( ei ) = cos n + i sin n i , a b= e bln a Z = x +iy = r ( cos +i sin )= r e 2 3 n z z z 1/ n= r cos + i sin . e z =1+ z + + ... n n 2! 3!

ax . /f t#e roots are 3+i D t#en t#e to t#e nearest singular point. a =b = a , y=( Ax+ B ) e 3x iD x i D x ")7O87> A: @A<@)RM9 ")7O87> . 5et @ be asmople close% solution is y =e ( A e + B e ) . or into forms li0e ax ax y =e ( c1 sin D x +c 2 cos D x )= y =c e sin (D x+ E ) . 97@O6:( curve wit# a continuousl. turning tangent except possibl. at a finite number of points (t#at is, we allow a finite number of O8:78 5/67A8 7;<A"/O6 4/") @O69"A6" corners, but ot#erwise t#e curve be smoot# . /f f(O is anal.tic on @O7**/@/76" A6: 8/F)" 9/:7 6O" G78O: f ( z ) dz =' an% insi%e @, t#en. /f c is not e2ual to eit#er a or bH C e cx , if c e2uals a or b, around C cx 2 cx Kap7:FOURIER TRANSFORM: Average of f(x on (a, b a b H , /f c!a!bH . T' ()*+ a C xe C x e ")7O87> T: @A<@)RM9 /6"7F8A5 *O8><5A . /f f(O is b f ( x ) dx ."#e average value of sin$2 nx an% cos$2 nx pa,-)./la, 0'l/-)'* '( ( D a )( Db ) y = sin 3 x (),0- 0'l1e anal.tic on insi%e a simple close% curve @, t#e value of f(O at a cos 3 x a point O!a insi%e @ is given b. t#e following contour integral along ! i3 x f (z ) b a 1 . an% t#en ta0e t#e real or imaginar. p. ( D a )( Db ) y = e f ( a )= dz . "#is is @auc#.Ms integral formula. @: & & cx 2 & i z a y p of ( D a)( D b) y =e P n ( x ) 1 1 A particular solution 2 2 sin nx dx= cos nx dx=1 /2 6ote carefull. t#at t#e point a is insi%e @H if a were outsi%e @, t#en 2& 2 pi P ( x ) w#ere is pol.nomial of %egree n is: if c /s not e2ual to & & n K ( z ) woul% be anal.tic ever.w#ere insi%e @ an% t#e integral cx "#e average value of sin mx cos nx over a perio% is: a b H eit#er a or bH y p=e Q n( x ) , if c e2ual a or b, woul% be Oero b. cauc#.Ms t#eorem. /f t#e values of f(O are given & ec 2 cx 1 y p=x e Q n( x ) if c! a!b. y p=x e Q n ( x ) . 4#ere on t#e boun%ar. of a region @, t#en t#e value of f(O at an. point a sin mx cos nx dx =' 2& Q ( x ) P ( x ) is pol.nomial of t#e same %egree as wit# & insi%e @. 4it# t#is interpretation .ou will fin% t#e cauc#.Ms integral n n un%etermine% coefficients to be foun% to satisf. t#e given & formula written wit# a replace% b. O, an% O replace% b. some ' m n 1 %ifferential e2uation. 6ote t#at sines an% cosines are inclu%e% in %ifferent %umm. integration variabl, sa. w. sin mx sin nx dx = 1/ 2 ' m=n ' cx 2& & b. use of complex exponentials. f (() e 1 ' m =n = ' f ( z )= d( . . O")78 97@O6:(O8:78 7;<A"/O69: /n case of %epen%ent 1 2 & (z & ' m n y ' = p , y' ' = p ' . /n%epen%ent variable x 5A<876" 978/79: 5et @S1 an% @S2 be two circles wit# center 1 variable . missing cos mx cos nx dx = 1/ 2 ' m= n ' . "#e fourier(coeff. dp dp dy dp 2& & at OS'. 5et f(O be anal.tic in t#e region 8 between t#e circles. y ' = p , y' ' = = =p missing . "o solve 1 m = n =' dx dy dx dy "#en f(O can be expan%e% in a series of t#e form 1 y ' ' + f ( y )=', y' y ' ' + f ( y ) y' =', y' dy ' + f ( y ) dy=' b b2 2 1: f ( x )= a '+a 1 cos x+ a 2 cos 2x ... b1 sin x+ b2sin 2x... An% f ( z )=a '+a1( z a' )+a2 ( z z ') +...+ 1 + ... conver 2 1 2 z z ' ( z z ')2 & y' + f ( y ) dy=const . 1 2 an= f ( x )cos nx dx an% bn is solving a n gent in 8. 9uc# a series is calle% a 5aurent series. "#e PbQ series is && 5A?5A@7 "8A69*O8>: calle% t#e principal part of t#e laurent series. & pt f ( z ) dz f ( z ) dz 1 ! ( f )= f ( t ) e dt = " ( p) . 1 1 bn= f ( x ) sin nx dx . )ere is an example for it. * is ' from an= , bn= ' 2 & i ( z z ') n+ 1 2 & i ( z z ')n + 1 /f all t#e bMs are & & t bn an% put @O6IO5<"/O6: # ( p ) $ ( P )= (pi +x+ ' an% 1 from '+x+pi.an% %o t#e same for e p t g ( t J) % ( J) d J dt Oero, f(O is anal.tic at O!OS', an% we call OS' a regular point. /f & ' t = ' J= ' b n ' but all t#e bMs after bSn are Oero, f(O is sai% to #ave a 1 t t#em toget#er in ,1:an= 'cos nx dx + 1cos nx dx pole of or%er n at O!OS'. /f n!1, we sa. t#at f(O #as a simple pole. & ' & w#ic# is e2ual to ! g ( t J ) % ( J) d J . /f t#ere are an infinite number bMs %ifferent from Oero, f(O #as an ' COMPLEX FORM OF FOURIER SERIES: essential singularit. at O!OS' . "#e coefficient bS1 of 1U(OSOS' Vs n= :/8A@ :75"A *<6@"/O6: ix ix 2i x in x calle% t#e resi%ue of f(O at O!OS'. . or in ot#er f ( x )=c' + c1 e +c 1 e +c 2 e ... = cn e (n ) n (n ) K( x) L ( x a) dx =(1) K ( a ) . 9O>7 *O8><5A9 ")7 879/:<7 ")7O87>: 5et OS' be an isolate% singular point n = & 1 of f(O . 4e are going to fin% t#e value of f ( z ) dz aroun% a 1 x> a wa. c n= f ( x ) e inx dx %o t#e same t#is as %one in /6IO5I/6F L . u ( x a )= . 2& & simple close% curve @ surroun%ing OS' but inclosing no ot#er ' x< a 2 singularities. 5et f(O be expan%e% in laurent series about O!OS' example 1: On OTHER INTERVALS t#e coefficients , t#e. can a x L ( x )=' ,b x L ' ( x )=L ( x ) , c x L ' ' ( x )=2 L ( x ) . & 2& t#at converges near O!OS'. 1 1 @O6IO5<"/O6 ")7O87>: b n= f ( x )sin nx dx be written a n= f ( x )cos nx dx t t f ( z) dz = 2 & iresidue of f ( z ) at t%e singular point insideC) && & ' 2& g ( t J ) % ( J ) d J = g ( J ) % ( t J ) d J 9um of resi%ues if we #ave more t#an one. /n general we write it 1 inx ' ' * ( z ' )=lim z z ( z z ') f ( z ) w#en OS' is a simple pole. an% t#e c n= f ( x ) e dx . or u can write t#em in form "8A69*O8>9 O* :78/IA"/I79 O* .: 2& ' ><5"/?57 ?O579: >ultipl. f(O b. (O(OS' $m, w#ere m is an l ! ( y ' )= p& y ' 1 n& x integer greater t#at or e2ual to t#e or%er n of t#e pole, %ifferentiate 2 an= f ( x ) cos dx an% same for bn an% c n ./f ! ( y ' ' )= p & p y' y' ' l l l t#e result m(1 times, %ivi%e b. (m(1 ! An% evaluate t#e resulting 3 2 ! ( y ' ' ' )= p & p y ' p y ' ' y '' ' expression at O!OS'. a function is even li0e x$2 or cos x, w#ose grap# for negativ x is KAP 2. CALCULUS OF VARIATIONS: 1ust a reflection in .(axis of its grap# for positive x d " " f ( x )= f ( x ) an% t#e opposite for o%% func. =' 7<578 7;<A"/O6 : f ( x )= f ( x ) . dx y ' y l KAP12. 3ENERALI4E PO5ER SERIES OR THE ' if f ( x ) isodd FRO6ENIUS: f ( x ) dx= l an% again if f(x is METHO OF 2 2 l 2 f ( x ) dx if f ( x ) iseven 7xample: x y ' ' +B xy ' +( x +2 ) y =' .

{ {

'

{
l

'

o%%,

b n=

2 x dx f ( x ) sin n & l ' l a n ='

y= a ' x s +a1 x s +1+ a2 x s +2 ...= an xn+ s

if f(x is even t#en.

y ' =s a ' x

s 1

+( s+ 1 ) a a x +( s +2 ) a 2x

s +1

...= ( n + s ) a n x
n='

n + s 1

KAP 17. PARTIAL IFFERENTIAL E$UATIONS: 5A?5A@7M9 7;<A"/O6 2 u =' t#e function u ma.be t#e gravitational potential in a region containing no mass. ?O/99O6M9 7;<A"/O6 2 u= f ( x , y , z ) "#e function u ma. represent t#e same p#.sical 2uantities liste% in laplacMs e2uatoin. Nut in a region containing mass, electric c#arge, or i 3x f ( x )= g (3 ) e d 3 ! FOURIER TRANSFORM. sources. ")7 :/**<9/O6 O8 )7A" *5O4 7;<A"/O6: 1 i3 x g ( 3 )= f ( x ) e dx . FOURIER SINE 1 u 2 2 & u= 2 . )ere u ma.be t#e non(stea%.(state 3 t f s ( x ) an% g s ( 3) a pair of TRANSFORMS. 4e %efine temperature, temperature var.ing wit# time, in a region wit# no fourier sine transforms representing o%% functions. ' #eat sourcesH or it ma. be concentration of %iffusing substance. 2 f s ( x )= g s (3 ) sin 3 xd 3 , Alp#a is %iffusivit. constant. & ' 1 u 2 ' 4AI7 7;<A"/O6: u= 2 . 2 v t g s ( 3)= f s( s ) sin 3 xdx . FOURIER COSINE & ' KAP 18: FUNCTIONS OF COMPLEX VARIA6LE: f x ( x ) an% g s ( 3) a pair of A function f(O is anal.tic in a region of t#e complex plane if it #as TRANSFORM. 4e %efine a %erivative at ever. point of region. "#e statement Pf(O is anal.tic fourier cosine transform representing even function. ' at a point O!aQ means t#at f(O #as a %erivative at ever. point 2 f s ( x )= g s (3 ) sin 3 xd 3 insi%e some small circle about O!a. & ' f ( z )=u ( x , y )+iv ( x , y ) is anal.tic in a ")7O87> 1: /f ' 2 u u v g c ( 3 )= f c ( s ) cos 3 x dx . = = region, t#en in t#at region . "#e e2uations are &' x x y KAP ". OR INAR# IFFERENTIAL E$UATIONS. calle% t#e @A<@)R(8/7>A66 on con%itions. 5/67A8 */89"(O8:78 7;<A"/O69 77=A>?57: ")7O87> 2: /f u(x,. an% v(x, . an% t#eir partial %erivatives dy wit# respect to x an% . are continuous an% satisf. t#e @auc#.( = P dx ln y= P dx y ' +Py =Q $%&! y 8ienmann con%itions in a region, t#en f(O is anal.tic at all points Pdx P dx + C insi%e t#e region. . 97@O6:(O8:78 5/67A8 y =e =Ae 9O>7 :7*/6/"/O69: A regular'point'of f(O is a point at w#ic# y ' ' +A. ' +B y =' , 7;<A"/O69 7=A>?57: 4e #ave 2 f(O is anal.tic. A singular'point'or singularity'of f(O is a point at dy 2 d dy d y Dy= = y' , D y= = = y ' ' D =d / dx . w#ic# f(O is not anal.tic. /t is calle% an isolate% singular point if dx dx dx d x 2 f(O is anal.tic ever.w#ere else insi%e some small circle about t#e 2 2 D1 y +A: y+ B.= ' ( D +A D+B ) y =' t#is is x$2 singular point. expression ")7O87> 3. /f f(O is anal.tic in a region 8, t#en it #as ( D +1)( D + B) y= ' ( D+ B ) y =' C ( D+1 ) y =' %erivatives of all or%ers at points insi%e t#e region an% can be B x x B x x . /n general it is: y =c1 e , y= c2 e , y =c1 e +c 2 e expan%e% in a ta.lor series about an. point OS' insi%e t#e region. ax bx y =c1 e +c 2 e solution of ( Da )( D b ) y =' /f we #ave "#e power series converges insi%e t#e circle about OS' t#at exten%s relation between t#e average of t#e s2uare of f(x an% t#e coefficients in t#e fourier series for f(x t#at t#e s2uare is finite. 1 f ( x )= a '+ a n cos nx+ bnsin nx . EFINITION OF 2 1 1

2 n& x a n= f ( x ) cos dx l ' l b n ='

y ' ' =s ( s 1) a ' xs 2+( s +1) s a 1 x s 1+( s +2)( s +1 ) a2 xs ! (n + s )(n+ s 1) a n x

. PARSEVAL'S THEOREM: A

( )

p & nx y CAUCH#'S FORMULA: i & ( y )=& n( y )=sin# . "#e final result can be writter as dz dx pie d f ( ( ) d( f ( z ) dz a 1 1 4e 0now t#e value of = + 2 2 2 2i f ' ( z )= f ' ( a )= 2 2 1+ z 1+ x 1+ p e n& x n& y 2 &i 2 & i C p ' (( z ) ( z a ) un ( x , y)= B n sin sin# . 4e will fin% now a solution a a t#e contour integral is & no matter #ow large p becomes since f ( ( ) d( f ( z ) dz ( n) n+ (n ) n+ f ( z)= f ( a )= x o x< a n+ 1 n+ 1 t#ere are no ot#er singular points besi%es O!i in t#e upper #alf( 2 & i ( ( z) 2 & i ( z a) w#ic# also satisfies t#e con%itions u ( x ,b )= . ' x =a p - t#en t#e secon% integral on t#e rig#t of t#e plane.5et 12 f ( z ) = LAURENT SERIES EXAMPLE: e2uation above ten%s to Oero since t#e numerator contains p an% 8emember t#at s2uare on t#e x(.(axis. 4e call t#is new con%ition z ( 2 z )( 1 +z ) a t#e %enominator p$2. "#us t#e first term on t#e rig#t ten%s to n& x n&b 2 "#is function #as singular points at O!', O!2, an% O!(1. 4e write it & (t#e value of t#e contour integral as p - an% we g(x . b n sin# a = a g ( x ) sin a dx g(x !x from ' to a. ' in a form. dx a 1 =& #ave . .$,/'01.$2D'CA3'B1'4/1D'.2 ' 2 n&b 2 n& x 1 + x bn= sin# x sin a dx integrerer og fZr a a' P( x ) a dx 15A!4A.1'A3&',3.1#*A!'2"'.$1'"*20)' 2 &n x & nx Q( x ) a sin a x cos 1 a a )'/f ?(x an% ;(x are pol.nomials wit# t#e %egree of ; at least two bn= sin# n & b 2 2 2 a a &n & n greater t#an t#e %egree of ?, an% if ;(O #as no real Oeros. /f t#e ' 1 2 n+ 1 integran% ?(x U;(x is an even function, t#en we can also fin% t#e n & b 2 a (1) b = sin# . "#en t#e solution is integral from ' to infinit.. COS:X; IS AN EVEN FUNCTION n a a &n AN SIN:X; IS AN O FUNCTION. 1 n+ 1 n & b 2 a (1) n& x n& y u ( x , y )= sin# sin# sin . a &n a a n= 1 SOME IMPORTANT IFFERENTIAL RELATIONS: 2 2 HERE IS ANOTHER EXERCISE FROM O6LI3 2: D =i ( an% t#e 4#en we #ave ( D +( ) y =' 4e get 2 i(t i(t u( x , y ) u ( x , y ) solution can be written in two forms: or y = Ae + B e y = 47 #ave a partial %ifferential: wit# 2 y =c1 sin ( t +c 2 cos ( t . y x t#e con%ition u ( x ,o )=L ( x ) . 4e will tr. to solve t#is A SECON OR ER IFFERENTIAL EXAMPLE: e2uation b. use of fourier transform, >)-< x a0 1a,)a?le a*+ @ a0 pa,ame-e,. 4e use now t#e fourier transform of t#e %erivative. 2 " [ f ' (t )] =i ( " [ f (t )] , " [ f ' ' ( t )]=( " [ f ( t )] . 4e ta0e t#e fourier transform of t#e e2uation. 4 ( x ,t ) 2 2 y ( 4 ( ( ,t )= y ( 4 ( ( , y ) dy=4 ( ( , y ) y 1 4 ( ( , y ) dy= y( 2 4 ( ( , y ) 4 ( ( , y )= A( ( ) e y ( 4 (( , y ) 2 1 /2 ( y + y ( 4 ( ( , y)= ' 4 ( ( , y )= A ( ( ) e y 1 u ( x , y )= 4 ( ( , y ) e i( x dx .we 0now t#at 2 & 1 x =' i( x L ( x )= 4 (( , ')= L(x )e b. ' xa 2 & L ( x ) f ( x ) dx = f ( ' ) %efinition #ence for <(w, ' we get 1 1 i( x 4 (( , ')= u ( x , ') e dx = = A( ( ) t#us we #ave 2 & 2& 1 1/ 2( y 4 (( , y )= e an% t#e final result is 2& 1 1 i ( x 1/ 2 y ( u ( x , y )= e e d( vi integrerer og bru0er 2 & 2 & formelen fra rottman. 8esten er en0elt.

( ( (

)[

) ) [

( )

( )

FOURIER TRANSFORM TIL 9 VISE: cos ( x & a( x ) = e . a 2+( 2 2 a

'

LAPLACE TRANSFORM OF ERIVATIVES: ! [ f ' ( t ) ] =s ! [ f ( t )] f ( ' ) *or /t is for t#e first %erivative 2 t#e secon% %erivative ! [ f ' ' ( t )] =s ! [ f ( t )] s f ( ') f ' ( ') SOME OTHER METHO S FOR FIN IN3 RESI UES:
2& 2&

,=

'

dt 1 = 1+b cos2 t 2

'

dt it . dz =i e d t =i z d t 1 +b cos2 t

or

1 1 it 1 dz we set z =e , cos t = ( z +z ) an% get iz 2 1 dt ,= #ere is not so muc# place so .. 2i b B b b z +( +1 ) z 2+ B 2 B 2 zd z 2 z W(=(b2 2 b+1 ) /b . *ire 2 2 2 2 ib ( z z W )( z z ( ) en0le poler, bare polene / z W er innenfor @ og begge 1 b & = ,= resi%.ene er . :ette gir 2 2( z 2 X b +1 1+b W z ( ) , =& blir integran%en 1, og dt =

RESI UE THEOREM SIMPLER 5A#: consi%er

dx 1+ x 2

. 4e
2

u= u 5e <a1e a +)((e,e*-)al e=/a-)'*: . on t#e 2 2 x y 4#ere @ is close% intervall ' x a , ' y b me% grensebetingelser boun%ar. of t#e semi u (', t )=u (', y )= u( a , y )=' . 4e will use separation of circle s#own #ere. *or variables to fin% t#e general e2uation. an. pY1, t#e semicircle 6 ' '( x) & ' '( y) incloses t#e singular point + =' u ( x , y )= 6 ( x ) & ( y )=' Or O!i an% no ot#ersH t#e 6 ( x) &(y) 6 '' (x) & ' ' ( y) resi%ues of t#e integran% at O!i is : = = (constant ) #ence we #ave 1 1 6 ( x) & ( y) * ( i )= ( z i ) = . "#en t#e value of z i n& x ( z i )( z +i ) 2 i 6 ' ' + 6 =' =(' ! =(a !' 6 ( x )= 6 n( x )=sin t#e contour integral is 2 & i ( 1 / 2i )=& . /ntegral in two parts.(1 a 2 2 an integral along t#e x(axis from (p to pH for t#is part O ! xH (2 an n & = n= 2 & ' ' & = ' , R(' ! ' . i , integral along t#e semicircle, w#ere . "#en we #ave z= p e a

2 1+ z

dz

lim

IT >a0 -<e )*1e,0e lapla.e -,a*0(',m '(:

2 p 3 ( p +2 )

USIN3 LAPLACE TRANSFORM TO O6TAIN SOLUTIONS FOR COUPLE E$UATIONS: 2 x ( t )+2 n x ( t )+n x ( t )= ' we will use laplace transfrom 2 y ( t )+2n y x(t ) ( t )+ n y ( t )=[ to obtain t#e solution of x(t an% .(t . /nitial con%itions x ( t )=\ 4e call 5(. ! R an% 5(x ! =. x(' !.(' !'. 2 x ' ' +2 nx ' +n x= ' "ransforms of %erivatives og .. 2 y ' ' + 2 ny ' +n y =[ x ' 2 ! ( y' )= p& y' , ! ( y ' ' )= p p y ' y '' we insert t#is into e2 6 ( p +2 n p + n ) \ =' 6 ( p + n) =\ 2 2 2 & ( p + 2 n p+n )[ p 6 =' & ( p +n) =[ p 6 \ p[ \ \ 1 1 &= , 6= ! 1 =\ ! B 2 2 2 ( p+ n ) ( p +n ) ( p +n ) ( p + n) 1 1 1 !1 = ! [ # ( p )$ ( p ) ]= g % b. 5S3B ( p + n ) ( p +n ) 1 1 n t n ( t J ) # ( p )= ! [ g ( t )] = g ( t )=e g ( t J )=e p +n p +n 1 1 nt n J $ ( p )= ! [ $ ( p )] = e %( J)=e p +n ( p +n )
2 2 2

/n @ase]3: t#e solution is given b. a a y =c1 x cos( D ln ( x ))+c 2 x sin (D ln ( x )) )ere is an example for Freen function.

] [

'

\e

(t J )

n J

d J=

'

\e

nt

d J x ( t )= \ t e

x ( ' )= \

&=
1

[ p\ p 1 [ \ ! 1 B 2 2 p +n ( p +n ) ( p+ n )

! [ # ( p )$ ( p )] =g % # ( p)=
$ ( p )=

]
2

( p +n )

( 1 n t ) e

nt

=g (t )

1 nt nt n t t % ( t )=t e y ( t )=[ \ e 2 2 T ( p +n ) THIS IS VER# IMPORTANT FOR IFFERENTIAL E$. y ' ' + y ' 2 y =B sin 2 x . instea% of tac0ling t#is problem %irectl., we are first going to solve t#e e2uation . 9ince exp(2ix ! 2i x cos 2x W / sin 2x is complex, t#e solution & ' ' +& ' 2 & =B e R ma. be complex also. "#en & =& *+i & , is e2uivalent to 2i x & * ' ' +& * ' ' 2 & *= B e =B cos 2x two e2uations 9ince t#e 2i x & , ' ' +. , ' 2& , = B e =B sin 2x secon% e2uation above is t#e same as t#e 2uestion, we see t#at t#e solution of t#e 2uestion is t#e imaginar. part of R. "#us to fin% y p for t#e 2uestion we fin% & p for t#e e2uation above an% ta0e its imaginar. part. 4e observe t#at 2i is not e2ual to eit#er of t#e roots of t#e auxiliar. e2uation in R e2uation. *ollowing t#e met#o% of t#e last paragrap#, we assume a solution 2 ix of t#e form & p= C e an% subsitute it into R e2uation to get:

(B+2i 2 ) C e & p=
1

2i x

=B e

2i x

C=

B (2i T) 1 B = = ( i +3 ) 2 i T B' A

( i +3 ) e ta0ing t#e imaginar. part of RSp we fin% A 1 3 y p= cos 2x sin 2x . A A

2ix

EN INHOMO3EN LI3NIN3EN. y ' ' + P ( x ) y ' +Q ( x ) y =* ( x ) Ii mer0er oss at #vis y %( x ) er en l^sning av %en #omogene ligningen 8(x ! ', og .Sp (x en eller annen l^sning av %en in#omogene (parti0ul_r y ( x )= y% ( x )+ y p( x ) en l^sning av %en l^sning , sZ er ogsZ: in#omogene ligningen. 9p^rsmZlet er %a om %enne l^sningen y ( x') og y ' ( x ') , passer me% gitte grensever%ier for sli0 at vi fZr %en ^ns0e%e ent.%ige l^sningen. >e% y %( x )=c1 y 1( x )+ c2 y 2( x ) fZr vi betingelsene

c1 y 1( x ')+ c2 y 2( x')= y ( x ') y p( x' ) c 1 y 1( x' )+c2 y ' 2( ' )= y' ( x ') y p( x ')

0rav til l^sning er

THE CAUCH#AEULER E$UATION. 2 d y +b y =' x2 d y +a x 4e assume a trial solution given b. 2 dx dx dy m 1 m =m x , %ifferentiating, we #ave. An% y=x dx 2 d y m 2 =m ( m1 ) x . 9ubstituting into t#e original e2uation 2 dx 2 m 2 m1 m x ( m( m1) x )+ a x ( m x )+b ( x )='

som 01ent at %eterminanten til ligningssettet mZ v_re uli0 null, %vs. y ( x ) y 2 ( x ') 7= 1 ' ' . Og %ette vet vi fra f^r er o0 y ' 1 ( x') y' 2 ( x ) y 1( x ) , y 2( x ) er line_rt uav#engige. ()vis #^.re si%en / si%en ligningsettet ovenfor s0ulle v_re li0 null, #ar vi %ire0te at y p( x ) er en l^sning som tilfresstiller 0ravene, og er %erme% %en s^0te ent.%ige l^sningen :erme% er problemet formelt l^st. :en generelle l^sningen av %en in#omogene %iff ligningen #ar y ( x )= c1 y 1( x )+ c2 y 2( x )+ y p ( x ) 6este problem blir formen y p( x ) . )er er %et imi%lerti% Z finne en parti0ul_r l^sning flere meto%er som 0an bru0es. 7t vanlig tilfelle er %iff ligninger #vor venstre si%en #ar 0onstante 0oeffis%ienter. y ' ' + ay ' +by = * ( x ) )er 0an vi 0omme langt me% lit strategis0 g1etting. rx rx 1. *( x )= A e , pr8v y p( x )=B e * ( x )= A sin rx+ B cos rx , pr8v y p ( x )=C sin rx +D cos rx 2. 3. * ( x )= polynomav grad 3 , pr8v y p( x )= polynom av grad 3 B x x * ( x )=e ( A sin rx + B cos rx ) , pr8v y p ( x )=e (C sin rx + D cos rx ) rx 70semple: . 5^sning av %en #omogen lig y ' ' 2. ' + y= A e x x rx vi pr^ver og finner y %( x )=c1 e +c 2 x e y p( x )= B e A A B= y ( x ) = og %erme% . p ( r 1 )2 ( r 1 )2 y ' ' + B y ' + B.= cos# x Anot#er example is: 4e want to fin% t#e general solution to t#e %ifferential e2uation, t#at is we want to fin% t#e solution to t#e #omogeneous %ifferential e2uation. y ' ' + B y' + B.= ' *rom t#e c#aracteristic e2uation 2 2 \ + B \ +B =( \ +2 ) = ' \ =2, 2 9ince we #ave repeate% root, we #ave to intro%uce a factor of x for one solution to ensure 2 x 2 x linear in%epen%ence. 9o we obtain u 1 =e ,u 2=x e "#e wrons0ian of t#ese two functions is

e xe 2x 2 x 2 x 2 x 8earranging gives : m 2+( a 1) m +b =' we t#en can solve for =e e ( 2 x 1 )+2x e e 2 x 2x 2 e e (2 x 1 ) m. "#ere are t#ree particular cases of interest. Bx B x Bx Bx @ase ]1: "wo %istinct roots, m 1 an% m 2 Necause t#e e ( 2 x 1 )+2 xe =(2 x +1 +2x ) e =e @ase ]2: One real repeate% root, m wrons0ian is non Oero, t#e two functions are linearl. in%epen%ent, @ase ]3: @omplex roots, a D i . so t#is is in face t#e general solution for t#e #omogeneous %iff m m /n @ase]1, t#e solution is given b.: y =c1 x +c2 x e2uation. 4e see0 functions A(x an% N(x so m m A ( x ) u 1+ B ( x ) u2 is a general solution of t#e non /n @ase]2: t#e solution is given b. y =c1 x ln( x )+c 2 x #omogeneous e2uation. 4e nee% onl. calculate t#e integrals
1 2

2x

2 x

1 1 u ( x ) B ( x ) dx , B ( x )= u1 b ( x ) dx 7 2 7 1 2x 2x A ( x )= Bx x e cos# x dx= xe cos# x dx e A ( x )=

i.e.

e (( V( x 1 )+e (3 x1 ))+ C )

2x

B ( x )=
x2

1 2x 1 x 2x 2x e cos# xdx = e cos# x dx = e ( 3+ e )+C T e B x

FROM CHAPTER 2: EULER E$UATION EXERCISES. , = x 1+ y ' 2 dx

" ( x , y , y ' )= x 1 + y' dx

1 arc cos#( x )= 2 x 1
y
=c cos# 1( x / c )

" ( x 1+ y' ) 2 d u 1 = u = y' +1 = y' y' du 2 u 2 2 d ( y ' +1 ) d(y ' ) d x x + (1 ) 2 1 + y' dy ' dy dy x = 2 2 y' 2 y ' +1 2 y ' +1 " x y' = from 7uler e2uation we #ave partia y' y' 2+1 x y' " " " =' =' =' x y' y x y' 2 +1 y x y' 2 2 2 =c ( constant ) ( x y' ) =c ( y' +1 ) y' 2+1 c dy c 2 2 2 2 2 x y ' = y ' c +c y ' = 2 2 , y' = = 2 2 dx d x x c x c c c dy= 2 2 dx y = 2 dx x c x 1 2 c " =', y

1 2 ' + f ( y ) dy=const) y @A97 (% : An e2uation of t#e form 2 d y +a y= f ( x ) a2 x 2 d y +a 1 x @alle% an 7uler Or @auc#. 2 dx ' dx e2uation, can be re%uce% to a linear e2uation wit# constant coefficients b. c#anging t#e in%epen%ent variable from x to O dy dy z x = w#ere. *or t#e we #ave an% x= e dx dz 2 2 d y d y x2 d y = 2 2 dx dz d z @A97 (e ! 8e%uction of or%er. "o fin% a secon% solution of y ' ' + f ( x ) y ' + g( x )=' given one solution u(x , substitute y =u ( x ) v ( x ) into e2uation above an% solve for v(x . A IFFERENTIAL EXERCISE 6# USE OF LAPLACEA ' t <' y '= y '' = ' f ( t )= 1 t >' get

T<e ,e0- '( .'l/m* 1 le(-. F'/,)e, -,a*0(',m example: 4ttry 'f9x:'ved'%;elp'av'uttry et'du'%ar'funnet'for'"9(:)' Ii mer0er oss at *(w er li0e fun0s1on / w, som ventet, sli0 at fourierintegralet for f(x ogsZ blir en cosinus(transformas1on 1 1 i(x f ( x )= " ( ( ) e d(= " ( ( ) cos( ( x ) d( 2 & 2 & 2 1 cos ( (a) cos ( ( x ) d( 2 & ( ANOTHER FOURIER TRANSFORM EXAMPLE: FINN FOURIERREKKA FOR FUNKSCONEN (:x; 3ITT 1 ( !+ x ) ! x <' sin n x 2 f ( x )= tilZfin% . n 1 n =1 ( ! x ) '< x ! 2 a n& x n& x f ( x )= ' + an cos + b sin Og 0oeffisientene 2 n=1 ! n= 1 n ! ! 1 n& x dx og bSn finnes av an= f ( x ) cos ! ! !

y x =cos# c c

( )

ax =cos# ( ay+ b )

1 !b = constants c an% 0 are %etermine% from given c c x1 ! x 2 . points HERE IS ANOTHER $UESTION SAME LIKE A6OVE:

a=

x2

, = (1+ y y ' )2 dx " ( x , y , y ' )=(1 + y y ' )2

t#en

" " =2. ' ( 1 + y y ' ) ! =2 y ( 1 + y y ' ) y y' d " " d =' ( 2 y (1+ y y ' ) )2 y ' ( 1 + y y ' )=' d x y' y dx

t#e %erivation ca n #appen wit# pro%uct t#ing. d dv du ( u v )= v +u An% at last we get. dx dx dx 2 2 2 2( y y ' ' + 2 y y' + y ' ) 2 y ' ( 1 + yy ' )=' 2 y ( y' ' y )=' f^lgen%e integraler: og 1 2 c' x = y an% we #ave to fin% t#e x w#ic# is e2ual to . 2c 2c

1 n& x f ( x ) sin ! dx Ii mer0er oss at f(x / %enne ! ! oppgaven er en o%%e(fun0s1on, sli0 at vi #ar an=' for alle n. Ii%ere finner vi for b n si%en f(x er o%%e fun0s1on. ! 2 x b n= ( ! x ) sin n & dx *or Z bestemme bSn be#^ver vi ! ' ! bn=
!

A long rectangular metal plate #as its two long si%es an% t#e far en% at ' @ an% t#e n& x ! n 2 2 RESI UE EXAMPLE: %en blir til base at 1'' @. "#e wi%t# of t#e plate is ( 1 B )( 1 +B z ) x sin dx= (1 ) ( delvis integras;on ) ctil C ! n& 1'cm. *in% t#e stea%. state temperature i pi x ' e . =' %istribution insi%e t#e plate. dz :er integras1onsveien @ er en lu00et 2 2 ! 2 2 C ( 1 B )(1 +B z ) sammen finnes %a b n= og fourierre00a for f(x blir . + . =' n & Or "o solve t#is 0urve ra ` 8 til 8 pZ %en reelle a0sen og en #alvsir0el me% ra%ius 2 2 x y 8 / ^vre #alvplan. 7tter aor%ans lemma vil integralet over ! 1 n&x f ( x ) = sin e2uation, we are going to tr. a solution of ! = & setter og #ar %a summen * . 9i%en integran%en #ar #alvsir0elen gZ mot null nZr & n ! . ( x , y )= 6 ( x ) & ( y ) . t#e form. n =1 poler pZ %en reelle a0sen for O ! (1U2 og O ! b, mZ vi ogsZ legge 2 2 1 d 6 1 d & fra oppgaven. 2 2 smZ #alvsir0ler run%t %isse / ^vre #alvplan. Ii%ere er %et / ^vre = =const = 0Y' . 6 ' ' = 6 An% IFFERENTIAL EXERCISE: 6 d x 2 & d y2 #alvplan en pol for O!iU2. ?rinsipal ver%ien av integralet er %a gitt 2 2 S'l1e )-D x y ' ' xy ' + y = x cos & x & ' ' = & "#e constant 0$2 is calle% t#e separation constant. P dx som er li0 ve% y y ( 1 Bx 2) ( 1+ Bx2 ) sin x e sin x e 4e first 6= , & = y . = 6& = Alle y 2 & i *es ( z =i /2 )+& i [ *es ( z =1 / 2 )+ *es ( z =1 / 2 ) ] cos x e cos x e & y 1 2 %iscar% t#e solutions containing since we are given polene er en0le, og vi #ar %en an%re e *es ( z =i / 2 )= i e X . ' as y . 6ext %iscar% solutions containing i cos(0x since "!', w#en x!'. "#is leaves us 1ust e y sin x , *es ( z=1/ 2)= *es ( z = 1/ 2 )= resultatet blir %a. X but t#e value of 0 is still to be %etermine%. 4#en x!1', we are to & pi x & n& dx = ( 1+e 2 ) cos = 2 2 #ave "!', t#is will be tru if sin(1'0 !', t#at is if for X C ( 1 B )(1 +B z ) 1' FOURIER TRANSFROM EXAMPLE: n!1,2,... t#us for an. integral n, t#e solution n& x n & y / 1' 2x +a a x ' .=e sin 9atisfies t#e given boun%ar. con%itions f ( x )= 2x + 2a ' x a :en fourier transformerte er %a. 1' ' ellers on t#e t#ree "!' si%es. *inall. we must #ave "!1'' w#en .!'
'

sin

n& x ! n dx= [(1 ) 1 ] ! n&

og

cos

pi x

dx

{ "{ "

. = b n e
n= 1 n= 1

n & y /1'

sin

n& x 1'

for .!' we must #ave "!1'' for f(x ! 1'' wit# l!1'.

. y = '= bn sin
bn=
.= OTHER SECON OR ER E$UATIONS: y ' = p , y' ' = p' @A97 (a : :epen%ent variable . missing: @A97 (b : /n%epen%ent variable x missing: dp dp dy dp y ' = p , y' ' = = =p dx d y d x dy y ' ' + f ( y )=' multipl. b. .M. @A97 d: "o solve y ' y ' ' + f ( y ) y' =' or y ' dy ' + f ( y ) dy=' t#en integrate to

n& x =1'' 1'

1' B'' 2 n&x 1'' sin dx= n & 1' ' 1' '

odd n even n

" becomes t#en

B'' & y /1' & x e sin n 1'

1 3 & y / 1' 3 & n e sin +... 3 1'

ANOTHER FRO6ENIBS METHO EXAMPLE.