Академический Документы
Профессиональный Документы
Культура Документы
075FALL 2004
10. Functions of a Complex Variable 10.15. Indented contours. . 110. By making use of integration around suitable indented contours in the complex plane, evaluate following integrals: the sin x (a) x(x 2 +a2 ) dx (a > 0), sin x (b) x( 2 x2 ) dx. Solution. (a) For R > 1 and 0 < < 1, dene the contour C = C 1 + C2 + C3 + C4 , where C1 is the real interval [R, ], C2 is the upper half of the circle |z | = with clockwise orientation, C3 is the real interval [, R], and C4 is the upper half of the circle |z | = R with counterclockwise orientation. Then eiz eix sin x dx = Im lim dz Im P dx. 2 2 2 2 2 2) R z ( z + a ) x ( x + a C + C x(x + a ) 1 3 0 Moreover, by Theorem 2 of classnotes, eiz lim dz = 0, 2 R C4 z (z + a2 ) and, by Theorem 4 of classnotes, eiz eiz i dz = i Res [ , 0] = 2 . lim 2 2 0 C2 z (z 2 + a2 ) z (z + a ) a So,
R 0
lim
C1 +C3
eiz dz = z (z 2 + a2 )
R 0
lim
eiz dz lim R z (z 2 + a2 ) 0
C2 +C4
eiz dz z (z 2 + a2 )
= 2i Res[ = i Hence,
eiz i , ia] ( 2 ) 2 2 z (z + a ) a
1 ea . a2 1 ea eiz dz = . z (z 2 + a2 ) a2
C1 +C3
(b) For large R > 0 and small > 0, dene the contour C = C 1 + C2 + C3 + C4 + C5 + C6 + C7 + C8 , where C1 , C3 , C5 and C7 are the real intervals [R, ], [ + , ], [, ] and [ + , R], C2 is the upper half of the circle |z + | = with clockwise orientation, C 4 is the upper half of the circle |z | = with clockwise orientation, C 6 is the upper half of the circle |z | = with clockwise orientation, and C 8 is the upper half of the circle |z | = R with counterclockwise orientation. Then eiz eix sin x dx = Im lim dz Im P dx. 2 2 2 2 2 2) R z ( z ) x ( x + C + C + C C x( x ) 1 3 5 7 0 By Theorem 2 of classnotes,
R C8
lim
eiz dz = 0, z ( 2 z 2 )
and, by Theorem 4 of classnotes, eiz 1 i lim dz = i Res[ , ] = , 2 2 2 2 0 C2 z ( z ) z ( z ) 2 eiz eiz i lim dz = i Res [ , 0] = , 2 2 2 2 0 C z ( z ) z ( z ) 4 eiz eiz i lim dz = i Res [ , ] = . 2 2 2 2 0 C6 z ( z ) z ( z ) 2 Since
eiz z ( 2 z 2 )
has no singularities in the upper half plane, we get eiz dz = 0, 2 2 C z ( z ) eiz dz = lim R z ( 2 z 2 ) 0 eiz 2i dz = . 2 2 z ( z )
So
sin x dx = Im lim R x( 2 x2 ) 0
2 eiz dz = . 2 2 z ( z )
i eitx dx = 0 x i
cos tx dx = 0, x
and
sin tx dx = 0 x
Solution. Case 1, t > 0. For large R and small , dene the contour C = C 1 + C2 + C3 + C4 , where C1 and C3 are the real intervals [R, ] and [, R], C 2 is the upper half of the circle |z | = with clockwise orientation, and C 4 is the upper half of the circle |z | = R with counterclockwise orientation. Then itx eitz e dx = lim dz. P R C1 +C3 z x 0 By Theorem 2 of classnotes,
R C4
lim
eitz dz = 0, z
C1 +C3
C2 +C4
Case 2, t < 0. For large R and small , dene the contour C = C 1 + C2 + C3 + C4 , where C1 and C3 are the real intervals [R, ] and [, R], C 2 is the lower half of the circle |z | = with counterclockwise orientation, and C 4 is the lower half of the circle |z | = R with clockwise orientation. Then itx eitz e P dx = lim dz. R C1 +C3 z x 0 By Theorem 2 of classnotes,
R C4
lim
eitz dz = 0, z
Since
has no singularities in the lower half plane, we get itz e dz = 0. C z eitx dx = lim R x 0
1 x
This gives P
C1 +C3
eitz dz = lim R z 0
C2 +C4
eitz dz = i. z
Case 3, t = 0. Since
eitx
Compare the real and imaginary parts on both sides. We get cos tx P dx = 0, x and P Note that the integral above becomes
i dx = 0 x i
sin tx x
dx actually converges in all three cases. So the last equation sin tx dx = 0 x (t > 0), (t = 0), (t < 0).
sin tx dx = 0 x
dx
[Notice that f (z ) = (eiaz eibz )/z 2 has a simple pole at the origin.] By taking a = 0 and b = 2, also deduce the formula sin2 x = . dx x2
Solution. If we naively set cos ax = Re(e iax ) and cos bx = Re(eibx ) and take the Re outside the integral sign, then the resulting integral doesnt make any sense as is: cos ax cos bx eiax eibx I= dx = Re dx = . x2 x2 More precisely, I itself is nite, since the integrand in the left-hand side is well-behaved for all x. For example, by expanding the cosines in the integrand near z = 0, we nd cos az cos bz z2 = [1
(az )2 2!
+ ... +
+ . . .] [1 z2
(bz )2 2!
+ ... +
+ . . .]
On the other hand, the integrand on the right-hand side, f (z ) at z = 0. Indeed zf (z ) reads as zf (z ) = eiaz eibz z [1 + iaz +
) + . . .] [1 + ibz + (ibz 2! + . . . + = z n bn a2 b 2 a = i(a b) z + . . . + in z n1 + . . . , 2! n! which is analytic at z = 0 (and furthermore lim z 0 [zf (z )] = 0 since a = b). Therefore, we read I = ReIp where Ip is the principal-value integral eiax eibx Ip P dx . x2
(iaz )2 2!
+ ... +
(iaz )n n!
(ibz )n n!
+ . . .]
We calculate Ip by closing the path by a small semicircle C + of radius around z = 0 in the upper half plane, and a large semicircle C R+ of radius R also in the upper half plane. The resulting closed contour does not contain any singularities of the integrand and has to be zero by the Cauchy integral theorem. In addition, by Theorems 4& 2 of classnotes, lim dz f (z ) = i Rez [f (z ), 0] = i i(a b) = (a b),
0+ C+ R CR+
lim
dz f (z ) = 0.
The last equation was obtained by noticing that a > 0 and b > 0 while 1/z 2 goes to 0 uniformly in |z | = R as R . It follows that Ip + lim dz f (z ) + lim dz f (z ) = 0
0+ C+ R CR+
Ip = (b a) I = ReIp = (b a). In the special case a = 0 and b = 2, the integrand of the original integral becomes 1 cos 2x 2 sin2 x cos ax cos bx = = . x2 x2 x2
Hence, the result of integration reads as 2 sin2 x dx = 2, x2 which agrees with the desired formula.