SOLUTION SET IV FOR 18.

075FALL 2004
10. Functions of a Complex Variable
10.12. Residues. .
In the following, I use the notation
Res
z=z
0
f(z) Res(z
0
) Res[f(z), z
0
],
where Res is the residue of f(z) at (the isolated singularity) z
0
.
82. Evaluate the integral
_
C
dz
z
2
1
when C is the curve sketched in Figure 10.21.
Solution.
1
z
2
1
has two simple poles. One is at z = 1, the other is at z = 1. Its easy to
check that Res[
1
z
2
1
, 1] =
1
2
, and Res[
1
z
2
1
, 1] =
1
2
. The pole at z = 1 is encircled in
the counterclockwise (positive) sense, while the pole at z = 1 is encircled in the clockwise
sense. Hence,
_
C
dz
z
2
1
= 2iRes[
1
z
2
1
, 1] 2iRes[
1
z
2
1
, 1] = i (i) = 2i.
88 Determine the residue of each of the following functions at each singularity:
(a) e
1
z
, (b) e
1
z
2
, (c) cos

z
, (d) (1 + z
2
)e
1
z
.
Solution. (a) We have
e
1
z
=

n=0
z
n
n!
= 1 + z
1
+

n=2
z
n
n!
.
So e
1
z
has an essential singularity at z = 0, and
Res[e
1
z
, 0] = 1.
(b) We have
e
1
z
2
=

n=0
z
2n
n!
= 1 + z
2
+

n=2
z
2n
n!
.
Date: October 9, 2002.
1
2 SOLUTION SET IV FOR 18.075FALL 2004
So e
1
z
2
has an essential singularity at z = 0, and
Res[e
1
z
2
, 0] = 0.
(c) We have
cos

z
=

n=0
(1)
n

2n
(z )
2n
(2n)!
.
(Note that the coecient of
1
z
is 0.) So cos

z
has an essential singularity at z = , and
Res[cos

z
, ] = 0.
(d) We have
(1 + z
2
)e
1
z
= (1 + z
2
)

n=0
z
n
n!
= (1 + z
2
)(1 + z
1
+
z
2
2
+
z
3
6
+

n=4
z
n
n!
)
=
3
2
+
7
6
z
1
+ [higher powers of z
1
].
So (1 + z
2
)e
1
z
has an essential singularity at z = 0, and
Res[(1 + z
2
)e
1
z
, 0] =
7
6
.
10.13. 10.13 Evaluation of Real Denite Integrals. .
90. Use residue calculus to evaluate the following integrals:
(a)
_
2
0
d
A+B sin
=
2

A
2
B
2
(A > |B|),
(b)
_
2
0
d
a
2
+sin
2

=
_
2
0
d
a
2
+cos
2

=
2
a

a
2
+1
(a > 0),
(c)
_

2
0
sin
4
d =
_

2
0
cos
4
d =
3
16
,
(d)
_
2
0
sin
2

5+4 cos
d =

4
.
Solution. (a) First make the substitution: z = e
i
, dz = ie
i
d.
Now the complex z describes the unit circle C
1
in the positive sense as varies from 0
to 2. So, as was discussed in class, the integral becomes
_
2
0
d
A + B sin
=
_
C
1
dz
iz
A + B
z
2
1
2iz
=
_
C
1
2dz
Bz
2
+ 2iAz B
SOLUTION SET IV FOR 18.075FALL 2004 3
The poles of the integrand are simple and occur when Bz
2
+ 2iAz B = 0, which in
turn gives
z

=
iA

B
2
A
2
B
Furthermore,
z
+
z

=
A
2
B
2
+
A
2
B
2
B
2
= 1.
Therefore, z
+
is a (simple) pole inside the unit circle. Now using the known formula
Res
z=z
0
g(z)
h(z)
=
g(z
0
)
h

(z
0
)
, where z
0
: simple zero of h(z), we get:
Res(z
+
) =
2
2Bz
+
+ 2iA
=
1
iA + i

A
2
B
2
+ iA
=
1
i

A
2
B
2
Thus, by the residue theorem,
_
2
0
d
A + B sin
=
2i
i

A
2
B
2
=
2

A
2
B
2
(b) We manipulate the integrand as follows:
1
a
2
+ sin
2

=
1
a
2
+
1cos 2
2
=
2
2a
2
+ 1 cos 2
.
So, with the new variable = 2,
_
2
0
d
a
2
+ sin
2

=
_
2
0
2d
2a
2
+ 1 cos 2
=
_
4
0
d
2a
2
+ 1 cos
= 2
_
2
0
d
2a
2
+ 1 cos
= 2
_
2
0
d
2a
2
+ 1 sin
,
where we shifted the integration variable by /2 in the integral of the third line and used
the periodicity of the integrand.
By using the result of part (a) above with A = 2a
2
+ 1 and B = 1, we get
_
2
0
d
2a
2
+ 1 sin
=
2
_
(2a
2
+ 1)
2
1
=

a

a
2
+ 1
.
4 SOLUTION SET IV FOR 18.075FALL 2004
Thus,
_
2
0
d
a
2
+ sin
2

=
2
a

a
2
+ 1
,
and, hence,
_
2
0
d
a
2
+ cos
2

=
_
2
0
d
a
2
+ sin
2
(

2
)
=
_ 3
2

2
d
a
2
+ sin
2

=
_
2
0
d
a
2
+ sin
2

=

a

a
2
+ 1
.
The integral of the third line ensues from the periodicity of the integrand.
(c) By subtracting the given integrals, we get
_
2
0
sin
4
d
_
2
0
cos
4
d =
_
2
0
(sin
4
cos
4
) d
=
_
2
0
(sin
2
cos
2
)(sin
2
+ cos
2
) d
=
_
2
0
(sin
2
cos
2
) d
=
_
2
0
cos 2 d
=
sin2
2
|

2
0
= 0.
So,
_
2
0
sin
4
d =
_
2
0
cos
4
d.
Then,
_
2
0
sin
4
d =
_
2
0
sin
4
d +
_

2
sin
4
d +
_ 3
2

sin
4
d +
_
2
3
2
sin
4
d
=
_
2
0
[sin
4
+ sin
4
( +

2
) + sin
4
( + ) + sin
4
( +
3
2
)] d
= 2
_
2
0
(sin
4
+ cos
4
) d
= 4
_
2
0
sin
4
d.
SOLUTION SET IV FOR 18.075FALL 2004 5
This shows
_
2
0
sin
4
d =
_
2
0
cos
4
d =
1
4
_
2
0
sin
4
d.
But
_
2
0
sin
4
d =
_
C
1
(
z
2
1
2iz
)
4
dz
iz
=
1
16i
_
C
1
(z
2
1)
4
z
5
dz
=
2i
16i
Res[
(z
2
1)
4
z
5
, 0]
=
2i
16i
6
=
3
4
,
where C
1
is the unit circle with center at origin. The residue was found easily by noticing
that
(z
2
1)
4
z
5
=
z
8
4z
6
+ 6z
4
4z
2
+ 1
z
5
,
by which the coecient of z
1
is 6. So,
_
2
0
sin
4
d =
_
2
0
cos
4
d =
3
16
.
(d) Again, by the usual replacement z = e
i
,
_
2
0
sin
2

5 + 4 cos
d =
_
C
1
(z
2
1)
2
(2iz)
2
5 + 4
z
2
+1
2z
dz
iz
=
_
C
1
(z
2
1)
2
2iz
2
(10z + 4z
2
+ 4)
dz
The simple poles of this integrand occur when 4z
2
+ 10z + 4 = 0, i.e., when z =
1
2
or
z = 2, while a double pole occurs at z = 0. Since z = 2 is not within the unit circle, we
disregard it.
6 SOLUTION SET IV FOR 18.075FALL 2004
Res(
1
2
) = lim
z
1
2
(z +
1
2
) (z
2
1)
2
2iz
2
(2z + 1)(z + 2)
=
(
3
4
)
2
2i(
3
2
)
=
3i
16
Res(0) =
1
(2 1)!
_
d
dz
(z
2
f(z))
_
z=0
=
_
d
dz
_
(z
2
1)
2
2i(4z
2
+ 10z + 4)
__
z=0
=
_
4i(2z
2
+ 5z + 2)(4z
3
4z) (z
4
2z
2
+ 1)(16iz 20i)
(4i)
2
(2z
2
+ 5z + 2)
2
_
z=0
=
5i
16
Alternatively, you may expand the integrand in z (considering |z| small) and nd the
coecient of z
1
. (Try it for practice!)
Thus,
_
2
0
sin
2

5 + 4 cos
d = 2i
_
3i
16

5i
16
_
=

4
91. Use residue calculus to evaluate the following integrals:
(a)
_

dx
(x+b)
2
+a
2
=

a
(a > 0),
(b)
_

0
dx
(x
2
+a
2
)(x
2
+b
2
)
=

2ab(a+b)
(a > 0, b > 0),
(c)
_

0
dx
x
4
+4a
4
=

8a
3
(a > 0),
(d)
_

0
dx
(x
2
+a
2
)
2
=

4a
3
(a > 0).
Solution.
(a) The degree of the denominator is 2 greater than the degree of the numerator and the
function is nite for all real values of x. Thus, we can employ the strategy given in class
by closing the original path with a large semicircle in the upper half plane (or lower half
plane). By shifting the integration variable by b, we get
_

dx
(x + b)
2
+ a
2
=
_

dx
x
2
+ a
2
=
_
C
1
F(z)dz = 2i

k
Res(z
k
)
where the points z
k
are the poles of F(z) =
1
z
2
+a
2
in the upper half-plane.
SOLUTION SET IV FOR 18.075FALL 2004 7
The (simple) poles occur when z
2
+ a
2
= 0, that is when z = i
_
(a
2
) = ia (since
a > 0). So there is one (simple) pole in the upper half-plane, namely, at z
1
= ia.
Res(z
1
) =
1
2z
1
=
1
2ia
.
Thus,
_

dx
(x + b)
2
+ a
2
= 2i
1
2ia
=

a
.
(b)
1
(z
2
+a
2
)(z
2
+b
2
)
has two singularities on the upper half plane. One of these is at z = ai,
the other is at z = bi; both of them are simple poles. Note that the denominator of
1
(x
2
+a
2
)(x
2
+b
2
)
is of degree 4. Accordingly,
_

dx
(x
2
+ a
2
)(x
2
+ b
2
)
= 2i (Res[
1
(z
2
+ a
2
)(z
2
+ b
2
)
, ai] + Res[
1
(z
2
+ a
2
)(z
2
+ b
2
)
, bi])
= 2i (
1
2ai(b
2
a
2
)
+
1
2bi(a
2
b
2
)
)
=

ab(a + b)
.
Since
1
(x
2
+a
2
)(x
2
+b
2
)
is even, we get
_

0
dx
(x
2
+ a
2
)(x
2
+ b
2
)
=
1
2
_

dx
(x
2
+ a
2
)(x
2
+ b
2
)
=

2ab(a + b)
.
(c)
1
z
4
+4a
4
has two singularities on the upper half plane. One of these is at z =

2e
i
4
a,
the other is at z =

2e
3i
4
a. Both of them are simple poles. Note that the degree of the
denominator of
1
x
4
+4a
4
is 4. Accordingly,
_

dx
x
4
+ 4a
4
= 2i (Res[
1
z
4
+ 4a
4
,

2e
i
4
a] + Res[
1
z
4
+ 4a
4
,

2e
3i
4
a])
= 2i (
e

i
4
8

2a
3
i

3i
4
8

2a
3
i
)
=

4

2a
3
[(

2
2

2
2
i) (

2
2

2
2
i)]
=

4a
3
.
Since
1
x
4
+4a
4
is even, we have
_

0
dx
x
4
+ 4a
4
=
1
2
_

dx
x
4
+ 4a
4
=

8a
3
.
(d) The degree of the denominator is greater than twice the degree of the numerator and
the function is nite for all real values of x. We can once again employ the strategy given
in class. Also, note that the integrand is even so that
_

0
dx
(x
2
+ a
2
)
2
=
1
2
_

dx
(x
2
+ a
2
)
2
8 SOLUTION SET IV FOR 18.075FALL 2004
The poles occur when (z
2
+ a
2
)
2
= 0, that is when z = ia. Thus, there is one pole in
the upper half-plane, i.e., at z = ia, and it is a double pole.
Res(ia) =
_
d
dz
(z ia)
2
(z
2
+ a
2
)
2
_
z=ia
=
_
d
dz
1
(z + ia)
2
_
z=ia
=
_
2
(z + ia)
3
_
z=ia
=
1
4ia
3
Therefore,
_

0
dx
(x
2
+ a
2
)
2
=
1
2
2i
1
4ia
3
=

4a
3
92. Use residue calculus to evaluate the following integrals:
(a)
_

0
x sinmx
a
2
+x
2
dx =

2
e
am
(a > 0, m > 0),
(b)
_

0
cos mx
(x
2
+a
2
)(x
2
+b
2
)
dx =

2(b
2
a
2
)
(
e
am
a

e
bm
b
) (a > 0, b > 0, m 0, b = a),
(c)
_

cos mx
(x+b)
2
+a
2
dx =

a
e
am
cos bm (a > 0, m 0),
_

sinmx
(x+b)
2
+a
2
dx =

a
e
am
sinbm (a > 0, m 0),
(d)
_

0
cos mx
(x
2
+a
2
)
2
dx =

4a
3
e
am
(1 + am) (a > 0, m 0),
(e)
_

0
cos mx
x
4
+4a
4
dx =

8a
3
e
am
(cos am + sin am) (a > 0, m 0),
(f)
_

0
x
3
sinmx
x
4
+4a
4
dx =

2
e
am
cos am (a > 0, m > 0).
Solution. (a)
z
a
2
+z
2
has a simple pole in the upper half plane, which is at z = ai.
_

xcos mx
a
2
+ x
2
dx + i
_

xsin mx
a
2
+ x
2
dx =
_

xe
mxi
a
2
+ x
2
dx
= 2i Res[
ze
mzi
a
2
+ z
2
, ai]
= 2i
aie
am
2ai
= i e
am
,
where we close the path in the upper half plane for the last integral involving e
imz
, since
m > 0. The rst integral in the rst line is of course 0 because the integrand is odd. So,
_

xsin mx
a
2
+ x
2
dx = e
am
.
SOLUTION SET IV FOR 18.075FALL 2004 9
Note that
x sinmx
a
2
+x
2
is even, we have
_

0
xsinmx
a
2
+ x
2
dx =
1
2
_

xsin mx
a
2
+ x
2
dx =

2
e
am
.
(b)
1
z
2
+a
2
has a simple pole on the upper half plane, which it at z = ai. Since m 0, we
get
_

cos mx
x
2
+ a
2
dx + i
_

sin mx
x
2
+ a
2
dx =
_

e
mxi
x
2
+ a
2
dx
= 2i Res[
e
mxi
z
2
+ a
2
, ai]
= 2i
e
am
2ai
=
e
am
a
,
where we close the path in the upper half plane. The second integral in the rst line
vanishes because the integrand is odd. So,
_

cos mx
x
2
+ a
2
dx =
e
am
a
.
Since
cos mx
x
2
+a
2
is even, we have
_

0
cos mx
x
2
+ a
2
dx =
1
2
_

cos mx
x
2
+ a
2
dx =
e
am
2a
.
Similarly,
_

0
cos mx
x
2
+ b
2
dx =
e
bm
2b
.
So,
_

0
cos mx
(x
2
+ a
2
)(x
2
+ b
2
)
dx =
_

0
1
b
2
a
2
(
cos mx
x
2
+ a
2

cos mx
x
2
+ b
2
) dx
=
1
b
2
a
2
(
_

0
cos mx
x
2
+ a
2
dx
_

0
cos mx
x
2
+ b
2
dx)
=
1
b
2
a
2
(
e
am
2a

e
bm
2b
)
=

2(b
2
a
2
)
(
e
am
a

e
bm
b
).
(c) The given integrals are evaluated by the standard prescription as follows.
_

cos mx
(x + b)
2
+ a
2
dx = Re
_

e
imx
(x + b)
2
+ a
2
dx = Re
_

e
im(xb)
x
2
+ a
2
dx
_

sin mx
(x + b)
2
+ a
2
dx = Im
_

e
imx
(x + b)
2
+ a
2
dx = Im
_

e
im(xb)
x
2
+ a
2
dx,
10 SOLUTION SET IV FOR 18.075FALL 2004
where
_

e
imx
x
2
+ a
2
dx = 2i Res
z=ia
e
imz
z
2
+ a
2
=

a
e
ma
.
Hence,
_

cos mx
(x + b)
2
+ a
2
dx = Re
_
e
imb

a
e
ma
_
=

a
e
ma
cos mb,
_

sinmx
(x + b)
2
+ a
2
dx = Im
_
e
imb

a
e
ma
_
=

a
e
ma
sinmb.
(d) We can calculate the real part of
_

e
imx
(x
2
+a
2
)
2
dx by noticing that our function is
even. As in question 91(d), we have one pole in the upper half-plane. This is a double pole
at z = ia.
Res(ia) =
_
d
dz
_
(z ia)
2
e
imz
(z
2
+ a
2
)
2
__
z=ia
=
_
d
dz
_
e
imz
(z + ia)
2
__
z=ia
=
_
(z + ia)
2
ime
imz
e
imz
2(z + ia)
(z + ia)
4
_
z=ia
=
(2ia)
2
ime
ma
e
ma
4ia
(2ia)
4
= ie
ma
_
ma + 1
4a
3
_
Therefore,
_

0
cos mx
(x
2
+ a
2
)
2
dx =
1
2
Re
__

e
imx
(x
2
+ a
2
)
2
dx
_
=
1
2
Re
_
2i ie
ma
_
ma + 1
4a
3
__
= e
ma
_
ma + 1
4a
3
_
(e) Clearly,
_

0
cos mx
x
4
+ 4a
4
dx =
1
2
Re
_

e
imx
x
4
+ 4a
4
dx.
We calculate the last integral involving e
imz
by closing the path in the upper half plane
since m > 0. The poles of the integrand occur at z
4
+ 4a
4
= 0 and are all simple. The
SOLUTION SET IV FOR 18.075FALL 2004 11
poles that lie in the upper half plane are z
1
= e
i/4

2a and z
2
= e
3i/4

2a. Accordingly,
_

e
imx
x
4
+ 4a
4
dx = 2i
_
Res
z=z
1
e
imz
z
4
+ 4a
4
+ Res
z=z
2
e
imz
z
4
+ 4a
4
_
= 2i
_
e
imz
1
4z
3
1
+
e
imz
2
4z
3
2
_
=
2i
16a
3
[(cosma + i sin ma)(1 + i) + (cos ma i sin ma)(1 i)]
=

4a
3
(cos ma + sin ma)
It follows that
_

0
cos mx
x
4
+ 4a
4
dx =

8a
3
(cos ma + sin ma).
(f) Once again, we can calculate the imaginary part of
_

0
x
3
e
imx
x
4
+4a
4
dx noting that our
function is even and m > 0. The poles occur when z
4
+ 4a
4
= 0 z =

2ia
2
, as in
part (e) above. Thus the roots are:
a
1
=

2ia
2
= a

2(
1

2
+ i
1

2
) = a + ia since a > 0.
a
2
= a
1
= a ia.
a
3
=
_
2ia
2
= ia a.
a
4
= a
3
= ia + a.
Thus there are two poles in the upper half-plane: a
1
and a
3
(both simple poles).
Res(a
1
) = lim
za
1
(z a
1
)f(z)
= lim
za
1
(z a ia)z
3
e
imz
z
4
+ 4a
2
= lim
za
1
z
3
e
imz
(z + a + ia)(z
2
+ 2ia
2
)
=
(a + ia)
2
e
im(a+ia)
8ia
2
Res(a
3
) = lim
za
3
z
3
e
imz
(z + ia a)(z
2
2ia
2
)
=
(ia a)
2
e
imama
8ia
12 SOLUTION SET IV FOR 18.075FALL 2004
Therefore,
_

0
x
3
sin mx
x
4
+ 4a
4
dx =
1
2
Im
__

x
3
e
imx
x
4
+ 4a
4
dx
_
=
1
2
Im
_
2
_
(a + ia)
2
e
imama
(ia a)
2
e
imama
8a
__
=
1
2
Im
_

2
i(e
ima
e
ma
+ e
ima
e
ma
)
_
=

2
e
ma
_
e
ima
+ e
ima
2
_
=

2
e
ma
cos ma