Você está na página 1de 6

Course 214: Hilary Term 2007

Complex Analysis
Section 8: Applications of Cauchys Residue
Theorem
David R. Wilkins
First Version - January 25, 2007
c David R. Wilkins 2007
Copyright

Applications of Cauchys Residue Theorem


Lemma 7.1 Let R be a positive real number, and let f be a continuous
complex-valued function defined everywhere on the semicircle SR , where
SR = {z C : |z| = R and Im[z] 0}.
Suppose that there exists a non-negative real number M (R) such that |f (z)|
M (R) for all z SR . Then
Z


M (R)
isz

f (z)e dz

s
R

for all s > 0, where R : [0, ] C is the path with [R ] = SR defined such
that R () = Rei for all [0, ].
Proof It follows from the definition of the path integral that
Z
Z
isz
f (z)e dz =
f (R ())eisR () R0 () d
R
Z0

=
f (Rei )riRs cos Rs sin iRei d.
0

Therefore
Z



Z

isz

f (z)e dz R

|f (Rei )||eiRs cos Rs sin | d


Z
RM (R)
eRs sin d.
0

Now sin 2/ when 0 /2, and therefore


Z
Z
2
2
2Rs

Rs sin
e
d
e d
.
2Rs
0
0
Also
Z

Rs sin

d =

eRs sin d.

(This follows on making the substitution that replaces by .) Therefore


Z

eRs sin d
.
Rs
0
It follows that

as required.

isz

f (z)e


M (R)
dz
,
s

Example We shall apply Cauchys Residue Theorem (Theorem 6.16) and


Lemma 7.1 in order to evaluate
Z
eisx
dx
2
2
x + a
when s > 0.
Let a be a positive real number, let R be a real number satisfying R > a,
and let R : [0, ] C be the path that sends [0, ] to Rei . (Thus R ()
traverses a semicircle of radius R in the upper half of the complex plane
from R to R as increases from 0 to . Now it follows from the Triangle
Inequality that |z 2 | |z 2 + a2 | + |a2 |, and thus |z 2 + a2 | |z|2 a2 for all
complex numbers z, and therefore


1
1


z 2 + a2 R 2 a2
for all complex numbers z satisfying |z| R. It now follows from Lemma 7.1
that
Z

isz


e



dz


2
2
2
s(R a2 )
R z + a
for all real numbers s and R satisfying s > 0 and R > a. Therefore
Z
eisz
lim
dz = 0.
R+ z 2 + a2
R
2

Now the function f has poles at ia and ia. Moreover




eisz
eisz
esa
lim (z ia) 2
=
lim
=
,
zia
zia z + ia
z + a2
2ia
eisz
as a simple
z 2 + a2
pole at ia with residue esa /2ia. Thus if we apply Cauchys Residue Theorem
(Theorem 6.16) in order to evaluate the path integral of this function around
the boundary of the set
and therefore the meromorphic function that sends z to

{z C : |z| R and Im[z] 0},


we find that
Z R
R

eisx dx
dx + lim
R+
x 2 + a2

Z
R

eisz
esa
esa
=
2i

=
z 2 + a2
2ia
a

when s > 0. If we then take the limit of the left hand side of this identity as
R +, we find that
Z
eisx
esa
dx
=
.
2
2
a
x + a
when s > 0. This formula does not hold when s 0. And indeed, if we take
the complex conjugate of the above identity we find that
Z isx
e
esa
dx
=
.
2
2
a
x + a
when s > 0. It follows that
Z

eisx
e|s|a
dx
=
.
x 2 + a2
a

when |s| =
6 0. This identity also holds when s = 0, though this does not
follow from the above calculations.
Example We can also evaluate the above integral by applying Cauchys
Residue Theorem to the path integral taken around the boundary of a rectangle in the complex plane with vertices at R, R, R + iM and R + iM ,
where R and M are large positive real numbers.
Let a be a real number, and let R and M be real numbers satisfying
R > a and M > a. The inequality


1
1


z 2 + a2 R 2 a2
3

is satisfied for all complex numbers z for which |z| > a. It follows from this
that
Z

Z M
Z M


eisz
1
1
is(Riy)


dz
|e
| dy = 2
esy dy

2 + a2
2 a2
2
z
R
R

a
[R,R+iM ]
0
0
M

.
s(R2 a2 )
Similarly
Z




eisz
M

dz
,

2
2
2
z +a
s(R a2 )


2M esM
eisz
dz
.
z 2 + a2 M 2 a2

[R,R+iM ]

and

[R+iM,R+iM ]

If we take, for example, M = R in these inequalities, and let R tend to +,


we find that
Z
eisz
lim
dz = 0,
R+ [R,R+iR] z 2 + a2
Z
eisz
lim
dz = 0,
R+ [R,R+iR] z 2 + a2
Z
eisz
lim
dz = 0.
R+ [R+iR,R+iR] z 2 + a2
Also, Cauchys Residue Theorem ensures that
Z
Z
esa
eisz
eisz
=
dz
+
dz
2
2
2
2
a
[R,R] z + a
[R,R+iR] z + a
Z
Z
eisz
eisz

dz

dz
2
2
2
2
[R+iR,R+iR] z + a
[R,R+iR] z + a
when s > 0. It follows that
Z isx
Z
e
eisz
esa
dx
=
lim
dz
=
2
2
R+ [R,R] z 2 + a2
a
x + a
when s > 0.
Example Let be a real number satisfying 0 < < 1. We evaluate the
integral
Z
x
dx
x(x + 1)
0
4

through an application of Cauchys Residue Theorem. Let


D = C \ {x R : x 0},
so that D is the open set obtained on removing the negative real axis from the
complex plane, let log: D C denote the principal branch of the logarithm
that sends rei to log r + i for all real numbers r and satisfying r > 0 and
< < , and let z = exp( log z) for all z D. Then the function f
z
that sends z D \{1} to
is a meromorphic function on D. The only
z(z 1)
pole of this function that lies within the open set D is a simple pole at z = 1
with residue 1. Let R and be real numbers satisfying Rp
> 1 and 0 < < 1,
3
and let [ 4 , ] be determined such that R + i = R2 + 2 exp(i ).
It follows from Cauchys Residue Theorem (Theorem 6.16) that
Z
Z
f (z) dz
f (z) dz
[R+i,+i]

Z
Z

f (z) dz +
f (z) dz
[Ri,i]

R,

= 2i,
where : [ 34 , 34 ] C is the path from i to + i that sends
t [ 43 , 34 ] to eit , and R, : [ , ] C is the path from R i
p
to R + i that sends t [ , ] to R2 + 2 eit . [Thus (t) traverses a
three-quarters of a circle of radius about zero in the anti-clockwise direction
as t increases from 34 to 43 , and R, (t) traverses most of a circle of rap
dius R2 + 2 about zero as t increases from to .] Now the inequality
> 0 ensures that
Z
lim
f (z) dz = 0.
0

Also
lim

f (z) dz =

R,

f (z) dz,

where R : [, ] C is the path that sends t [, ] to Reit . It follows


that
Z

Z
Z
2i
f (z) dz = lim+
f (z) dz
f (z) dz
0

lim

0+

[R+i,+i]
0

Z

f (x + i) dx

[Ri,i]

f (x i) dx

=
=

lim (f (x + i) f (x i)) dx

+
R 0
Z R

lim (f (x + i) f (x i)) dx

0+

Now
(x + i)
0 (x + i)(x 1 i)
lim+ exp( log(x + i))

lim+ f (x + i) =

lim+

x(x + 1)
exp((log x + i))
exp( log x) exp(i))
=
=
x(x + 1)
x(x + 1)

x
.
= ei
x(x + 1)

Similarly
lim exp( log(x i))

lim f (x i) =

0+

x(x + 1)
exp((log x i))
exp( log x) exp(i))
=
=
x(x + 1)
x(x + 1)

x
.
= ei
x(x + 1)

0+

It follows that
lim+ (f (x + i) f (x i)) =

=
Therefore

ei ei

R+

It follows that
Z +
0

x
x(x + 1)

2ix sin
x(x + 1)

x
dz = 2i
2i sin()
0 x(x + 1)
Now the inequality < 1 ensures that
Z
lim
f (z) dz = 0.
Z

f (z) dz.

x
dz = lim
R+
x(x + 1)

when 0 < < 1.


6

Z
0

x
dz =
x(x + 1)
sin