Applications of Residue

UNIT 8 APPLICATIONS OF RESIDUE

Structure Page No.
8.1 Introduction 35
Objectives
8.2 Proper Definite Integerals 35
8.3 Improper Definite Integrals 39
Improper Integrals Involving Trigonometic Functions
Indented Contour Integrals
8.4 Summary 56
8.5 Solutions/Answers 57

8.1 INTRODUCTION
An important feature of the complex analysis is its ability to solve the problems
of real analysis. In this unit we will discuss the applications of complex
analysis methods to solve the problems related to integration of real valued
functions of a real variable. You know that the Cauchy’s residue theorem gives
the evaluation of the contour integral of the function in terms of the sums of the
residue of each of its singular point inside the contour. In this unit the Cauchy’s
residue theorem is applied for the evaluation of definite integrals, trigonometric
integrals and improper integrals occurring in real analysis and applied
mathematics.

We have started the unit discussing the evaluation of proper definite integrals
involving trigonometric functions in Sec. 8.2. Improper definite integrals are
discussed in Sec. 8.3. Here we have introduced the notion of Cauchy principal
value and used it to evaluate improper real integrals and improper integrals
involving trigonometric functions which usually occur in the application of the
Fourier transforms. Indented contour integrals which help us in finding the
contour integrals when a singularity is lying on the real axis are also discussed
in this section.

Objectives
After studying this unit, you should be able to:
• use the Cauchy’s residue theorem for evaluating definite integrals,
trigonometric integrals and improper integrals of real functions;
• obtain the Cauchy principal value of an improper integral;
• compute improper integrals involving trigonometric functions; and
• compute indented contour integrals.

8.2 PROPER DEFINITE INTEGRALS

As an application of the residue theorem for the evaluation of real integrals, we
first consider real definite integrals involving trigonometric functions. The
basic idea here is to first transform the given integral to associated contour
integral which is then evaluated by using the residue theorem. Let us now see
how this is done.
35
Applications of Let R (sin θ , cosθ ) be a rational function. Then integrals of the form
Analytic Functions
2π

∫ R (sin θ , cosθ ) dθ
0
can be transformed by writing
z − z −1 z + z −1
z = eiθ (0 ≤ θ ≤ 2π ), sin θ = , cosθ =
2i 2
into integrals of the form
∫ f ( z ) dz
C
where C is the positively oriented unit circle (see Fig.1) and
1  z − z −1 z + z −1 
f ( z ) = R  , (1)
iz  2 i 2 
Then by the residue theorem
2π
Fig. 1
∫ R (sin θ , cosθ ) dθ = 2π i ∑ Res [ f ( z )] .
0 C
(2)

The sum ∑ extends over all the residues of f (z ) inside the unit circle C.
C

We illustrate the method through some examples.

Example 1: Evaluate the integral

2π
cos 2θ dθ
I= ∫ 5 − 4 cosθ . (3)
0
Solution: For the values of z lying on the unit circle C : { | z | = 1} , we have
z = eiθ , (0 ≤ θ ≤ 2π ) and we can write
z 2 = cos 2θ + i sin 2θ , z −2 = cos 2θ − i sin 2θ .
and
z 2 + z −2 z 2 − z −2
cos 2θ = , sin 2θ = .
2 2
The integrand in Eqn. (3) then becomes
1 ( z 2 + z −2 )
i ( z 4 + 1)
f ( z) = 2 = .
iz [5 − 2 ( z + z −1)] 2 z 2 ( z − 2) (2 z − 1)

The singularities of f (z ) lying inside C are the poles at the points z1 = 0 of

order 2 and z2 = 1 , a simple pole. Now we compute residues at these points.
2
d
Res [0, f ( z )] = lim z 2 f ( z )
z → 0 dz

d  i ( z 4 + 1)  5i
= lim  = .
z →0 dz 2 ( 2 z 2 − 5 z + 2) 
  8
1   1
Res  , f ( z ) = lim1  z −  f ( z )
2  z →
2  2
 i ( z 4 + 1)  − 17i
= lim1  2 = .
z → 2 4 z ( z − 2)  24
36
 
Then the residue theorem can be used to conclude that Applications of Residue
2π
cos 2θ dθ  5i 17i 
∫ 5 − 4 cosθ = 2π i  8 − 24 
0
π
= .
6
***

Example 2: Evaluate the integral

π
dθ
I =∫ 2 2
, (a > 0) .
0
a + cos θ
Solution: By using the identity 2 cos2 θ = 1 + cos 2θ and the substitution
φ = 2θ , the given integral reduces to
2π
dφ
I= ∫ (2a 2 + 1) + cos φ .
0

Using z = eiϕ , we find that

1 dz
I =∫
 z + z  iz
−1
C
(2a 2 + 1) +  

 2 
dz
or I = −2i ∫
C
z + 2 (2a 2 + 1) z + 1
2

dz
or I = −2i ∫ .
2 2 2 2
C ( z + 2 a + 1 + 2 a a + 1) ( z + 2 a + 1 − 2 a a + 1)
We find that for the integrand
1
f ( z) =
( z + 2a 2 + 1 + 2a a 2 + 1) ( z + 2a 2 + 1 − 2a a 2 + 1)
singular points are the poles given by

z1 = − (2a a 2 + 1 + 2a 2 + 1), z2 = 2a a 2 + 1 − (2a 2 + 1) .

Recalling that a is a positive real number, we find that | z1 | > 1 and | z2 | < 1 .
In other words, z2 lies inside the contour C . Therefore the residue at z2 is
given by
Res [ z 2 , f ( z )] = b1 = lim ( z − z 2 ) f ( z )
z →z2

 1 
or b1 = lim  
(
z → 2a a 2 +1 − ( 2 a 2 +1)  2 2 
 z + 2 z a + 1 + 2a + 1 
1
= .
4a a 2 + 1
From the residue theorem we then conclude that
1 π
I = (2π i ) (−2i ) = .
4a a + 1 a a 2 + 1
2

***
37
 Applications of Example 3: Evaluate the integral
Analytic Functions 2π
I = ∫ cos 2n θ dθ .
0
Solution: We transform the above integral to the form
2n
 z + z −1  dz − i ( z 2 + 1) 2 n
I = ∫  
 = ∫ dz .
C 
2  i z 2 2n C
z 2n +1
( z 2 + 1) 2n
The integrand f ( z ) = has a pole of order 2n + 1 at z = 0 . Thus by
z 2n +1
the residue theorem,
−i
I = 2π i 2 n Res [0, f ( z )] .
2
Using the method of finding residues, we have
π 1 d 2n 2
I = 2n −1 lim 2 n
( z + 1)2 n . (4)
2 z →0 ( 2n)! dz
Expanding the r.h.s of Eqn. (4), we get
π d 2n  2n   2n   2n  
I = 2 n −1 lim 2n   ( z 2 )2 n + ⋅ ⋅ ⋅ +   ( z 2 ) n + ⋅ ⋅ ⋅ +   
2 (2n)! z → 0 dz  0  n  2n  
n n n!
  = Cm =  2n  
m m !(n − m) ! π  2n 
= lim    4 n ( 4 n − 1) ⋅ ⋅ ⋅ ( 2 n + 1) z 2n
+ ⋅ ⋅ ⋅   2 n ( 2 n − 1) ⋅ ⋅ ⋅ 2 . 1 + 0 + ⋅ ⋅ ⋅ 0 
2 2n −1 (2n)! z →0  0  n
  
π  2n  π  2n 
or I = 2n −1   (2n) (2n − 1) ... 2.1 = 2n −1   .
2 (2n) !  n  2 n
***

You may observe here in Example 3 that the Res [0, f ( z )] of the function
( z 2 + 1) 2n
f ( z) = may also be obtained easily by using the Laurent series.
z 2n+1

Example 4: Evaluate the integral

ez
I = ∫ n +1 dz .
C
z
where C is positively oriented unit circle. Hence deduce that
2π
2π
∫e
cos θ
cos (n θ − sin θ ) dθ = .
0
n!
ez
Solution: The integrand f ( z ) = has a pole of order n + 1 at z = 0 . Then
z n +1
by the residue theorem,
1 d n n +1 2π i
I = 2π i Res [0, f ( z )] = (2π i ) lim n
( z f ( z )) = . (5)
z →0 n ! dz n!
Putting z = eiθ = cosθ + i sin θ , the given integral reduces to
2π
 e cos θ + i sin θ ie iθ  2π

∫ ∫e
cos θ i (sin θ − nθ )
I=  ( n +1) iθ  dθ = i e dθ
0  e  0
2π

∫e
cos θ
or I =i [cos (sin θ − nθ ) + i sin (sin θ − nθ )] dθ . (6)
0
38
Comparing Eqns. (5) and (6) and by equating the imaginary parts, we have Applications of Residue
2π
2π
∫e
cos θ
cos (n θ − sin θ ) dθ = .
0
n!
***

You may now check your understanding of the method while solving the
following exercises:

E1) Evaluate the integral

2π
dθ
∫ a + b cosθ (a > b > 0) .
0

E2) Show that, if a > b > 0 ,

2π

∫
sin 2 θ dθ 2π
(
= 2 a − a 2 − b2 .
a + b cosθ b
)
0

E3) Evaluate the following integrals:

2π 2π
cos 2θ dθ dθ
i) ∫ 13 − 12 cosθ
ii) ∫ (1 + 2a cosθ + a 2 ) 2 (−1 < a < 1) .
0 0

E4) Prove that, for any positive integer n ,

2π
2π  2n 
∫ (sin θ ) dθ = 4n  n  .
2n

b
You know from your knowledge of real analysis that a real integral ∫ f ( x) dx
a
is called an improper integral if
i) one or both of the limits of integration are not finite
ii) the integrand has infinite discontinuities at a or at b (a, b finite) or at
some point c, a < c < b .

In the next section we shall take up the improper integrals of the continuous
function f ( x) of real variable x over the semi-infinite and infinite intervals.

8.3 IMPROPER DEFINITE INTEGRALS

∞
First we consider the improper integrals of the type ∫ f ( x) dx .
−∞
Let f (x) be a continuous function of the real variable x on the interval [0, ∞[ .
From your knowledge of real analysis, you may recall that, the improper
integral of f over [0, ∞[ is defined as
∞ r

∫ f ( x) dx = lim
r →∞ ∫ f ( x ) dx (7)
0 0
provided the limit on the right exists. 39
Applications of If f is defined for all real x , then the existence (convergence) of the improper
Analytic Functions
∞
integral ∫ f ( x) dx , is equivalent to the existence of the two limits
−∞
∞ r1 0 0

∫ f ( x) dx = lim
r1 → ∞ ∫ f ( x) dx and ∫ f ( x)dx = lim
r2 → ∞ ∫ f ( x) dx .
0 0 −∞ − r2
and we write
∞ r1 0

∫ f ( x) dx = lim
r1 → ∞ ∫ f ( x ) dx + lim
r2 → ∞ ∫ f ( x) dx . (8)
−∞ 0 − r2
To apply the residue theorem, we reformulate the above integral as the limit
r ∞
lim
r →∞ ∫ f ( x) dx = P.V . ∫ f ( x ) dx (provided the limit exists). (9)
−r −∞
The limit is called the Cauchy principal value (abbreviated as P.V.). The
∞
existence of improper integral ∫ f ( x ) dx implies the existence of the Cauchy
−∞
principal value (9) and that value is the number to which the integral (8)
converges. This is because
r 0 r

∫ f ( x) dx = ∫ f ( x ) dx + ∫ f ( x) dx
−r −r 0
and the limit as r → ∞ of each of the integrals on the right exists when the
integral (8) converges.

Remark: The existence of P.V. does not guarantee the convergence of

improper integral as illustrated in the following example.

Example 5: Consider the function f ( x ) = x over ] − ∞, ∞ [ . It can be seen that

∞
the improper integral ∫ f ( x ) dx does not exist. This is because
−∞
r1
1 2
∫ x dx =
2
(r1 − r22 ) , which does not tend to a limit as r1, r2 → ∞
− r2
r
1
r →∞ ∫
2
independently. On the other hand lim x dx = lim [r − r 2 ] = 0 . Thus, the
r →∞ 2
−r
∞
improper integral ∫ xdx does not exist/converge but the Cauchy principal
−∞
value of this integral exists and is zero.
***

However, you may note that when f (x) for − ∞ < x < ∞ , is an even function
and the Cauchy principal value (9) exists, then both the integrals (7) and (8)
∞ ∞ ∞
converge and the P.V . ∫ f ( x ) dx = ∫ f ( x ) dx = 2 ∫ f ( x ) dx . (10)
−∞ −∞ 0
The computation of improper integral is based on the following idea. Let
D ⊆ C be the domain containing the closed upper half plane
H u = {z : Im z ≥ 0} . Let z1, K, z k be the distinct points in the open upper half
40
 Applications of Residue
plane H u and let
f : D \ {z1, K, zk } → C
be an analytic function. Now we choose r > 0 large enough so that r > | zα |
for 1 ≤ α ≤ k . We consider then the positively oriented contour γ contained in
D and is the composition of the segment [−r , r ] and the semicircle Cr from
r to − r (see Fig. 2).

Fig. 2

Therefore, the Cauchy residue theorem can be applied to the function f and
we have
r k

∫ f ( x) dx + 2 ∫ f ( z ) dz = 2 ∫ f ( z ) dz = 2π i ∑ Res [ z j , ( f )] .
−r Cr γ j =1

If we can ensure that the contribution from the semicircle integral ∫ f ( z ) dz is

Cr
negligible (zero) as r tends to infinity then
∞ r k
P.V . ∫ f ( x ) dx = lim ∫ f ( x) dx = 2π i ∑ Res [ z j , ( f )] .
r →∞
−∞ −r j =1
∞
If the improper integral ∫ f ( x) dx exists, then
−∞
∞ k

∫ f ( x ) dx = 2π i ∑ Res [ z j , ( f )] . (11)
−∞ j =1

In case f (x ) is an even function then Eqn. (11) reduces to

∞ ∞ k
1
∫ f ( x ) dx =
2 −∫∞
f ( x ) dx = π i ∑
j =1
Res [ z j , ( f )] . (12)
0

We now illustrate the method through an example.

Example 6: Evaluate the integral

∞
x 2 dx
∫ ( x 2 + 9) ( x 2 + 4) 2
.
0

Solution: Let us first show that the integral

41
Applications of ∞ ∞
x 2 dx 1 x 2 dx
Analytic Functions
∫ ( x 2 + 9) ( x 2 + 4) 2 2 ∫ ( x 2 + 9) ( x 2 + 4) 2
= (13)
0 −∞

converges and then we find its value.

We can observe that the integral on the r.h.s. of Eqn. (13) represents an
integration of the function
z2
f ( z) = 2
( z + 9) ( z 2 + 4 ) 2
along the entire real axis in the complex plane. This function has isolated
sigularities at z = ± 3i (simple poles) and z = ± 2i (double poles) and is
analytic everywhere else. When r > 3 , the singular points z = 3i and z = 2i of
f lie in the upper half plane in the interior of the semicircular region bounded
by the segment [−r , r ] of the real axis and the upper half Cr of the circle
| z | = r from − r to r (see Fig. 3).
Fig. 3
Integrating (using the residue theorem) f counterclockwise around the
boundary of this semicircular region, we see that
r

∫ f ( x) dx + ∫ f ( z ) dz = 2π i (b1 + b2 )
−r Cr

where b1 = Res [2i, f ( z )] and b2 = Res [3i, f ( z )] . Now we find b1 and b2 ,

d z2 
b1 = lim  ( z − 2i )2 2 2 2 
z → 2i dz
 ( z + 9) ( z + 4) 
d  z2 
= lim  2 
z → 2 i dz  ( z + 9) ( z + 2i ) 
2
 
 2 z ( z + 9) ( z + 2i) − z 2{2 z ( z + 2i )2 + 2( z 2 + 9) ( z + 2i)} 
2 2
= lim  
z → 2i
 ( z 2 + 9) 2 ( z + 2i ) 4 
− 13i
= .
200
and
z2 3i
b2 = lim ( z − 3i) 2 2
= .
z → 3i ( z − 3i) ( z + 3i) ( z + 4) 50
Thus
r
 − 13i 3i  π
∫ f ( x) dx + ∫ f ( z ) dx = 2π i  200 + =
50  100
.
−r Cr
Therefore, we have
r
π
∫ f ( x) dx =
100
− ∫ f ( z ) dz (14)
−r Cr
which is valid for all values of r greater than 3.
We now show that the value of the integral on the right of Eqn. (14)
approaches 0 as r → ∞ .
If z is a point on the half circle C r then | z 2 | = | z |2 = r 2 and by using the
triangle inequality | z + w | ≥ || z | − | w | | , we get

| ( z 2 + 9) ( z 2 + 4) 2 | ≥ (|| z |2 −9 |) (|| z |2 −4 |) 2 = (r 2 − 9) (r 2 − 4) 2 .

42
Consequently, we get the required estimate as follows
 Applications of Residue
z 2 dz r2
∫ f ( z ) dz = ∫ 2
( z + 9 ) ( z 2
+ 4 ) 2
≤
( r 2
− 9 ) ( r 2
− 4 ) 2
L (C r )
Cr C r

where L (Cr ) = π r is the length of the semicircle Cr . We thus have

z 2 dz π r2
∫ ( z 2 + 9) ( z 2 + 4) 2 ≤ ( r 2 − 9) ( r 2 − 4) 2 .
C r

Now as r → ∞ , the right hand side of the above inequality goes to 0 and
therefore ∫ f ( z )dz = 0 .
Cr
We then have from Eqn. (14)
∞ r
π
P.V . ∫ r →∞ ∫
f ( x) dx = lim
100
. f ( x) dx =
−∞ −r
Since integrand is even and the Cauchy principal value exists therefore
∞
x2 π
∫ 2
( x + 9) ( x + 4) 2
dx =
200
.
0
***

We now state a result in the form of a theorem which can be used for finding
the Cauchy principal value of the integral of rational function f , of the
p ( x)
form f ( x) = , where p (x ) and q (x) are polynomials, over ] − ∞, ∞ [ . If
q ( x)
the assumptions of this theorem are satisfied then the condition
lim ∫ f ( z ) dz = 0 is trivially satisfied.
r →∞
Cr

Theorem 1: Let p and q be two polynomials having real coefficients with

deg (q) ≥ deg ( p ) + 2 . Suppose q ( x) ≠ 0 for all real x . Let z1 , z 2 , ..., z k be
p ( z)
the complete set of poles of the function f ( z ) = in the upper half plane.
q ( z)
Then
∞ k
P.V . ∫ f ( x) dx = 2π i ∑ Res [ z j , f ( z )] . (15)
−∞ j =1

We shall not be proving the theorem here but illustrate it through examples.
– –

Example 7: Find the Cauchy principal value of the integral

∞
x dx
∫ ( x + 1) ( x 2 + 2 x + 2)
2
.
−∞

Solution: We have p ( z ) = z and q ( z ) = ( z 2 + 1) ( z 2 + 2 z + 2) . Clearly

deg (q) = 4 > deg ( p) + 2 = 3 . The zeros of q (z ) are z = ± i and z = −1 ± i (thus
q ( x) ≠ 0 for any real x ). All the zeros of q(z ) are the poles of the function
p ( z)
f ( z) = . Of these poles, z1 = i (simple pole) and z2 = −1 + i (simple pole)
q ( z)
lie in the upper half plane. Note that p ( z1 ) ≠ 0 ≠ p ( z 2 ) . Also note that
q′ (i) ≠ 0 and q′ (−1 + i ) ≠ 0 . We now compute the residues at these poles.
43
Applications of
Analytic Functions  p ( z )  p (i ) 1 − 2i
Res i, = =
 q ( z )  q′ (i ) 10
and
 p ( z )  p (−1 + i ) − 1 + 3i
Res (−1 + i) = = .
 q ( z )  q′ (−1 + i ) 10
Now using Theorem 1, we have from Eqn. (15)
∞
x dx 1 − 2i − 1 + 3i  − π
P.V . ∫ 2 2
( x +!) ( x + 2 x + 2)
= 2π i 
 10
+
10 
=
5
.
−∞

***

Let us look at another example.

Example 8: Let k , n ∈ Z, 0 ≤ k < n . Then show that

∞
x 2k π
∫ 1 + x 2n
dx =
 π 
.
−∞ n sin (2k + 1) 
 2n 
Solution: We consider p ( z ) = z 2 k and q ( z ) = 1 + z 2n . The roots of
q ( z ) = 1 + z 2n in the upper half plane are
 ( 2 j + 1) π i 
 
zj = e  2n
 , 0 ≤ j < n, j an integer.
The derivatives q′ (z ) are not equal to zero at all these points of the upper half
plane. Hence, all z ′j s are the simple roots of q (z ) and therefore the simple
z 2k
poles of f ( z ) = . Moreover, p ( z j ) ≠ 0 . By computing the residues we
1 + z 2n
have
p ( z j ) 1 2 k − 2 n +1 1
Res [ z j , f ( z )] = = zj = − z 2j k +1 (Q z 2j n = −1) . (16)
q′ ( z j ) 2n 2n
Further, we have
n −1 n −1  π i ( 2 j +1) ( 2 k +1) 
 2n 
∑z 2 k +1
j = ∑e  

j =0 j =0

 ( 2 k +1)π i  n −1  π i ( 2 k +1) j 
   
=e  2n 
∑e n 

j =0

 2 k +1) π i 

 2n

 1 − e(( 2k +1)π i )
=e  πi
 ( 2 k +1) 
 n 
1− e
1 − (−1)2 k +1
=
 π 
− 2i sin (2k + 1) 
 2n 
2 i
= = . (17)
 π   π
2i sin (2k + 1)  sin  (2k + 1) 
 2n   n
Now using Theorem 1, we get
44
 ∞ Applications of Residue
x 2k n −1

∫ 1 + x 2n dx = 2π i ∑
j=0
Res [ z j , f ( z )]
−∞
n −1
1
= 2π i ∑ − 2n z 2j k +1 [using Eqn. (16)]
j =0

 
 
 1  i 
= (2π i )  −  [using Eqn. (17)]
 2 n  
 sin (2k + 1) π  
  
  n 
π
= .
 π 
n sin  (2k + 1) 
 2n 
***

It is now time for you to check your understanding of the method discussed
above. You may try the following exercises:

E5) Evaluate
∞ ∞
dx x 2 dx
i) ∫ (x + 1) 2
2
ii) ∫ x4 + 1
.
0 0

E6) Show that

∞
1 π (2n − 2) !
∫ 2
( x + 1) n
dx = 2 n − 2
2 [(n − 1)!]2
.
−∞

E7) Show that

∞ ∞
x 2 dx π dx π
i) ∫ 2 2 2
(x + a )
=
2a
, (a > 0) ii) ∫ x 4 + x2 + 1 = 2 3
.
−∞ 0

E8) Find the Cauchy principal value of the integral

∞
dx
∫ 2
x + 2x + 2
.
−∞

E9) Show that

∞
dx π
∫ ( x 2 + a 2 ) ( x 2 + b 2 ) = ab (a + b) , a > 0 and b > 0 .
−∞

We shall now take up in the next sub-section the integrals of the type
∞ ∞

∫ f ( x ) sin x dx or ∫ f ( x ) cos x dx that occur in the theory and applications of

−∞ −∞
the Fourier transforms.

8.3.1 Improper Integrals Involving Trigonometric Functions

We assume that p and q are two polynomials of degree m and n ,

45
Applications of respectively having no factor in common and n ≥ m + 1 . If q (x) has no real
Analytic Functions
zeroes then
∞ ∞
p( x) p( x)
P.V . ∫ q( x) cos x dx, P.V . ∫ q( x) sin x dx
−∞ −∞
are convergent improper integrals. The methods discussed in Sec. 8.3 cannot be
y
e −e
−y applied here directly since
sinh y = 2
2 | sin z |2 = sin 2 x + sinh 2 y and | cos z | = cos2 x + cosh 2 y
increase like sinh y or e y as y → ∞ .
Therefore, the method has to be modified and the modifications are motivated
by the fact that
r r r
p( x) p( x) p( x)e iα x
∫ q( x)
cos α x dx + i ∫
q( x)
sin α x dx = ∫
q( x)
dx
−r −r −r

together with the fact that | eiz | = e − Im z = e − y is bounded in the upper half plane
Hu where Im z ≥ 0 .

We shall explain the details of the method through an example.

Example 9: Show that

∞
cos x π
∫ 2
( x + 1) 2
dx = .
e
−∞

Soution: As we notice that the given integrand is even, it is sufficient to show

that the Cauchy principal value of the integral exists and then to find that value.
We consider the function
eiz
f ( z) = 2
( z + 1) 2
which is analytic everywhere on and above the real axis except at the point
z = i . The singularity z = i (a double pole) lies in the interior of the
semicircular region whose boundary consists of the segment − r ≤ x ≤ r of the
real axis and the upper half plane Cr of the circle | z |= r , (r > 1) from z = r to
Fig. 4 z = − r (see Fig. 4).
Integrating f (z ) anticlockwise along the boundary of the oriented contour
γ = {Cr }U {[−r , r ]} , we get by using the Cauchy’s residue theorem
r

∫ f ( z) dz + ∫ f ( x) dx = 2π i b1 (18)
Cr −r
where
d e iz
( z − i )2 2
b1 = Res [i, f ( z )] = lim
z → i dz ( z + 1) 2
d  e iz 
= lim  
z → i dz  ( z 2 + i ) 2 
 
−i
= .
2e
We have from Eqn. (18)
r
Re ∫ f ( x ) dx = Re (2π ib1 ) − Re ∫ f ( z ) dz .
−r Cr
46
Thus, Applications of Residue
r
π ei z
Re ∫ f ( x ) dx = − Re ∫ dz .
−r
e Cr
( z 2 + 1)2
Now | eiz | = e − y ≤ 1 and | z 2 + 1 | ≥ || z |2 −1 | = (r 2 − 1) .
When z is on Cr , y ≥ 0 then for the estimate of the integral, we have
ei z ei z πr
Re ∫ dz ≤ ∫C ( z 2 + 1)2 dz ≤ (r 2 − 1) 2 → 0, (r → ∞) .
Cr
( z 2 + 1)2 r

Hence
∞
cos x π
∫ f 2
( x + 1) 2
dx = .
2
−∞
***
The evaluation of the integral as discussed above is quite involved and lengthy.
We now state a result which is a generalisation of Theorem 1.
This result can be used to evaluate the integrals directly, especially, avoiding
going into the last step of Example 9.

Theorem 2: Let p and q be the two polynomials with real coefficients and let
α ≥ 0 . Assume that q ( x) ≠ 0 for all real x and also that deg (q) ≥ 2 + deg ( p)
when α = 0 , and deg (q) ≥ 1 + deg ( p) for α ≠ 0 . Let z1 , z 2 , ..., z k be all the
roots of q in the upper half plane. Then
∞ k
p ( x) i α x
P.V . ∫ e dx = 2π i ∑ Res [ z j , f ( z )] (19)
−∞
q( x) j =1

p ( z )e i α z
where f ( z ) = .
q ( z)
This result can be written explicitly as follows:
∞ k
p ( x)
P.V . ∫ cos(a x ) dx = −2π ∑ Im ( Res [ z j , f ( z )]) . (20)
−∞
q ( x) j =1

∞ k
p( x)
P.V . ∫ sin (a x) dx = 2π ∑ Re ( Res [ z j , f ( z )] ) . (21)
−∞
q( x) j =1

We shall not be proving these results here but illustrate them through the
following examples.

Example 10: Evaluate the integral

∞
cos x dx
∫ ( x + a 2 ) ( x 2 + b2 )
2
(a > b > 0) .
−∞

Solution: Let p ( x) = 1 and q ( x ) = ( x 2 + a 2 ) ( x 2 + b 2 ) . We see that

deg (q ) = 4 > 1 + deg ( p) and x = ± a i, ± b i are all the zeros of q (x) . We
further note that q ( x) ≠ 0 for all real x and x = ai and x = b i are the zeros of
q (x) which lie in the upper half plane. Consider the function
p ( z) i z ei z
f ( z) = e = 2 , (α = 1 > 0) .
q( z) (z + a2 ) (z 2 + b2 )

47
Applications of The function f (z ) has the simple poles at z = ai and z = bi which lie in the
Analytic Functions
upper half plane. We now compute the residues at these simple poles.
e iz
Res [ai, f ( z )] = lim ( z − ai) 2
z → ai ( z + a 2 ) ( z 2 + b2 )
eiz
= lim
z → ai ( z + ai ) ( z 2 + b 2 )
− ie − a
= .
2a (b 2 − a 2 )
Similarly,
− ie − b
Res [bi, f ( z )] = .
2b (a 2 − b 2 )
Using the result (20) we can write
∞
cos x dx  − ie − a − ie −b 
P.V . ∫ 2 = − 2π Im 
 2a (b 2 − a 2 ) 2b(a 2 − b 2 ) 
+
−∞
( x + a 2 ) ( x 2 + b2 )  
π e −b
e 
−a
= 2  − 
a − b 2  b a 
but as the integrand is an even function, we get the desired result as
∞
cos x dx π  e −b e − a 
∫ ( x 2 + a 2 ) ( x 2 + b2 ) = a 2 − b 2  b − a  .
−∞

***

You must have observed in Example 10 that you merely have to satisfy the
hypothesis of Theorem 2 for the given integral and then use the results (19) or
(20) or (21), as per the requirement, directly .

Let us consider another example.

Example 11: Find the Cauchy principal value of the improper integral
∞
sin x dx
∫ 2
x + 4x + 5
.
−∞

Solution: Here p ( x ) = 1, q ( x ) = x 2 + 4 x + 5 . Consider the function

eiz
f ( z) = 2 .
z + 4z + 5
The function f (z ) has only a simple pole in the upper half plane, namely,
− 2 + i . We have
eiz
Res [(−2 + i ), f ( z )] = lim ( z + 2 − i ) 2
z →( −2 + i ) z + 4z + 5
iz
e
= lim
z → ( −2 + i ) z + 2 + i

e −2i −1
= .
2i
Using Result (21), we get
∞
sin x dx
P.V . ∫ 2
= 2π Re ( Res [(−2 + i), f ( z )])
−∞
x + 4x + 5
48
 Applications of Residue
 e −2i −1 
= 2π Re  
 2i 
π
=− sin 2 .
e
***

After going through Examples 10 and 11, you must have realised how much
the application of Theorem 2 simplifies the evaluation of improper integrals.
Sometimes, for the evaluation of these integrals, it becomes necessary to use
the application of a lemma called Jordan’s lemma, which we are stating now.

Jordan’s Lemma: Let p and q be two polynomials with real coefficients of

degree m and n , respectively, where n ≥ m + 1 . If Cr is the upper semicircle
z = reiθ for 0 ≤ θ ≤ π , then
ei z p ( z )
r →∞ ∫
lim dz = 0 . (22)
C
q( z)
r

p ( z)
Note that since n ≥ m + 1 , it follows that → 0 as | z | → ∞ .
q ( z)
We shall not be proving the lemma here but illustrate it through examples.
However, we would mention that the proof of the lemma is based on an
inquality, known as
Jordan’s Inequality, according to which
π
π
∫e
− r sin θ
dθ < (r > 0) . (23)
O
r
You may observe in Fig. 5 from the graphs of the functions y = sin θ and
2θ
y=
π

Fig. 5
2θ π
that, sin θ ≥ for 0 ≤ θ ≤ . Thus, if r > 0 , we have
π 2
−2 rθ
− r sin θ π  π
e ≤e 0 ≤ θ ≤ 
 2
and
π π
2 2 − 2 rθ
π
∫e dθ ≤ ∫ e
− r sin θ π dθ = (1 − e − r ) .
0 0
2r
49
Applications of π
Analytic Functions 2
π
∫e
− r sin θ
Hence, dθ < , ( r > 0) .
0
2r
Since the graph of y = sin θ is symmetric with respect to the vertical line
π
θ= in the interval 0 ≤ θ ≤ π , we get Inequality (23).
2

Let us consider the following example.

Example 12: Find the Cauchy principal value of the improper integral
∞
x 3 sin x dx
∫ x4 + 4
.
−∞

Solution: We write
z 3eiz z 3eiz
f ( z) = 4 = .
z + 4 ( z − 1 − i) ( z − 1 + i) ( z + 1 + i ) ( z + 1 − i)
The points z = 1 + i and z = −1 + i , which lie in the upper half plane, are the
simple poles of f , with residues
ei −1 e − i −1
b1 = , b2 = .
4 4
Hence, when r > 2 = | 1 + i | = | −1 + i | and Cr denotes the upper half of the
positively orinted circle | z | = r then by the Cauchy residue theorem, we get
for the contour shown in Fig. 6
r
x 3eix dx
∫ x4 + 4
= 2π i (b1 + b2 ) − ∫ f ( z ) dz . (24)
−r C r

It implies
r
x 3 sin x dx  
Fig. 6 ∫ x4 + 4
= Im [ 2π i (b1 + b2 )] − Im  ∫ f ( z ) dz . (25)
−r C r 
 
Now we find the estimate of Im  ∫ f ( z ) dz  . We have
C r 
z eiz r3
| f ( z) | = 2 ≤ 4
| e iz |
z + 2 z + 2 (r − 2 )

When z is on Cr and since | eiz | ≤ 1 for such a point z , we cannot conclude

that integral of f (z ) along Cr → 0 as r → ∞ . (this is because
π r4
→
/ 0 as r → ∞ ).
(r − 2 ) 4
If we put z = reiθ where 0 ≤ θ ≤ π , r > 2 , we get for z on Cr
π

∫ ∫ | f (re
iθ
f ( z ) dz ≤ r ) | dθ
Cr 0
π
r4
4 ∫
≤ e − r sin θ dθ
(r − 2 ) 0
50
 π Applications of Residue
4 2
2r
∫e
− r sin θ
≤ dθ
(r − 2 ) 4 0
3
πr
≤ (using Jordan’s inequality)
(r − 2 ) 4
and we have

∫ f ( z ) dz → 0 , as r → ∞ .
Cr
Thus
 
Im  ∫ f ( z ) dz  ≤ ∫ f ( z ) dz → 0 , as r → ∞ .
 
 Cr  Cr

Hence, taking r → ∞ in Eqn. (25), we get

∞
x 3 sin x dx
P.V . ∫ = Im [2π i (b1 + b2 )]
−∞
x4 + 4

 2π i i 
= Im  (e + e − i ) 
 4e 
π cos1
= .
e
***

You may now test your understanding of how much you have learnt by doing
the following exercises.

E10) Show that

∞
cos x dx π
∫ = .
( x 2 + 1) 2 e
−∞

E11) Find the Cauchy principal value of the improper integral

∞
( x + 1) cos x
∫ x2 + 4x + 5
dx .
−∞

E12) Evaluate the improper integrals

∞ ∞
cos a x x 2 sin x dx
i) ∫ ( x 2 + b2 )2
dx (a > 0, b > 0) ii) ∫ ( x 2 + 1) ( x 2 + 9) .
0 0

E13) Evaluate the improper integral

∞ ∞
cos n x cos x dx
i) ∫ x 4 + 1 dx ii) ∫ ( x + a)2 + b 2 (b > 0) .
0 −∞

So far, we considered the integrals in which the integrands had no poles lying
on the contour or to be precise, on the real axis. In the next section we shall
consider problems in which the integrands have simple poles lying on the real
axis.
51
Applications of
Analytic Functions
8.3.2 Indented Contour Integrals

Let us consider the evaluation of the improper integrals with singularities

which are simple poles lying on the real axis. How do we modify the methods
of the previous sections? A possible solution is to use indented path, i.e., to
avoid the singularity by moving along a circular arc of small radius around it.
To be more clear consider a function f with the only singularity to be a simple
pole on the real axis at the origin. Then the indented curve C is as shown in
Fig. 7.

Fig. 7

Here C consists of the line segment [− r , − ε ], ε < r , the semicircle γ ε from

− ε to ε , the line segment [ε , r ] , and the semicircle Cr from r to − r . Here
small semicircle γ ε called indentation is introduced to avoid the singularity
of f at the origin. Thus
C = line segment [− r , − ε ] ∪ γ ε ∪ line segment [ε , r ] ∪ Cr
−ε r
and ∫ f ( z ) dz = ∫ f ( x ) dx + ∫ f ( z ) dz + ∫ f ( x) dx + ∫ f ( z ) dz . (26)
C −r γ∈ ε Cr
We hope here that the limit of the integral round the indentation exists as radius
ε tends to zero. We prove here the following result.

Theorem 3: (Indentation Lemma) Suppose that f has a simple pole at

z = z1 , with residue b , and let γ ε be a circular arc with radius
ε : γ ε (θ ) = z1 + ε eiθ (α ≤ θ ≤ β ) . Then

lim
ε→0 ∫ f ( z ) dz = ib ( β − α ) . (27)
γε

Proof: Since Res f ( z ) = b therefore, b = lim ( z − z1 ) f ( z ) . This shows that

z = zi z → z1
given η > 0 , there exists δ > 0 such that | ( z − z1 ) f ( z ) − b | < η whenever
0 < | z − z1 | < δ .
Let 0 < ε < δ . When z = γ ε (θ ) = z1 + ε eiθ , we have γ ε′ (θ ) = i ε eiθ = i ( z − z1 )
and thus
52
 β
Applications of Residue

∫ f ( z ) dz − ib ( β − α ) = ∫ [( f (γ ε (θ )) γ ε′ (θ ) − ib)] dθ
γε α

β
= ∫ [i (γ ε (θ ) − z1 ) f (γ ε (θ )) − ib] dθ
α

< η ( β − α ).
Consequently, lim
ε →0 ∫ f ( z ) dz = ib ( β − α ) .
γε
– –

Let us evaluate an integral using indentation lemma.

Example 13: Evaluate the integral

∞
sin 2 x
∫ x2
dx .
0

Solution: First of all we have to find a complex function f (z ) whose real part
sin 2 x
is when z = x is real. Our experience says that it must be a function
x2
involving exponential. The identity 2 sin 2 x = 1 − cos 2 x , suggests that our
1 − e 2iz
function should be f ( z ) = . The isolated singularity of f (z ) is z = 0 .
2z2
The Laurent series expansion of f (z ) about z = 0 is given by

1 − e 2iz 1
2
= 2 [1 − (1 + 2iz − 2 z 2 + L) ]
2z 2z
−i
= + 2 − L.
z
Thus, the pole at z = 0 is simple, with residue − i . Since the only pole (as it
has no other singularity) z = 0 lies on the real axis we cannot use the Fig. 8
semicircular contour as used in the last section. We make an indentation at 0
as shown in Fig. 8. Our contour C consists of semicircle Cr (r > 2) from r to
− r , segment [−r , − ε ], (0 < ε < r ) , inner semicircle γ ε from ε to − ε (note
here that in the figure it has been traced from − ε to ε , while integrating its
contribution is with minus sign) and segment [ε , r ] . f (z ) is analytic inside
and on C thus by the Cauchy-Goursat theorem
−ε r

∫ f ( x ) dx − ∫ f ( z )dz + ∫ f ( x ) dx + ∫ f ( z )dz = 0 . (28)

−r γε ε C
The first and the third integrals combine to give
r r r
1 − e − 2i x 1 − e 2i x 1 − cos 2 x
∫ x 2 dx + ∫ x 2 dx = ∫ x 2 dx
ε ε ε
r
sin 2 x
= 2∫ 2 dx . (29)
ε
x
As we have seen that z = 0 is a simple pole so applying indentation lemma, we
get
53
Applications of
Analytic Functions
lim
ε→0 ∫
γε
f ( z ) dz = i (π − 0) Res [0, f ( z )] = π . (30)

Finally, we have to estimate the integral ∫ f ( z ) dz . We see that on Cr

Cr
2iz
1− e 2
| f ( z) | = ≤
2z 2 2r 2
π
⇒ ∫ f ( z ) dz ≤
r
Cr

⇒ ∫ f ( z ) dz → 0 as r → ∞ (31)
Cr

Letting r → ∞, ε → 0 , we get from Eqns. (28)-(31)

∞ ∞
sin 2 x sin 2 x π
∫ 2 x 2 dx = π ⇒ ∫ x 2 dx = 2 .
0 0
***

We now state without proof two more results in the form of theorems which
are used in the evaluation of improper integrals. We shall illustrate these
results through examples.

Theorem 4: Let p and q be two polynomials with real coefficients of degrees

m and n , respectively and n ≥ m + 2 . If q has simple zeros at the points
p( z)
x1, x2 , K , xk on the real axis and z1, z2 , K , zl are the poles of f ( z ) =
q( z)
in the upper half-plane {z : Im z > 0} , then
∞ k l
p( x)
P.V . ∫ dx = π i ∑ Res [ x j , f ( z )] + 2π i ∑ Res [ z j , f ( z )] .
−∞
q ( x) j =1 j =1
– –

Theorem 5: Let p and q be two polynomials with real coefficients of degrees

m and n , respectively and n ≥ m + 1 . If q has simple zeros at the points
x1, x2 , K , xk on the real axis and z1, z2 , K , zl are the poles of
e i α z p( z )
f ( z) = , α ≥ 0 in the upper half-plane {z : Im z > 0} , then
q( z )
∞ k l
p ( x)
P.V . ∫ cos (a x) dx = −π ∑ Im ( Res [ x j , f ( z )]) − 2π ∑ Im ( Res [ z j , f ( z )])
−∞
q ( x) j =1 j =1

and
∞ k l
p( x)
P.V . ∫ sin (a x)dx = π ∑ Re ( Res [ x j , f ( z )]) + 2π ∑ Re ( Res [ z j , f ( z )]) .
−∞
q ( x) j =1 j =1

Let us consider the following examples to illustrate the theorems above.

Example 14: Show that

∞
dx
P.V . ∫ x ( x − 1) ( x − 2)
=0.
−∞
54
 p ( z) 1 Applications of Residue
Solution: Let us consider f ( z ) = = . Here we have
q ( z ) z ( z − 1) ( z − 2)
p ( z ) = 1 and q ( z ) = z ( z − 1) ( z − 2) , deg q ( z ) = 3 ≥ deg p ( z ) + 2 and f (z ) has
simple poles, namely z = 0, z = 1, z = 2 , all lying on the real axis (no pole in
the upper half plane). We compute the residues of f (z ) at these simple real
poles.
1 1
Res [0, f ( z )] = lim z =
z → 0 z ( z − 1) ( z − 2) 2
1
Res [1, f ( z )] = lim ( z − 1) = −1
z →1 z ( z − 1) ( z − 2)
1 1
Res [2, f ( z )] = lim ( z − 2) = .
z →2 z ( z − 1) ( z − 2) 2
Using Theorem 4, we get
∞
dx 1 1
P.V . ∫ x ( x − 1) ( x − 2)
= π i  −1+  = 0 .
2 2
−∞
***

Example 15: Show that

∞
sin x dx  1
P.V . ∫ x (1 + x 2 )
=π 1 −  .
 e
−∞

eiz p ( z )
Solution: Consider f ( z ) = . Here p ( z ) = 1 and q ( z ) = z (1 + z 2 ) and
q ( z)
clearly, deg q ( z ) = 3 ≥ deg p ( z ) + 1 . Observe that f has the simple poles at
z = 0, z = ± i . We observe that z = 0 lies on the real axis and z = i lies in the
upper half plane. We have
eiz
Res [0, f ( z )] = lim z = 1 , and
z → 0 z (1 + z 2 )

eiz 1
Res [i, f ( z )] = lim ( z − i ) 2
=− .
z →1 z (1 + z ) 2e
Using Theorem 5, we get
∞
sin x dx
P.V . ∫ = 2π Re ( Res[1, f ( z )]) + π Re ( Res [0, f ( z )])
−∞
x (1 + x 2 )
 1
= π 1 −  .
 e
***
And now some exercises for you.

E14) Show that

∞
dx π
P.V . ∫ 2
( x − i ) ( x − 1)
= dx .
2
−∞

E15) Derive the integration formula

∞
cos (ax) − cos (bx) π
∫ x 2
dx = (b − a ) (a ≥ 0, b ≥ 0) .
2
−∞
55
Applications of E16) Show that
Analytic Functions ∞
cos x dx π sin a
P.V . ∫ a 2
− x 2
=
a
, (a > 0) .
−∞

We now end this unit by giving a summary of what we have covered in it.

8.4 SUMMARY
In this unit, we have covered the following:

1. Applications of complex analysis methods to solve the problems related to

the integration of real valued functions of real variables.

2. The Cauchy’s residue theorem is applied for the evaluation of definite

integrals, trigonometric integrals and improper integrals occurring in real
analysis and applied mathematics.

2π
3. The integral ∫ F (sin θ , cosθ ) dθ can be transformed to an integral
0

∫ f ( z ) dz where C is the unit circle, and z = eiθ (0 ≤ θ ≤ 2π ) . Then

C
2π

∫ F (sin θ , cosθ ) dθ = 2π i × (sum of the residues of f (z ) inside C ).

0

4. If f is a continuous function of the real variable x over the interval

∞
] − ∞, ∞[ , then the improper integral ∫ f ( x) dx exists if
−∞
∞ 0 r

∫ f ( x) dx = lim
r→∞ ∫ f ( x )dx + lim
r→∞ ∫ f ( x) dx
−∞ −r 0
provided the two limits on the right exist.

5. Assigned to an improper integral is the Cauchy principal value which is

defined as the limit
∞ r
P.V . ∫ f ( x ) dx = lim
r→∞ ∫ f ( x) dx
−∞ −r
provided the limit exists.

6. In the case when f ( x), − ∞ < x < ∞ , is an even function and

∞ ∞ ∞ ∞
P.V . ∫ f ( x) dx exists then P.V . ∫ f ( x) dx = ∫ f ( x) dx = 2 ∫ f ( x ) dx .
−∞ −∞ −∞ 0

7) If p and q be two polynomials with real coefficients where q ( x) ≠ 0 for

all real x such that deg (q) ≥ deg ( p ) + 2 and z1 , z 2 , K, z k be the set of
p ( z)
all the poles in the upper half plane of the function f ( z ) = . then
q ( z)
56
 ∞ k Applications of Residue
P.V . ∫ f ( x ) dx = 2π i ∑ Res [ z j , f ( z )] .
−∞ j =1

8) If p and q be two polynimials with real coefficients where q ( x) ≠ 0 for

all real x such that deg (q) ≥ 1 + deg ( p ) and if z1, z 2 ,K, zk be all the poles
p ( z ) iα z
of f ( z ) = e ,α ≠ 0 in the upper half plane, then
q ( z)
∞ k
p( x)
P.V . ∫ cos (a x) dx = −2π ∑ Im[ Res [ z j , f ( z )]]
−∞
q( x ) j =1
∞ k
p( x)
P.V . ∫ sin (a x ) dx = 2π ∑ Re[ Res [ z j , f ( z )]]
−∞
q( x) j =1

9) Let p and q be two polynomials with real coefficients of degrees m and

n , respectively and n ≥ m + 2 . If q has simple zeros at the points
x1, x2 , .K, xk on the real axis and z1 , z 2 , K, z k are the poles of
p ( x)
f ( z) = in the upper half plane, then
q ( x)
∞ k I
p ( x)
P.V . ∫ dx = π i ∑ Res [ x j , f ( z )] + 2π i ∑ Res [ z j , f ( z )] .
−∞
q ( x) j =1 j =1

10) Let p and q be two polynomials with real coefficients of degrees m and
n , respectively and n ≥ m + 1 . If q has simple zeros at the points
x1, x2 , K , xk on the real axis and z1 , z 2 , K , z I are the poles of
e iαz p ( x)
f ( z) = , α ≥ 0 in the upper half plane, then
q ( x)
∞ k I
p ( x)
P.V . ∫ cos (ax)dx = −π ∑ Im( Res[ x j , f ( z )]) − 2π ∑ Im ( Res( z j , f ( z )])
−∞
q ( x) j =1 j =1

and
∞ k I
p ( x)
P.V . ∫ sin (ax)dx = π ∑ Re ( Res[ x j , f ( z )]) + 2π ∑ Re ( Res ( z j , f ( z )]).
−∞
q ( x) j =1 j =1

8.5 SOLUTIONS/ANSWERS

E1) Using the substitution z = eiθ (0 ≤ θ ≤ 2π ) , the integrand takes

the form
1 − 2i
f ( z) = −1
= 2
iz [a + b ( z + z ) / 2] (bz + 2az + b)
The singularities of f (z ) are
− a + (a 2 − b 2 ) − a − (a 2 − b 2 )
α= ,β = .
b b
You can see that
a + (a 2 − b 2 ) a
|β |= > > 1.
b b
57
Applications of It means the point z = β lies outside the unit circle C . Since αβ = 1
Analytic Functions
therefore | α | < 1 and the point z = α is a simple pole lying inside the
unit circle. We thus have
− 2i − 2i
Res [α , f ( z )] = lim ( z − α ) = .
z →α b ( z − α ) ( z − β ) b (α − β )
Thus
2π
dθ 2π
∫0 a + b cosθ (a > b > 0) = 2π i × Res[α , f ( z )] = (a 2 − b2 ) .

E2) Substituting z = eiθ , (0 ≤ θ ≤ 2π ) , we get

2π
sin 2 θ dθ −1 ( z 2 − 1) 2 dz
∫ a + b cosθ 2i C∫ z 2 (bz 2 + 2az + b)
=
0

−1 ( z 2 − 1)2 dz
2bi C∫ z 2 ( z − α ) ( z − β )
=

where,
− a + (a 2 − b 2 ) − a − (a 2 − b 2 )
α= ,β = .
b b
From E1), we have already seen that | α | < 1 lies inside the unit circle C .
α is a simple pole and 0 is a double pole. Now
 ( z 2 − 1) 2  (α 2 − 1) 2
Res α , 2  = 2
 z ( z − α ) ( z − β )  α (α − β )
2 a 2 − b2
=
b
and
 ( z 2 − 1) 2  d 2 ( z 2 − 1)2
Res 0, 2  = lim z × .
 z ( z − α ) ( z − β )  z → 0 dz z2 (z − α ) (z − β )
− 2a
=α + β =
b
Hence
2π sin 2 θ dθ − 1  − 2a 2 a 2 − b 2 
∫
0 a + b cosθ
= 2π i × ×
2bi  b
+
b 
 
2π
= 2  a − a 2 − b 2  .
b  

E3) i) Substituting z = eiθ , (0 ≤ θ ≤ 2π ) ,

z 2 + z −2
2π
cos 2θ dθ 2 dz
∫ =∫
13 − 12 cosθ C  z + z  iz
−1
0
13 − 12 

 2 
i 1 + z4
= ∫ 2 dz .
2 C z (3z − 2) (2 z − 3)
2
Observe that the integrand f (z ) has the simple poles at z = and
3
58
 3 2 Applications of Residue
z= and the double pole at z = 0 . Only z = 0 and z = lie
2 3
inside the unit circle C .
d 2 1 + z4 13
Res [0, f ( z )] = lim z 2 = .
z → 0 dz z (3z − 2) (2 z − 3) 36
and
2   2 1 + z4 − 97
Res  , f ( z ) = lim2  z −  2 = .
3  z→ 3  3  z (3z − 2) (2 z − 3) 180
Hence
2π
cos 2θ dθ  i   13 97  8π
∫ 13 − 12 cosθ
= (2π i ) ×    − =
 2   36 180  45
.
0

ii) Substituting z = eiθ , (0 ≤ θ ≤ 2π ) ,

2π
dθ dz
∫ =∫
(1 + 2a cosθ + a ) C iz (1 + a ( z + z −1 + a 2 ))2
2
0

− iz dz
=∫ 2 2 2
.
C
( az + ( a + 1) z + a )
−1
Singularities of f (z ) are α = − a and β = , both are of
a
multiplicity 2.
1
Also | β | = > 1 , so z = β lies outside the unit circle C . As
a
| α | = a < 1 , and α is a double pole, we have
d − iz
Res [− a, f ( z )] = lim ( z + a) 2
z → −a dz ( z + a )2 ( z + 1a )2
d − iz
= lim
z → −a dz z + 1a

− i (a 2 + 1)
= .
(1 − a 2 )3
Hence
2π
dθ − i (a 2 + 1) (a 2 + 1)
∫ (1 + 2a cosθ + a 2 ) 2
= 2π i ×
(1 − a 2 )3
= 2π
(1 − a 2 )3
.
0

E4) Proceed as in Example 3

1
E5) i) Consider f ( z ) = which has the double poles at z = ± i .
( z 2 + 1) 2
The singularity z = i of f in the upper half plane lies in the
interior of the semicircular region bounded by the segment [−r , r ]
of the real axis and the upper half Cr of the circle | z | = r from Fig. 9
− r to r (see Fig. 9). We have
d 1  −i
Res [i, f ( z )] = lim  ( z − i ) 2 2 = .
z→i
 dz ( z + 1) 2  4
59
Applications of This implies
Analytic Functions r
−i π
∫ f ( x) dx + ∫ f ( z ) dz = 2π i × 4
= .
2
−r Cr
Thus
r
π
∫ f ( x) dz = 2 − ∫ f ( z) dz
−r Cr
which is valid for all values of r > 1 . Moreover,
| ( z 2 + 1)2 | ≥ (|| z |2 −1 |2 ) = (r 2 − 1)2
for z ∈ Cr and
dz πr
∫ ( z 2 + 1)2 ≤
(r − 1) 2
2
Cr

which tends to 0 as r → ∞ . Hence

∞ r
π
P.V . ∫ f ( x) dx = lim
r→∞ ∫
f ( x ) dx =
2
−∞ −r

which implies that

∞
dx π
∫ ( x 2 + 1)2 = 4 .
0

ii) We have p ( z ) = z 2 and q ( z ) = z 4 + 1 . Clearly,

deg q ( z ) = 4 = deg p ( z ) + 2 . The zeros of q (z ) are
1
z= (±1 ± i) . Also q ( x) ≠ 0 for any real x . The singularities of
2
p ( z) 1
f ( z) = are all simple poles. Of these poles, α = (1 + i )
q ( z) 2
1
and β = (−1 + i) lies in the upper half plane. Observe that,
2
p (α ) ≠ 0, p ( β ) ≠ 0 . Also, q′ (α ) ≠ 0, q′ ( β ) ≠ 0 . As a result,

 p ( z )  p (α ) 1 − i
Res α , = =
 q ( z )  q′ (α ) 4 2
and
 p ( z )  p (β ) − 1 − i
Res  β , = =
 q ( z )  q′ ( β ) 4 2
Hence
∞
x 2 dx 1 − i −1 − i  π
P.V . ∫ 4
( x + 1)
= 2π i  +
4 2 4 2 
= 2.
−∞

As integrand is an even function and P.V. exists, therefore

∞ ∞
x 2 dx 1 x 2 dx π
∫ ( x 4 + 1) 2 ∫ ( x4 + 1) = 2 2 .
=
0 −∞

60
 1 Applications of Residue
E6) For f ( z ) = 2
, z = ± i are the poles of order n of which z = i lies
( z + 1)n
in the upper half plane.
1 d n −1  1 
Res [i, f ( z )] = lim  .
(n − 1)! z →i dz n −1  ( z + i )n 
d n −1  1 
  − 2 n +1
dz n −1  ( z + i) n  = (− n) (−n − 1) ...(−2n − 2) ( z + i )
 
(2n − 2)!
= (−1)n −1 ( z + i) − 2n +1
(n − 1)!
which implies that
(2n − 2)!
Res [i, f ( z )] = −2− 2 n +1 i .
[(n − 1)!]2
Therefore,
r
π (2n − 2)!
∫− r f ( x) dx + C∫ f ( z ) dz = 2π i × Res[i, f ( z )] = 22 n − 2 [(n − 1)!]2 .
r

It can be easily seen that

dz πr
∫ ( z 2 + 1) n ≤ 2
(r − 1) n
Cr

which tends to 0 as r → ∞ . Also,

∞ r
(2n − 2)! π
P.V . ∫ f ( x) dx = lim
2 r→∞ ∫
f ( x) dx =
[(n − 1)!]2
. 2n − 2
−∞ −r
As integrand is even and the Cauchy principal value exists hence,
∞
π (2n − 2)!
∫ f ( x) dx =
2 2n − 2
[(n − 1)!]2
−∞
or
∞
π (2n − 2)!
∫ f ( x) dx = 22 n −1 [(n − 1)!]2
.
0

E7) i) We have p ( z ) = z 2 and q ( z ) = ( z 2 + a 2 ) 2 . Clearly,

deg q ( z ) = 4 deg p ( z ) + 2 . The zeros of q (z ) are z ± ai and
q′ ( x) ≠ 0 for any real x . Also z = ± ai are the double poles of
p ( z)
f ( z) = . Of these poles, z1 = ai lies in the upper half plane.
q ( z)
p ( z) d 2 z2 i
Res = lim ( z − ai) 2 2
=− .
z = z1 q ( z ) z → ai dz (z + a ) 4a
Using Theorem 1
∞
x 2 dx i π
P.V . ∫ 2 2 2
(x + a )
= 2π i × − =
4a 2a
.
−∞
∞ 2
x π
or ∫ 2 2 2
(x + a )
dx =
2a
.
−∞

ii) Proceed as in i) above.

61
Applications of ∞
dx
Analytic Functions E8) Proceed as in Example 6 and obtain P.V . ∫ 2
x + 2x + 2
=π .
−∞

E9) We have p ( z ) = 1 and q ( z ) = ( z 2 + a 2 ) ( z 2 + b 2 ) . Clearly,

deg q ( z ) = 4 > deg p ( z ) + 2 = 2 . The zeros of q (z ) are z = ± ai and
z = ± bi (q ( x ) ≠ 0 for any real x) . Poles, z1 = ai (simple pole) and
z2 = bi (simple pole) lie in the upper half plane. Note that
p ( z1 ) ≠ 0, p ( z 2 ) ≠ 0 and q′ (ai ) ≠ 0, q′ (bi ) ≠ 0 .
1 i
Res [ai, f (z)] = lim ( z − ai ) f ( z ) = lim 2 2
=
z → ai z → ai ( z + ai ) ( z + b ) 2a (a − b 2 )
2

and
i
Res [bi, f (z)] = − .
2b (a − b 2 )
2

Using Theorem 1
∞
dx  i i  π
∫ ( x 2 + a 2 ) ( x 2 + b 2 ) = 2π i  2a (a 2 − b2 ) − 2b(a 2 − b2 )  = ab(a + b) .
−∞

E10) It is sufficient to show that P.V. of the integral exists as the integrand is
eiz
an even function. Consider f ( z ) = 2 which is analytic
( z + 1) 2
everywhere on and above the real axis except at the point z = i . The
singularity z = i (is a double pole) lies in the interior of the semicircular
region whose boundary consists of the segment − r ≤ x ≤ r of the real
axis and the upper half circle Cr of the circle | z | = r , (r > 1) from z = r
to z = − r . Now
d eiz
Res [i, f ( z )] = lim ( z − i)2 2
z → i dz ( z + 1) 2

d  eiz  − i
= lim  2 2
= .
z →i dz  ( z + 1)  2e
Integrating f (z ) along the boundary, we get
r
π
∫
−r
f ( x)dx + ∫ f ( z )dz = 2π i × Res[i, f ( z )] = e
Cr

Thus,
r
π ei z
Re ∫ f ( x) dx = − Re ∫ dz .
−r
e Cr
( z 2 + 1) 2
Now | eiz | = e − y ≤ 1 as y ≥ 0 and

| z 2 + 1 | ≥ | z |2 −1 = r 2 − 1 .

Hence we obtain
ei z ei z πr
Re ∫ dz ≤ ∫ dz ≤ 2 → 0 as r → ∞
Cr
( z 2 + 1)2 Cr
2
( z + 1) 2
(r − 1) 2
Hence,
62
 ∞ Applications of Residue
cos x π
∫ 2
( x + 1) 2
dx = .
e
∞

( z + 1) eiz ( z + 1) eiz
E11) We write f ( z ) = = .
z 2 + 4 z + 5 ( z + 2 + i ) ( z + 2 − i)
The point z = −2 + i , lying in the upper-helf plane, is a simple pole of f ,
with residue
b = e −1 (1 + i ) (cos 2 − i sin 2) .
For r > 5 = | −2 + i | and Cr denoting the upper half of the positively
oriented circle | z | = r (see Fig. 10) the Cauchy residue theorem yields
r
( x + 1) e ix dx
∫ x2 + 4x + 5
= 2π ib − ∫ f ( z ) dz .
−r Cr

We shall show that ∫ f ( z ) dz → 0 as r → ∞ . When z is a point on Cr Fig. 10

Cr

( z + 1) e iz (r + 1)
| f ( z) | = 2
≤ | e iz |
z + 4 z + 5 (r − 5 ) 2
and | eiz | ≤ 1 for such a point z . We cannot conclude that integral of
π r (r + 1)
f (z ) along Cr → 0 as r → ∞ (because / 0 as r → ∞) .
→
(r − 5 )2
If we put z = reiθ where (0 ≤ θ ≤ π , r > 5 ) , we get for z on Cr (using
Jordan Inequality)
π

∫ f ( z) ≤ r ∫ | f (re iθ ) | dθ
Cr 0
π
r (r + 1)
∫e
− r sin θ
≤ dθ
(r − 5 ) 2 0
π /2
2r (r + 1) π (r + 1)
≤
(r − 5 ) 2 ∫ e − r sin θ dθ ≤
(r − 5 ) 2
0
and, we have

∫ f ( z ) dz → 0 as r → ∞ . Thus,
Cr
∞

∫ f ( x) dx = 2π ib .
−∞

Hence, by equating the real parts

∞
( x + 1) cos x dx π (sin 2 − cos 2)
P.V . ∫ 2
= Re [2π ib] = .
−∞
x + 4x + 5 e

e aiz
E12) i) Let f ( z ) = dz . The Singularities of this function are at
( z 2 + b 2 )2
the points z = ± bi . Hence, z = bi is a double pole lying in the
63
Applications of upper-half plane. Residue at this point is given by
Analytic Functions
d e aiz −i
B = lim ( z − bi) 2 2 2
= 3 (1 + ab)e − ab
z →bi dz (z + b ) 4b
Using Theorem 2
∞
cos ax π
P.V . ∫ 2 2 2
dx = −2π Im B = 3 (1 + ab) e − ab
−∞
(x + b ) 2b
Since the integrand is an even function and P.V. exists therefore
∞
cos a x π
∫ ( x 2 + b2 ) 2 dx = 4b3 (1 + ab) e
− ab
.
0

∞
x 3 sin x π  9 1
ii) ∫ 2 2
( x + 1) ( x + 9)
dx =  − .
16  e 3 e 
0

E13) i) Consider the following contour γ (see Fig. 11) and the function
einz
f ( z) = 4 .
z +1
ei n z | ei n z | πr
We have ∫ 4
z +1
dz ≤ ∫ r 4
− 1
dz ≤ 4
r − 1
→ 0 as r → ∞
Cr C r
inz − Im nz
(note that for z ∈ Cr , | e | = e ≤ 1).
Considering the real part and using the fact that for these roots
1
= − z1 , we have
Fig. 11 z13
∞
cos nx  ei n z 
 
∫ x4 + 1
dx = lim
r→∞
dx = lim
r→∞
Re
 ∫ z4 + 1 
dz
−∞ γ 
  einz   einz  
= Re  2π i  Res 4 + Res
  z = e i π z + 1   3 iπ z 4 + 1  
  4   z =e 4 
  einz   einz  
= −2π Im   4  iπ +  4 + 1 3iπ 
  z + 1 4  z  z =e 4 
 z =e 
n
−
πe 2   n   n 
= cos   + sin   .
2   2  2 

eiz
ii) Consider the function f ( z ) = . The singularities of
( z + a) 2 + b 2
f (z ) are z = − a ± bi . Only z = − a + bi lies in the upper-half plane
and is a simple pole. Compute the residue at this point and use
Theorem 2 to get the answer
∞
cos a x cos a
P.V . ∫ 2
( x + a) + b 2
dx = − b .
be
−∞

64
 p ( z) Applications of Residue
E14) Let us consider f ( z ) = . Hence we have p ( z ) = 1 and
q ( z)
q ( z ) = ( z − i )2 ( z − 1) .
Note that deg q = 3 ≥ deg p + 2 as well as f (z ) has simple pole namely,
x = 1 lying on the real axis and a double pole z = i lying in the upper
half plane. Now we compute the residue of f (z ) at these poles.
1 i
Res [1, f ( z )] = lim ( z − 1) 2
= .
z →1 ( z − i ) ( z − 1) 2
d 1 i
Res [i f ( z )] = lim ( z − i )2 2
=− .
z →1 dz ( z − i ) ( z − 1) 2
Using Theorem 3, we get
∞
dx  i i π
P.V . ∫ 2
( x − i) ( x − 1)
= 2π i  −  + π i   = .
 2 2 2
−∞

e aiz − ebiz
E15) We choose a function f ( z ) = whose real part is
z2
cos ax − cos bx
. The singularity of f (z ) is z = 0 . The Laurent series
x2
expansion f (z ) about z = 0 is given by
e aiz − ebiz 1  (a 2 − b 2 ) 2 
f ( z) = = ( a − b ) iz − z + ⋅ ⋅ ⋅
z2 z 2  2 
= ( a − b) i
So that the pole at z = 0 is simple, with residue (a − b)i . Since the pole
z = 0 lies on the real axis, we make an indentation at 0 as shown in
Fig.12. Our contour Γ consists of semicircle Cr , (r > | a − b |) from r to
Fig. 12
− r , segment [−r , − ε ] (0 < ε < r ) , inner semicircle Cε from − ε to ε
and segment [ε , r ] . f (z ) is analytic inside and on Γ so, by the Cauchy-
Goursat theorem
−ε r

∫ f ( x)dx − ∫ f ( z ) dz + ∫ f ( x ) dx + ∫ f ( z ) dz = 0 .
−r Cε ε Cr
The first and the third integrals combine to give
r r r
e − ax − e − aix e aix − e bix cos ax − cos bx
∫ x2 dx + ∫ x 2 dx = 2 ∫ x 2
.
ε ε ε
As we have seen that z = 0 is a simple pole then applying indentation
lemma, we get
lim ∫ f ( z ) dx = i (π − 0) Res [0, f ( z )] = π (b − a ) .
ε →0
Cε

Finally, we have to estimate the integral ∫ f ( z ) dz . We see that on Cr

Cr

e aiz − ebiz e − ay + e −by 2

| f ( z) | = 2
≤ 2
≤ 2 , ( y ≥ 0)
z r r
2π
⇒ ∫ f ( z ) dz ≤
r
Cr
65
Applications of
Analytic Functions
⇒ ∫ f ( z ) dz → 0 as r → ∞
Cr

Letting r → ∞, ε → 0 ,
∞ ∞
cos ax − cos bx cos ax − cos bx π
∫2 x 2
dx = π (b − a ) ⇒ ∫
x 2
dx = (b − a) .
2
0 0

p ( z)
E16) Let us consider f ( z ) = eiz . Here we have p ( z ) = 1 and
q ( z)
q ( z ) = a 2 − z 2 . Note that deg q = 3 ≥ deg p + 1 . f (z ) has the simple
poles, namely z = ± a , lying on the real axis. Compute the residue of
f (z ) at these poles
eiz eia
Res [a, f ( z )] = lim ( z − a) = .
z→a (a 2 − z 2 ) 2a
eiz e − ia
Res [− a, f ( z )] = lim ( z + a) = .
z →− a ( a 2 − z 2 ) 2a
∞
cos x  − sin a sin (− a)  π sin a
∴ P.V . ∫ 2
(a − x )2
dx = −π 
 2a
+
2a 
=
a
.
−∞

–x–

66