SEOUL NATIONAL UNIVERSITY
School of Mechanical & Aerospace Engineering
400.002
Eng Math II
31 Residue Integration Method
31.1 Residue Theorem Theorem 1 [Residue theorem]
Let f(z) be analytic inside a simple closed path C and on C, except for finitely many singular points z1, z2,· · · , zkinsideC. Then the integral off(z) taken counterclockwise around C equals 2πi times the sum of the residues of f(z) at z1,· · · , zk.
6C uz1
uz2 · · · uzk '
&
$
H %
Cf(z)dz= 2πiPk
j=1Resz=zjf(z)
Proof. I
C
f(z)dz+ I
C1
f(z)dz+ I
C2
f(z)dz+· · ·+ I
Ck
f(z)dz = 0 C : ccw, C1, C2,· · ·, Ck: cw
I
C
f(z)dz= I
C1
f(z)dz+ I
C2
f(z)dz+· · ·+ I
Ck
f(z)dz C, C1, C2,· · · , Ck: ccw
I
Cj
f(z)dz = 2πiResz=zjf(z), j= 1,· · ·, k Example 1. Integration by the residue theorem.
I
C
4−3z
z2−zdz (C : ccw) Solution.
Resz=0 4−3z z(z−1) =
·4−3z z−1
¸
z=0
=−4, Resz=1 4−3z z(z−1) =
·4−3z z
¸
z=1
= 1 (a) 0 & 1 are inside C : 2πi(−4 + 1) =−6πi
(b) 0 is inside, 1 is outside : 2πi(−4) =−8πi (c) 1 is inside, 0 is outside : 2πi(1) = 2πi (d) 0 & 1 are outside : 2πi(0) = 0
Example 2. Poles and essential singularities.
C : 9x2+y2 = 9 (ccw) I
c
µ zeπz
z4−16 +zeπ/z
¶ dz Solution.
±2i insideC ; ±2 outsideC Resz=2i zeπz
z4−16 =
·zeπz 4z3
¸
z=2i
=−1
16 ; Resz=−2i zeπz z4−16 =
·zeπz 4z3
¸
z=−2i
=− 1 16 z·eπ/z =z
µ 1 +π
z + π2
2!z2 + π3 3!z3 +· · ·
¶
=z+π+π2 2 ·1
z +· · ·
∴ 2πi µ
−1 16 − 1
16 +π2 2
¶
=π µ
π2−1 4
¶
i= 30.221i . 31.2 Integrals of Rational Function of cosθ and sinθ
I = Z 2π
0
F(cosθ,sinθ)dθ =? F : real rational ftn Seteiθ =z, thendz/dθ=ieiθ ⇒ dθ=dz/iz, and
cosθ= 1
2(eiθ+e−iθ) = 1 2(z+1
z) sinθ= 1
2i(eiθ−e−iθ) = 1 2i(z−1
z)
F Ãf(z)
∴ I = I
C
f(z)dz
iz C:|z|= 1. Example 3.
Show that Z 2π
0
√ dθ
2−cosθ = 2π Solution.
cosθ= 1
2(z+ 1/z), dθ=dz/iz I
c
dz/iz
√2−12(z+ 1/z) = I
c
dz
−2i(z2−2√
2z+ 1) =−2 i
I
c
dz (z−√
2−1)(z−√ 2 + 1) z1 =√
2 + 1 : ( outside of C), z2=√
2−1 (inside of C)
Resz=z2 1
(z−√
2−1)(z−√
2 + 1) =
· 1
z−√ 2−1
¸
z=√ 2−1
=−1 2
∴ Z 2π
0
√ dθ
2−cosθ = 2πi(−2/i)(−1/2) = 2π .
31.3 Improper Integrals of Rational Functions.
When the interval of integration is not finite:
Z ∞
−∞
f(x)dx= lim
a→−∞
Z 0
z
f(x)dx+ lim
b→∞
Z b
0
f(x)dx If both limits exist, Z ∞
−∞
f(x)dx= lim
R→∞
Z R
−R
f(x)dx
If f(x) is a real rational function whose denominator is different from zero for all realx and is of degree at least two units higherthan the degree of the numerator, then the both limits exist.
Consider I
C
f(z)dz around a pathC .
Choose R large enough so that C encloses all the upper-half plane (UHP) poles (finitely many), then I
C
f(z)dz = Z
S
f(z)dz+ Z R
−R
f(x)dx= 2πiX
Resf(z)
=⇒ Z R
−R
f(x)dx= 2πiX
Resf(z)− Z
S
f(z)dz . (1)
Setz=Reiθ, then onS,R=const, 0≤θ≤π. By assumption,
|f(z)|< k
|z|2 (|z|=R > R0) for somek and R0. ¯
¯¯
¯ Z
S
f(z)dz
¯¯
¯¯< k
R2πR= kπ
R (R > R0) Take R−→ ∞ in (1). Then¯
¯R
Sf(z)dz¯
¯−→0.
∴ R∞
−∞f(x)dx= 2πiP
U HPResf(z) . (summation over all UHP poles)
−R R
-x 6y
S (semicircle) o
R
-
C (contour) uz1
uz2 · · ·
uzk
Example 4. An improper integral from 0 to∞ Z ∞
0
dx
1 +x4 = π 2√
2 Solution.
f(z) = 1/(1 +z4) ⇒z2 =±i Four simple poles z1=eπi/4, z2=e3πi/4, z3 =e−3πi/4, z4 =e−πi/4
z1 & z2 lie in the upper half-plane Resz=z1f(z) =
· 1 (z+z4)0
¸
z=z1
=
· 1 4z3
¸
z=z1
= 1
4e−3πi/4 =−1 4eπi/4 Resz=z2f(z) =
· 1 (z+z4)0
¸
z=z2
=
· 1 4z3
¸
z=z2
= 1
4e−9πi/4= 1 4e−πi/4 (∵ eπi =−1, e−2πi= 1)
Z ∞
−∞
dx
1 +x4 =−2πi
4 (eπi/4−e−πi/4) =−2πi
4 2isinπ 4 = π
√2
∴ Z ∞
0
dz 1 +x4 = 1
2 Z ∞
−∞
dz
1 +x4 = π 2√
2 . 31.4 Fourier Integrals
f: rational ftn satisfying the degree assumption as before Z ∞
−∞
f(x)eisxdx= 2πiX
Res[f(z)eisz] (s >0)
⇒
½ R∞
−∞f(x) cosxdx=−2πP
ImRes[f(z)eisz] R∞
−∞f(x) sinxdx= 2πP
ReRes[f(z)eisz] Since s >0 and S lies in the upper half-plane y≥0,
|eis(x+iy)|=|eisx| |e−sy|=e−sy ≤1 (s >0, y ≥0) AsR → ∞, the value of the integral overS →0.
Example 5.
Z ∞
−∞
cossx
k2+x2dx= π ke−ks,
Z ∞
−∞
sinsx
k2+x2dx= 0. (s >0, k >0) Solution. z=ik in the upper half-plane.
Resz=ik eisz k2+z2 =
·eisz 2z
¸
z=ik
= e−ks 2ik Z ∞
−∞
eisx
k2+x2dx= 2πie−ks 2ik = π
ke−ks Z ∞
−∞
cossx
k2+x2dx= π ke−ks,
Z ∞
−∞
sinsx
k2+x2dx= 0
31.5 Another Kind of Improper Integral.
A≤α≤B, lim
x→α|f(x)|=∞, Z B
A
f(x)dx=?
Z B
A
f(x)dx= lim
ε→0
Z α−ε
A
f(x)dx+ lim
η→0
Z B
α+η
f(x)dx Cauchy principal value of the integral:
pr.v.
Z B
A
f(x)dx= lim
ε→0
·Z α−ε
A
f(x)dx+ Z B
α+ε
f(x)dx
¸
For example,
pr.v.
Z 1
−1
dx x3 = lim
ε→0
·Z −ε
−1
dx x3 +
Z 1
ε
dx x3
¸
= 0 Theorem 1 [Simple poles on the real axis]
If f(z) has a simple pole at z=aon the real axis, then
r→0lim Z
c2
f(z)dz=πiResz=af(z) Proof.
f(z) = b1
z−a+g(z), b1 = Resz=af(z) g(z) is analytic on the semicircle of integration.
C2 :z=a+reiθ 0≤θ≤π, Z
C2
f(z)dz = Z π
0
b1
reiθireiθdθ+ Z
C2
g(z)dz Z π
0
b1
reiθireiθdθ= Z π
0
ib1dθ=b1πi Z
C2
g(z)dz ≤M πr asr→0, M πr→0.
Q. principal value of the integral of a rational functionf(z) from−∞ to∞? AsR−→ ∞,R
Sf(z)−→0.
Asr−→0,R
C2f(z)−→ −πiResz=af(z).
−R R
-x 6y
¾»
u a -C2
S (semicircle) o
R
- -
C (contour) uz1
uz2 · · ·
uzk
∴ pr.v.
Z ∞
−∞
f(x)dx= 2πi X
U HP
Resf(z) +πiX
real
Resf(z)
(the first sum over all the UHP poles, the second sum over all poles on the real axis) Example 6. Poles on the real axis
pr.v Z ∞
−∞
dx
(x2−3x+ 2)(x2+ 1) =?
Solution. Simple polesz= 1, z= 2 (real),z=i(in the UHP) Resz=1f(z) =
· 1
(z−2)(z2+ 1)
¸
z=1
=−1 2 Resz=2f(z) =
· 1
(z−1)(z2+ 1)
¸
z=2
= 1 5 Resz=if(z) =
· 1
(z2−3z+ 2)(z+i)
¸
z=i
= 1
6 + 2i = 3−i 20
∴ pr.v.
Z ∞
−∞
dx
(x2−3x+ 2)(x2) = 2πi µ3−i
20
¶ +πi
µ
−1 2+ 1
5
¶
= π 10 .