BAR CODE
Tutorial Letter 103/3/2015
Complex Analysis
MAT3705
Semesters 1 & 2
Department of Mathematical Sciences
This tutorial letter contains the memorandum for the October 2012
examination paper.
MAT3705/103/3/2015
QUESTION 1 1.1
(z − 4
z + 4) = (z − 4)(z + 4) (z + 4)(z + 4)
= (z − 4)(z + 4)
|z + 4|2
= zz + 4z − 4z − 16
|z + 4|2
= |z|2− 16
|z + 4|2 + 4(z − z)
|z + 4|2
Since z − z is purely imaginary (take z = x + iy then z − z = 2iy) we have that
Re(z − 4
z + 4) = |z|2− 16
|z + 4|2 and if
|z|2− 16
|z + 4|2 = 0 then |z| = 4.
1.2
Suppose that z = x + iy where x, y ∈ R Then
z + 4 z − 4
< 1 ⇔ |z + 4| < |z − 4|
⇔ p
(x + 4)2+ y2 <p
(x − 4)2+ y2
⇔ (x + 4)2+ y2 < (x − 4)2+ y2
⇔ 8x < −8x
⇔ 16x < 0
⇔ x < 0
So the region is {x + iy | y ∈ R,x < 0 } which is the region left of the y=axis excluding the y=axis.
The set is open since it contains none of its boundary points.
MAT3705/103
QUESTION 2 2.1
sin z = i ⇐⇒ 1
2i(eiz − e−iz) = i ⇐⇒ eiz− e−iz = −2
⇔ (eiz)2+ 2(eiz) − 1 = 0 Substitute w = eiz to obtain the equation w2+ 2w − 1 = 0.
Then
w = −2 ± √
4 + 4 2
= −1 ± 2√ 2 2
= −1 ±√ 2 so that
eiz = −1 ±√ 2
⇔ iz = ln
−1 +√ 2
+ i2nπ or
iz = ln
−1 −√ 2
+ i(2n + 1)π
⇔ z = −i ln(√
2 − 1) + 2nπ or z = −i ln(√
2 + 1) + (2n + 1)π Then since
√ 1
2 − 1 =√ 2 + 1 we can write the answer as
z = i ln(√
2 + 1) + 2nπ or z = −i ln(√
2 + 1) + (2n + 1)π 2.2 For z = x + iy where x, y ∈ R we have
eiz = ei(x+iy) = e−y+ix = e−yeix = e−y(cos x + i sin x) Hence
Re eiz = e−ycos x and
Re eiz = e−ycos x = 0
⇔ cos x = 0 and y ∈ R
⇔ x = ±π
2 + 2nπ, y ∈ R Hence
z = (π
2 + nπ) + iy, n ∈ Z, y ∈ R
From f (z) = (2x3+ y3− 7y − x) + i(x3− 2y3+ 5y + 4x), we have
u = 2x3+ y3− 7y − x, v = x3− 2y3+ 5y + 4x Then
ux = 6x2− 1 vx= 3x2+ 4 uy = 3y2− 7 vy == −6y2+ 5
These first partial derivatives are all continuous everywhere, so g will therefore be differentiable wherever the Cauchy-Riemann equations are satisfied.
(See the section on sufficient conditions for differentiability) .Now ux = vy ⇔ 6x2− 1 = −6y2+ 5
⇔ x2+ y2 = 1 Similarly
uy = −vx⇔ 3y2− 7 = 3x2+ 4
⇔ x2+ y2 = 1
f (z) is differentiable on the circle x2+ y2 = 1 and if (x0, y0) satisfies x20+ y20 = 1 we get f0(x0+ iy0) = ux(x0+ iy0) + ivx(x0+ iy0)
= 6x20− 1 + i(3x20+ 4) which is also
= 6x20− 1 + i(3(1 − y20) + 4)
= 6x20− 1 + i(−3y02+ 7) or
f0(x0+ iy0) = uy(x0+ iy0) + ivy(x0+ iy0)
= 3y02− 7 + i(−6y02+ 5 ) which is also
= 3(1 − x20) − 7 + i(−6y20+ 5 )
= −3x20− 4 + i(−6y02+ 5 ) 3.2
We get
∂
∂x(eusin v) = euuxsin v + euvxcos v and
∂2
∂x2(eusin v) = eu(ux)2sin v + euuxxsin v + euuxvxcos v + euuxvxcos v + euvxxcos v − eu(vx)2sin v
MAT3705/103
Similarly
∂
∂y(eusin v) = euuysin v + euvycos v
∂2
∂y2(eusin v) = eu(uy)2sin v + euuyysin v + euuyvycos v + euuyvycos v + euvyycos v − eu(vy)2sin v Hence
∂2
∂x2(eusin v) + ∂2
∂y2(eusin v) = eusin v((ux)2− (vy)2) + eusin v(uxx+ uyy) + eucos v(vxx+ vyy) + 2eucos v(uxvx+ uyvy) + eusin v((uy)2− (vx)2)
Since u + iv is analytic, it follows that u and v are harmonic and we get uxx+ uyy = 0,
vxx+ vyy = 0,
ux = vy and uy = −vx Hence also
(ux)2 = (vy)2 (uy)2 = (vx)2 and
uxvx = −uyvy Hence
∂2
∂x2(eusin v) + ∂2
∂y2(eusin v) = 0 and so eusin v is harmonic.
QUESTION 4 4.1 We have that
|z|2 = zz = (√
2)2 = 2 Hence in the circle |z| =√
2 we get
z = 2 z hence
(z)2 = 4
z2 on the circle C : |z| =√ 2 Hence
Z
C
4 z2dz =
Z
C
z2dz
C
f (z)
(z − z0)n+1dz = 2πi
n! f(n)(z0) where C is the circle |z| = 3.
Hence z0 = −1 and n = 3.
Clearly z0 = −1 lies inside the circle |z| = 3.
Also
f (z) = e2z, f0(z) = 2e2z, f00(z) = 4e2z and f000(z) = 8e2z Hence f000(−1) = 8e−2 and so
Z
|z|=3
e2z
(z + 1)4dz = 2πi 3! (8
e2) = 8πi 3e2 4.3
We have that f (z) is analytic in the domain |z| < 1 and |f (0)| = |1 + i| =√ 2.
Since |f (z)| ≤√
2 = |f (0)| for |z| < 1 and 0 lies in the domain |z| < 1, it follows from the Maximum Modulus Principle that f is constant in the domain |z| < 1.
Hence f (z) = 1 + i for |z| < 1 and so f0(z) = 0 for |z| < 1 and f0(0) = 0.
4.4
Since |g(z)| > |f (z)| ≥ 0 it follows that |g(z)| > 0 and so g(z) 6= 0 for every z ∈ C.
Since f and g are analytic on C, f (z)g(z) is also analytic on C.
But
f (z) g(z)
< 1 on C so by Louisville’s Theorem f (z)g(z) is constant on C.
QUESTION 5 Partial fractions:
1
(z − 1)(z − 2) = A
z − 1 + B z − 2 i.e.
1 = A(z − 2) + B(z − 1) If z = 1 then A = −1; if z = 2 then B = 1.
Hence
1
(z − 1)(z − 2) = −1
z − 1 + 1 z − 2 Since 1 < |z| < 2 we get
1
z
< 1 and z
2
< 1.
Hence
−1
z − 1 = − 1
z(1 − 1z) = −1 z
∞
X
n=0
(1
z)n= −
∞
X
n=0
1
zn+1 valid for |z| > 1 Also
1
z − 2 = 1
−2(1 − z2) = −1 2
∞
X
n=0
(z 2)n
= −
∞
X
n=0
zn
2n+1valid for |z| < 2
MAT3705/103
Hence
1
(z − 1)(z − 2) = −
∞
X
n=0
1 zn+1 −
∞
X
n=0
zn
2n+1 valid for 1 < |z| < 2
= ... − 1 z4 − 1
z3 − 1 z2 − 1
z − 1 2− z
22 −z2 23 − ...
QUESTION 6
6.1 z2− 1
z2− 5iz − 4 = z2− 1 (z − i)(z − 4i) Hence we have isolated singularities at z = i and z = 4i.
Let p(z) = z2− 1 and q(z) = z2− 5iz − 4.
Since p(i) 6= 0, p(4i) 6= 0 , q0(i) = 2i − 5i 6= 0, q0(4i) = 8i − 5i 6= 0, both of these singularities are simple poles.
Hence
Resz=ip(z)
q(z) = p(i)
q0(i) = i2− 1
2i − 5i = −2
−3i = −2i
3 and
Resz=4ip(z)
q(z) = p(4i)
q0(4i) = (4i)2− 1
2(4i) − 5i = −17
3i = 17i 3 and 6.2.1
Both i and 4i lie inside the circle |z| = 6.
Hence
Z
|z|=6
z2− 1
z2− 5iz − 4dz = 2πi(17i 3 − 2i
3) = 2πi(15i
3 ) = −10π 6.2.2
Since |4i − i| = 3 > 1 and |i − i| = 0 < 1 only the singularity at z = i lies inside the circle |z − i| = 1.
Hence
Z
|z−i|=1
z2− 1
z2− 5iz − 4dz = 2πi(−2i
3) = 4π 3 QUESTION 7.
7.1 We use the parametric representation z = eiθ(0 ≤ θ ≤ 2π) to describe C . Furthermore dzdθ = ieiθ = iz i.e. dθ = dziz
and from cos θ = 12(eiθ+ e−iθ),we obtain cos θ = 12(z + z−1)
Z 2π 0
dθ
3 − 2 cos θ = Z
|z|=1
1
3 − 2 12(z + z−1) 1 izdz
= Z
|z|=1
dz iz(3 − z − 1z)
= Z
|z|=1
dz i(3z − z2− 1)
= Z
|z|=1
idz z2− 3z + 1
z = 3 ± 9 − 4
2 = 3 ± 5
2 Note that 3+
√ 5
2 > 1 so this singularity lies outside |z| = 1) Then
Res 3−√ 5 2
i
z2− 3z + 1 = = i 2(3−
√ 5
2 ) − 3 = − i
√5 Therefore by the residue theorem we get
Z 2π 0
dθ
3 − 2 cos θ = Z
|z|=1
idz z2− 3z + 1
= 2πi (− i
√5)
= 2π
√5
7.2 Question: Show that f (z) − z4 has exactly four solutions inside the unit circle under the condition that |f (z)| < 1 for |z| = 1.
Let
g(z) = −z4
Then g(z) is analytic inside and on |z| = 1 and has 4 zeros (counting multiplicities) inside |z| = 1.
Also
|g(z)| = −z4
= 1 > |f (z)| on |z| = 1.
Thus by Rouche’s theorem g(z) and f (z) + g(z) = f (z) − z4 have the same number of zeros inside
|z| = 1
i.e. f (z) − z4 has exactly four solutions inside the unit circle.