Scribd
Upload
40 views

Solu 5

This document provides solutions to problems from Math 3X03 assignment 5. It discusses the following key points: 1) A series converges conditionally if the terms do not converge absolutely but the even and odd partial sums converge. 2) A sequence converges to 0 if for any epsilon greater than 0, there exists an N such that the terms are within epsilon of 0 when n is greater than N. 3) A power series represents an analytic function on both the interior and exterior of the unit circle if the terms converge uniformly on each region.

Uploaded by

nikolaimuthama
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
40 views

Solu 5

This document provides solutions to problems from Math 3X03 assignment 5. It discusses the following key points: 1) A series converges conditionally if the terms do not converge absolutely but the even and odd partial sums converge. 2) A sequence converges to 0 if for any epsilon greater than 0, there exists an N such that the terms are within epsilon of 0 when n is greater than N. 3) A power series represents an analytic function on both the interior and exterior of the unit circle if the terms converge uniformly on each region.

Uploaded by

nikolaimuthama
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

SOLUTIONS FOR ASSIGNMENT 5 (MATH3X03)

3.1.7. (4pts)
P
remark that
it is a well-known fact in classical analysis that for a series
an , if
n=0 P
PWe
P

a
and
a
converge
to
A
and
A
respectively,
then
the
original
an
even
odd
k=0 2k
k=0 2k+1
converges to Aeven + Aodd . This fact applies to complex series as well. (And you should think
about the proof.)
P
2k
(a)PSince |in / log n| > 1/n, the series is not absolutely convergent. But both
k=1 i / log(2k)

2k+1
and k=1 i
/ log(2k + 1) converge by the alternating series test. Thus the series in question converges.
(b) The series does not absolutely converge but conditionally converges for the same
reason.
3.1.10.
P (3pts)
If
ak converges, then for any  > 0, there exists some natural number N such that
whenever m, n > N , we have
n

X


ak < .



k=m

In particular, if m = n, then |am | <  (for m > N ). This is exactly the definition of
ak P
0, k .
If
gk (z) converges uniformly, then for any  > 0, there exists some natural number N
such that whenever m, n > N , we have


n
X



gk (z) < ,



k=m

for all z (in the suitable domain). In particular, if m = n, then |gk (z)| <  for all z. This is
exactly the definition of gk 0 uniformly when k .
3.1.14. (4pts)
zn
First we note that for each n, the function 1+z
2n is analytic both on the interior and the
exterior of the unit circle, since all the roots of 1 + z 2n are on the unit circle.
If |z| for some such that 0 < < 1. Then


zn
n
n
n

.
1 + z 2n |1 + z 2n |
1 |z|2n
1 2n
Date: 16 Mar, 2012.
Totally 25 points.
1

SOLUTIONS FOR ASSIGNMENT 5 (MATH3X03)

2 of 3

P n
By the root test, one knows that
/(1 2n ) converges if 0 < < 1. Thus, by
the Weierstrass M -Test (Theorem 3.1.7), the series in question is uniformly and absolutely
convergent on the disc of radius centered at 0, which in turn, by the Analytic Convergence
Theorem (Theorem 3.1.18), implies that it converges and represents an analytic function on
the interior of the unit circle, since every closed disc contained in the interior of the unit
circle is contained in the set {z C | |z| < } for some 0 < < 1.
If |z| for some > 1. Then
zn
wn
z n /z 2n
=
=
,
1 + z 2n
1/z 2n + 1
1 + w2n
where w = 1/z and |w| 1/. Applying the above argument to w, we conclude1 that the
series in question also converges and represents an analytic function on the exterior of the
unit circle.
3.2.2. (4pts)
2
(a) R = limn n
/(n + 1)2 = 1.
2n
(b) R = limn 4n = 2.
(c) R = limn (n!)/((n
+ 1)!) = 0.

n
n
(d) R = limn 1 + 2 = 2.
3.2.6. (3pts)
Since the function 1/(1 + ez ) is analytic as long as 1 + ez 6= 0, i.e. when z 6= (2k + 1)i,
k Z, the radius of convergence is R = . (The nearest singularities to 0 is .)
To find the first four terms of the Taylor series around z0 = 0, we do the following (there
are many ways to do it, as usual). Suppose the Taylor series of it is
1
= a0 + a1 z + a2 z 2 + a3 z 3 + .
1 + ez
By expanding ez into its Taylor series, we have
1 = (a0 + a1 z + a2 z 2 + a3 z 3 + )(1 + 1 + z + z 2 /2 + z 3 /6 + ).
By comparing the coefficients on both sides, we get
2a0 = 1, a0 + 2a1 = 0, a0 /2 + a1 + 2a2 = 0, a0 /6 + a1 /2 + a2 + 2a3 = 0.
Therefore,
1
1 1
1
= z + z3 + .
z
1+e
2 4
48
1Again,

every closed disc contained in the exterior of the unit circle is contained in {w C | 0 < |w| 1/}
for some > 1; and then apply the Analytic Convergence Theorem. One thing you need to be careful about
is how to handle 0 and . But this argument just said takes care of it.

SOLUTIONS FOR ASSIGNMENT 5 (MATH3X03)

3.2.8. (4pts)
P
P
2 2n1
n+1 (z )
(a) sin z 2 =
=
n=1 (1)
n=1
(2n1)!
P
2n z n
(b) e2z =
,
for
all
z.
n=0 n!

z 4n2
,
(2n1)!

3 of 3

for all z.

3.2.18. (3pts)
Just follow the hint.
Since cos(0) 6= 0, tan(z) is analytic at z = 0. So its Taylor expansion exists at z = 0 (in
fact, with a radius of convergence /2). Suppose the Taylor expansion is a0 +a1 z +a2 z 2 + ,
then
sin(z) = cos(z)(a0 + a1 z + a2 z 2 + ).
That is


z3 z5
z2 z4
z
+
= 1
+
+ (a0 + a1 z + a2 z 2 + a3 z 3 ).
3!
5!
2
4!
So we have
1
1
1
a0 = 0, a1 = 1, a2 = 0, a3 = a1 = .
2
3!
3
That is
1
tan(z) = z + z 3 + .
3
3.2.20. (3pts Extra credit)
P
Suppose
the
series
is
an (z z0 )n . On the circle of convergence, say with radius R,
P
P
n
|an (z z0 ) | =
|an |Rn does not depend on the choice of z. Thus the series either
absolutely converges
P everywhere or nowhere on its circle of convergence.
For example, P z n /n2 converges absolutely everywhere on its circle of convergence (of
radius 1), while
z n /n does not absolutely converge anywhere on its circle of convergence
(of radius 1).

You might also like