Mathematics 185 – Intro to Complex Analysis
Fall 2009 – M. Christ
Solutions to Selecta from Problem Set 12
4.4.4
Show that
π
2
sin
2
(
πz
)
=
∑
∞
n
=
∞
(
z

n
)

2
for all
z
∈
C
\
Z
.
Solution.
We deduce this from the formula
π
cot(
πz
) =
z

1
+
X
n
6
=0
±
(
z

n
)

1
+
n

1
²
,
(1)
where the sum is over all integers
n
6
= 0, positive and negative. We know the latter formula
to be valid for all
z
∈
C
\
Z
.
The connection between the two formulas is that
π
2
sin
2
(
πz
)
is the derivative of
π
cot(
πz
),
while
∑
∞
n
=
∞
(
z

n
)

2
is obtained, at least formally, by diﬀerentiating the righthand side
∑
∞
n
=
∞
(
z

n
)

1
term by term. Thus our task is to justify this term by term diﬀerentiation.
Let
F
N
(
z
) =
∑
N
n
=

N
(
z

n
)

1
. It was implicitly shown in the course of the proof of (1)
that
F
N
(
z
)
→
π
cot(
πz
)
uniformly on
D
,
as
N
→ ∞
. (The diﬀerence
F
N
(
z
)

π
cot(
πz
) is
expressed as an integral over
C
N
, and the bound established for this integral tends to zero
as
N
→ ∞
, uniformly for
z
∈
D
.)
Therefore the sequence of derivatives also converges uniformly on any disk centered at
w
of strictly smaller radius;
F
0
N
→
d
dz
(
π
cot(
πz
)) uniformly for all
z
∈
B
(
w,r/
2),
by one of our general theorems. Since
F
0
N
(
z
) =
∑
N
n
=

N
(
z

n
)

2
, this is the desired
conclusion.
4.4.8
Prove that
∑
∞
n
=0
(

1)
n

1
n

2
=
π
2
12
.
Solution.
Apply the residue theorem to
f
(
z
) =
π
z
2
sin(
πz
)
=
z

2
π
csc(
πz
), using the familiar
contours
C
N
. Calculate
Res(
f,n
) =
n

2
1
cos(
πn
)
= (

1)
n
n

2
for 0
6
=
n
∈
Z
;
using csc instead of cot is natural, because it gives the alternating
±
signs which we’re
aiming for.
Use the expansion sin(
πz
) =
πz
(1

π
2
z
2
/
6 +
O
(
z
4
)) to calculate
π
π
sin(
πz
)
=
π
πz
(1

π
2
z
2
/
6 +
O
(
z
4
))

1
=
z

1
(1 +
π
2
z
2
/
6 +
O
(
z
4
))
to obtain
z

2
π
π
sin(
πz
)
=
z

3
(1 +
π
2
z
2
/
6 +
O
(
z
4
)) =
z

3
+
π
2
6
z

1
+
O
(
z
)
1
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Documentand therefore
Res(
f,
0) =
π
2
6
.
By the Residue Theorem, for each integer
N
≥
1,
(2
πi
)

1
Z
C
N
f
(
z
)
dz
=
π
2
6
+

1
X
n
=

N
(

1)
n
n

2
+
N
X
n
=1
(

1)
n
n

2
.
If we can show that
R
C
N
f
(
z
)
dz
→
0 as
N
→ ∞
then we obtain the desired formula.
Now

Z
C
N
f
(
z
)
dz
 ≤
‘
(
C
N
) max
z
∈
C
N
h

z


2
π

sin(
πz
)


1
i
.
Since
‘
(
C
N
) = 8
N
+ 4 and

z


2
≤
N

2
for every
z
∈
C
N
, it suﬃces to prove that

sin(
πz
)


1
≤
1000 for every
z
∈
C
N
, for every integer
N
≥
1.
This is very similar to a corresponding bound established in class and in our text for
This is the end of the preview.
Sign up
to
access the rest of the document.
 Fall '07
 Lim
 Math, Power Series, FN, Rouch´’s Theorem

Click to edit the document details