Math 310 Problem Set 3

Jiyang Xie 1

Problem 1 (L p.46)
P
P
n z (2n)
(n−1) z (2n−1)
and cos(z) = ∞
(b) We have that sin(z) = ∞
n=0 (−1) (2n)! . Finding the
n=1 (−1)
(2n−1)!
small order terms of the product, we get
(1 −

z2
z3
2z 3
)(z − ) = z −
+ O(z 5 )
2
6
3

Thus, the order ≤ 3 terms of sin(z) cos(z) = z −

2z 3
.
3

(h) We will multiply the series expansions for ez and
all partial sums of sin(z) ≤ 1.

1
.
sin(z)

Note that the absolute value of

1
1
1
1
=
=
5
3
2
z
z
z
sin(z)
z 1 − ( 3! − z5!4 + ...)
z − 3! + 5! − ...
1
z2 z4
z2 z4
= [1 + ( −
+ ...) + ( −
+ ...)2 + ...]
z
3!
5!
3!
5!
1 x 7x2
= + +
+ ...
x 6 360
Thus
ez

x2 x3 x 4
1 x 7x2
1
= (1 + x +
+
+ ...)( + +
+ ...)
sin(z)
2
6
24
x 6 360
2x x2 13x3
1
+
+
+ ...
=( +1+
x
3
3
90

.........
Problem 4 (L p.58-60)
(a) By Hadamard’s we have that

1
R

= lim supn→∞ (nn )1/n = n. So, by convention, R = 0.

.........
Problem 5 (L p.58-60)
(a) Again by Hadamard’s, we have that
1
1
= lim sup(nan )1/n = lim sup n1/n lim sup(an )1/n = (1)( ) =⇒ R = r
R
r
n→∞
n→∞
n→∞
as desired.
(b) Again by Hadamard’s, we have that
1
= lim sup(n2 an )1/n = lim sup n2/n lim sup(an )1/n
R
n→∞
n→∞
n→∞
Note that
2 log(n)
=0
n
n→∞
=⇒ lim sup n2/n = 1

lim sup log(n2/n ) = lim sup
n→∞

n→∞

1
=⇒ lim sup n2/n lim sup(an )1/n = (1)( )
r
n→∞
n→∞
=⇒ R = r
September 25th, 2014

To find its radius of convergence. we have that 1 = lim sup(nd an )1/n = lim sup nd/n lim sup(an )1/n R n→∞ n→∞ n→∞ Note that d log(n) =0 lim sup log(nd/n ) = lim sup n n→∞ n→∞ =⇒ lim sup nd/n = 1 n→∞ 1 =⇒ lim sup nd/n lim sup(an )1/n = (1)( ) r n→∞ n→∞ =⇒ R = r as desired.. P n | n2 |. Thus s = 1 r lim sup |(an bn )1/n | = lim sup |(an )1/n ||(bn )1/n | n→∞ n→∞ ≤ lim sup |(an )1/n | lim sup |(bn )1/n | n→∞ n→∞ 1 = rs Thus we have that the radius of convergence must be greater than rs... So. . 1 P we1 note that along the boundary of the unit disc of convergence..6 in Stein. where |z| = 1... this is simply f 0 (x). Thus... by | zn2 | = Theorem 2. We claim that f (z) = ... an and bn can cancel each other to give an infinite radius of convergence... which is a p-series with p = 2 and thus converges absolutely. we get that f (z) is continuous and holomorphic along the entire unit disc.6 we know that it shares the same radius of convergence as f (x). 2014 1/n and lim supn→∞ bn = ..58-60)P n P (a) We have that (an +bn )z = an z n + bn z n . we refer to Hadamard’s to get lim supn→∞ n12 n = 1 = 1... If r = s. then the radius of convergence must only be at least r . Next.... We are given lim supn→∞ an 1 .. as desired.. . Also from Theorem 2. as desired. Problem 6 (L p.Math 310 Problem Set 3 Jiyang Xie 2 as desired. September 25th. (d) Note that by Theorem 2.. Problem 10 P (L p. Thus it is clear that the left side converges only if both of the terms of the right side converge...for example.. the radius of convergence is equal to the minimum of r and s if r 6= s.6 in Stein’s. (c) Again by Hadamard’s.58-60) P zn n2 has radius of convergence 1 and is continuous on the closed unit 1 disc. 1/n (b) We again refer to Hadamard’s.

the original function)..Math 310 Problem Set 3 Jiyang Xie 3 Problem 20 (S p. Problem 23 (S p. Since this equation has m−1+d .. Problem 26 (S p. We first will show that given any n > 0. . we claim that limx→0+ f n (x) = 0 as well. This yields q(x) + [ −p(x) q(x)2 p0 (x) −2 −1/x2 ] e . q(x) x3 which satisfies the form that we claimed..30-31) 2 Since both 0 and e−1/x are infinitely differentiable. F2 and F1 can differ only by a constant in Ω..   m−1+d d (d + m − 1)(d + m − 2). w2 ∈ Ω. we have R that F1 (w2 )−F1 (w1 ) = γ f (x)dx = F2 (w2 )−F1 (w1 ) =⇒ F1 (w2 )−F2 (w2 ) = F1 (w1 )−F1 (w1 ) for all w1 ...30-31)  P m 1 1 i m (a) Note that (1−z) = ( ∞ It is easy to collect like power terms = 1−z m i=0 z ) . by viewing for non-negative solutions to the Diophantine P this as a generating function d equation m a = d for the term x ...2 from Stein’s... we focus only on x = 0.. .. ..30-31) Suppose f has two primitives F1 and F2 such that F1 6= F2 .. since limx→0−1 f n (x) clearly equals 0. which is a contradiction. 2014 ...(d + 1) lim z = lim d→∞ d→∞ m−1 (m − 1)! m−1 d = (m − 1)! as desired. By Theorem 3. it would have to be identically 0.. we can find the nth derivative.... where each ai corresponds the xai term being sei=1 i lected from the ith entry in the product.. Assume it holds for the k derivative. we get that m−1 P∞ m−1+d d 1 = d=0 m−1 z . We can conclude that f (x) has no power series centered at 0 because since f n (0) = 0 for all n ≥ 1.. 2 we take the derivative with respect to x of a function of the form p(x) e−1/x . For the k + 1th derivative. we have that f n (0) = 0 for all n ≥ 1. (1−z)m Furthermore. September 25th. We proceed on induction on the 2 2 claim that the N th derivative of e−1/x has the form p(x) e−1/x for some polynomials p(x) q(x) and q(x) such that deg(p(x)) ≤ deg(q(x)). lim+ h→0 f n−1 (h) f n−1 (0 + h) − f n−1 (0) = lim+ h→0 h h p(h) −1/h2 = lim+ e h→0 hq(h) =0 Thus.... This is trivially true for the 0th derivative (here. Now. Take any two points w1 and w2 in the region Ω and consider a curve γ between them.. Thus.