0% found this document useful (0 votes)
85 views7 pages

MAT2125 Winter 2018 Assignment 3

This document outlines 5 questions for a math assignment due on February 15. The questions cover topics like convergence of series, norms, and convex hulls. Students are asked to prove several statements and show that specific sets are closed or sequences converge.

Uploaded by

kang
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
0% found this document useful (0 votes)
85 views7 pages

MAT2125 Winter 2018 Assignment 3

This document outlines 5 questions for a math assignment due on February 15. The questions cover topics like convergence of series, norms, and convex hulls. Students are asked to prove several statements and show that specific sets are closed or sequences converge.

Uploaded by

kang
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/ 7

MAT2125 Winter 2018

Assignment 3
Due Thursday February 15,
in the assignment drop box in King Edward 585 by 11:30
or
in the lecture by 13:00.

Each question is worth 10 marks.

1. Determine the set of x ≥ 0 for which the series



X 1
n=1
n + xn

converges.

2. Do Exercise 3.2.7: Prove that a series ∞


P
n=1 an converges
P∞ absolutely if and

only if for every subsequence (ank )k=1 , the “subseries” k=1 ank converges.

3. Let k · k and k · k0 be norms on Rd . Define k · k00 : Rd → [0, ∞) by

kak00 := max{kak, kak0 }.

Prove that k · k00 is a norm.

4. Define
A := {x ∈ Rd : kxk1 ≤ 1}.
(Note that we are using k · k1 here, not the Euclidean norm.)
Prove that A is a closed set.

5. For a set S ⊆ Rd , define the convex hull of S to be

co(S) := {t1 x1 +· · ·+tk xk : t1 , . . . , tk ∈ [0, 1], t1 +· · ·+tk = 1, x1 , . . . , xk ∈ S}.

Suppose we have k sequences in Rd , namely (a1,n )∞ ∞


n=1 , . . . , (ak,n )n=1 . Let
(bn )∞ d d
n=1 be another sequence in R , and let L ∈ R . Suppose that for each
n ∈ N≥1 ,
bn ∈ co{a1,n , . . . , ak,n }.
Suppose also that for each i = 1, . . . , k,

lim ai,n = L.
n→∞

Prove that
lim bn = L.
n→∞

This can be thought of as a version of the Squeeze Theorem for Rd .

Bonus (Up to 5 bonus marks) Prove that there is a sequence (an )∞


n=1 of
numbers in (0, ∞) such that:
an+1 an+1
lim sup = ∞, lim inf = 1,
n→∞ an n→∞ an
yet

X
an converges.
n=1

Solutions.
1. This converges for x ∈ (1, ∞) and diverges for x ∈ [0, 1]. (1 mark)
For x ≤ 1, we have
1 1
n
≥ . (1 mark)
P∞ n 1+ x n+1
The Harmonic Series n=1 n diverges. (1 mark)
Therefore, so does its tail
∞ ∞
X 1 X 1
= . (1 mark)
n=2
n n=1 n + 1

Hence by the Comparison Test,



X 1
diverges. (1 mark)
n=1
n + xn

For x > 1, we use the Ratio Test. (1 mark)


We have
1
n+1+xn+1 n + xn
1 =
n+xn
n + 1 + xn+1
1 xnn + 1
= n+1 (1 mark)
x xn + 1
By the Algebra of Limits, we thus get
1 n
n+1+xn+1 1 xn
+1
lim 1 = lim n+1
n→∞
n+xn
x n→∞ xn + 1
10+1
=
x0+1
1 (2 marks)
= < 1.
x
Hence by the Ratio Test, the series converges. (1 mark)

2. ⇒: Suppose that ∞
P
n=1 an converges absolutely, i.e.,


X
|an | converges.
n=1

Set L := ∞
P
n=1 |an |. (1 mark)
∞ ∞
Take a subsequence
P∞ (a )
nk k=1 of (a )
n n=1 . We’ll use the Boundedness Test
to prove that k=1 |ank | converges. (1 mark)
It will follow by the Absolute Convergence Test that ∞
P
a
k=1 nk converges.
(1 mark)
For any K ∈ N≥1 , we have
K
X nK
X
|ank | ≤ |an | ≤ L. (1 mark)
k=1 n=1

Hence L is an upper bound for the partial sums of the series ∞


P
k=1 ank ,
so by the Boundedness Test, this series converges. (1 mark)
⇐: We will prove the contrapositive: we assume that

X
|an | diverges,
n=1
and we will prove that there is a subsequence (ank )∞
k=1 for which


X
ank diverges. (1 mark)
k=1

As in the proof of the Absolute Convergence Test, define

(an )+ := max{an , 0} and


(an )− := max{−an , 0}. (1 mark)
Then since |an | = (an )+ + (an )− , it follows that either

X ∞
X
(an )+ diverges or (an )− diverges
n=1 n=1

(possibly
P both). (1 mark)
If ∞ (a
n=1 n + ) diverges, then define the subsequence (a ∞
nk k=1 Pconsist
) to
of all the positive terms of the sequence (an )n=1 . It follows that ∞

k=1 ank
diverges (to ∞).P∞ (1 mark)

Similarly, n=1 (an )+ diverges, then define the subsequence (ank )k=1 to
consist
P∞ of all the negative terms of the sequence (an )∞ n=1 . It follows that
a
k=1 nk diverges (to −∞). (1 mark)

3. Positive definiteness: For a ∈ Rd ,

kak00 = 0 ⇐⇒ kak = 0 and kak0 = 0 (1 mark)


⇐⇒ a = 0. (1 mark)
Homogeneity: For a ∈ Rd and c ∈ R,
kcak00 = max{kcak, kcak0 } (1 mark)
= max{|c| · kak, |c| · kak0 } (1 mark)
= |c| max{kak, kak0 } (1 mark)
= |c| · kak00 .
Triangle inequality: For a, b ∈ Rd ,

ka + bk00 = max{ka + bk, ka + bk0 } (1 mark)


≤ max{kak + kbk, kak0 + kbk0 } (1 mark)
≤ max{kak + kbk, ka0 k + kb0 k,
ka0 k + kbk, kak + kb0 k} (1 mark)
= max{kak, kak0 } + max{kbk, kbk0 } (1 mark)
= kak00 + kbk00 . (1 mark)

4. We will prove this using Proposition 4.3.9, namely by showing that if


(an )∞ d
n=1 is a sequence in A which converges to a ∈ R then a ∈ A. (2 marks)
For each n we have kan k1 ≤ 1. (1 mark)
Hence,

kak1 ≤ kan k1 + kan − ak1 (∆-ineq. for k · k1 ) (2 marks)


≤ kan k1 + dkan − ak2 (Exercise 4.1.3) (2 marks)
≤ 1 + dkan − ak2 . (1 mark)
Since an → a, it follows that

kan − ak2 → 0 as n → ∞, (1 mark)

and therefore kak1 ≤ 1, as required. 1 mark)


(Alternatively, one can use that kan − ak1 → 0 by Proposition 4.2.2.)

Alternate solution.
We will prove that Rd \ A is open. (2 marks)
We have
Rd \ A = {x ∈ Rd : kxk1 > 1}. (1 mark)
Let a ∈ Rd \ A. Set r := 1−kak
d
1
> 0. (1 mark)
Then if x ∈ B(a; r), we have
kxk1 ≥ kak1 − kx − ak1 (2 marks)
≥ kak1 − dkx − ak2 (2 marks)
> kak1 − dr = 1. (1 mark)
Therefore, B(a; r) ⊆ Rd \ A, as required. (1 mark)

5. Let us write
(1) (d)
ai,n = (ai,n , . . . , ai,n )
for each i = 1, . . . , k and n ∈ N≥1 ,

an = (a(1) (d)
n , . . . , an )

for n ∈ N≥1 , and


L = (L1 , . . . , Ld ). (2 marks)
Fix j ∈ {1, . . . , d} for the moment. Since (ai,n )∞
n=1 converges to L, it
follows from Proposition 4.2.3 that
(j)
lim ai,n = Lj ,
n→∞

for each i = 1, . . . , k. (2 marks)


Therefore, if we define
(j) (j)
m(j)
n := min{a1,n , . . . , ak,n }

(j)
for each n ∈ N≥1 , then we see that mn → Lj as n → ∞. (2 marks)
Likewise,
(j) (j)
Mn(j) := max{a1,n , . . . , ak,n } → Lj .
From the definition of the convex hull, we can see that

m(j) ≤ a(j)
n ≤ M
(j)
. (2 marks)
By the Squeeze Theorem, it follows

lim a(j)
n → Lj . (2 marks)
n→∞

By Proposition 4.2.3, we now conclude that an → L. (2 marks)

Alternate solution.
For each n, set

Mn := max{kai,n − Lk2 : i = 1, . . . , k}.


(1 mark)
Then since ai,n → L, we have Mn → 0. (2 marks)

Claim. kan − Lk2 ≤ Mn for each n. (1 mark)


To prove the claim, since an ∈ co{a1,n , . . . , ak,n }, write

an = t1 a1,n + · · · + tk ak,n

for some t1 , . . . , tk ∈ [0, 1] such that t1 + · · · + tk = 1. (1 mark)


Since t1 + · · · + tk = 1, we also have

L = t1 L + · · · + tk L.

Putting these together,


k
X
kan − Lk2 = ti (ai,n − L)


i=1 2
k
X
≤ kti (ai,n − L)k2 (2 marks)
i=1
Xk
= ti k(ai,n − L)k2
i=1
Xk
≤ ti Mn (1 mark)
i=1
= Mn .
This proves the claim.
From the claim and the Squeeze Theorem, it follows that kan − Lk2 → 0,
and thus, an → L. (2 marks)

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy