1. hw 5
Recall: Let S be a subset of R. We say that S is bounded below if there exists a ∈ R such that x ≥ a for any x ∈ S. In this case, a is called a lower bound for S. Assume that S is bounded below.
We say that a lower bound L for S is the greatest lower bound for S if a is a lower bound for S, then a ≤ L. The greatest lower bound for S is denoted by inf S.
(1) A subset S of Rn is finite if number of elements of S are finite. Show that any finite subset of Rn is closed.
Proof. Let p be a point of Rn. Let us show that {p} is a closed subset of Rn. It is equivalent to show that Rn\ {p} is open. For q ∈ Rn\ {p}, choose = d(p, q)/2. We leave it to the reader to verify that B(q, ) is contained in Rn\{p}. (We have done similar examples before.) We show that Rn\ {p} is open.
Let S be a finite subset of Rn. Then S = {p1, · · · , pk}. Since any finite union of closed sets is closed, S =Sk
i=1{pk} is closed.
(2) Let A be a finite subset of Rn and
B = {x ∈ Rn: d(x, y) ≤ 1 for some y ∈ A}.
Show that B is closed.
Proof. For any y ∈ A, show that the set D(y, 1) = {x ∈ Rn : d(x, y) ≤ 1} is closed. Since A is a finite subset of Rn and any finite union of closed set is closed,
B = [
y∈A
D(y, 1)
is closed.
(3) Let A be a subset of Rn and A0 be its derived set (the set of all accumulation points) of A.
Prove that A0 is closed. Is (A0)0= A0 for all A?
Proof. Let A = {1/n : n ∈ N} be a set considered as a subset of R. The set of accumulation point of A consists of a single point A0 = {0}. We see that A0 consists of an isolated point of A0. Thus (A0)0 = ∅ 6= A0.
(4) Let S = {(r, s) ∈ R2: r, s ∈ Q}.
(a) Find S.
S = R2.
(b) Find Sc. Here Sc= R2\ S.
Sc= R2. (c) Find ∂S.
Using the previous results, we find ∂S = R2.
(5) We have two equivalent definitions of convergence of sequences in Rp.
1
2
Definition 1.1. (Defn I) A sequence (an) in Rp is convergent to a point a ∈ Rp if for any
> 0, there exists a natural number N (this natural number depends on the choice of ) so that
an∈ B(a, ) whenever n ≥ N.
Definition 1.2. (Defn II) A sequence (an) in Rpis convergent to a point a ∈ Rpif for any neighborhood U of a, there exists a natural number NU of a such that
an∈ U whenever n ≥ NU.
We have proved in class that Defn I is equivalent to Defn II. Students may not be familiar with Defn II. In this exercise, we will let you experience how NU depends on the choice of U.
Let (an) be the sequence defined by an=
sin1
n,1 n
, n ∈ N.
(a) Use Defn I to show that (an) is convergent to 0 = (0, 0).
Proof. By 0 < sin(1/n) < 1/n for any n ≥ 1, kank =
r sin2 1
n+ 1 n2 <
r2 n2 < 2
n.
For any > 0, we choose N = [2/] + 1. Then for any n ≥ N, kank < . In other words, an∈ B(0, ) whenever n ≥ N. Here 0 = (0, 0). (b) Let A = {(x, y) : x2+ 4y2≤ 10−5}. Show that A is a neighborhood of 0, i.e. find > 0 so that B(0, ) ⊂ A. (A is a neighborhood of 0 if and only if 0 is an interior point of A.) Use this to find a natural number N so that an∈ A for all n ≥ N. This number will denoted by NA.
Proof. The key point is this: 0 is an interior point of A. Find an such that B(0, ) ⊂ A. For example, you might choose = 1/104. Assuming (x, y) ∈ B(0, 1/104). Then
|x| < 1/104 and |y| < 1/104 and hence x2+ 4y2< 5
108 < 1 105.
This implies B(0, 1/104) ⊂ A. Using the result in (a) and = 1/104, we take N = [2 · 104] + 1 = 2 · 104+ 1. For any n ≥ 2 · 104+ 1, one has
an∈ B(0, 1/104) ⊂ A.
Hence an ∈ A whenever n ≥ 2 · 104+ 1. We take NA= 2 · 104+ 1.
(c) Let B =
(x, y) : − 1
106 ≤ x, y ≤ 1 106
. Show that B is a neighborhood of 0 i.e. find δ > 0 so that B(0, δ) ⊂ B. Use δ to find a natural number M so that an ∈ B for all n ≥ M. This number will denoted by M = NB.
Proof. We simply take = 1/106. For (x, y) ∈ B(0, 1/106), we have |x|, |y| < 1/106. Thus B(0, 1/106) ⊂ B. Using the result in (a), we can choose N= [2·106]+1 = 2·106+1.
Then
an ∈ B(0, 1/106) ⊂ B
whenever n ≥ 2 · 106+ 1. This implies that an∈ B whenever n ≥ 2 · 106+ 1. We thus take NB= 2 · 106+ 1.
3
(6) Prove that the following sequences have a convergent subsequence in the given spaces.
We use the Bolzano-Weierstrass Theorem to prove the results. We show that these se- quences are bounded.
(a) Let (an) be the sequence of real numbers in R defined by an= sin n, n ∈ N.
Proof. Since |an| ≤ 1 for any n ≥ 1, (an) is bounded.
(b) Let (bn) be the sequence of real numbers in R defined by
bn= (−1)nsin en, n ∈ N.
Proof. We know |bn| = | sin(en)| ≤ 1 for any n ≥ 1. Hence (bn) is bounded.
(c) Let (an) be the sequence in R2defined by
an = (cos n, sin n), n ∈ N.
Proof. Since kank = 1 for all n ≥ 1, (an) is bounded.
(d) Let (bn) be the sequence in R3 defined by
bn=
cos n, e−ncos n2, cos1 n
, n ∈ N.
Proof. Since | cos n| ≤ 1 and |e−ncos n2| ≤ e−1 and | cos(1/n)| ≤ 1 for all n ≥ 1, (bn) is bounded.