Skip to main content

Homework 5

Exercise: Use the definition of convergence to determine the following limits in R\mathbb{R}:

  1. limn3n+1n+5\lim_{n\rightarrow\infty}\frac{3n+1}{n+5}
  2. limn(n2+nn)\lim_{n\rightarrow\infty}(\sqrt{n^2+n}-n)

Proof.

  1. limn3n+1n+5=3\lim_{n\rightarrow\infty}\frac{3n+1}{n+5}=3
  2. limn(n2+nn)=nn2+n+n=12\lim_{n\rightarrow\infty}(\sqrt{n^2+n}-n)=\frac{n}{\sqrt{n^2+n}+n}=\frac{1}{2}

Exercise:

  1. Give an example of a convergent sequence {sn}\{s_n\} of positive real numbers such that limnsn+1sn=1\lim_{n\rightarrow\infty}\frac{s_{n+1}}{s_n}=1
  2. Give an example of a divergent sequence {sn}\{s_n\} of positive real numbers such that limnsn+1sn=1\lim_{n\rightarrow\infty}\frac{s_{n+1}}{s_n}=1

Proof.

  1. sn=1s_n=1
  2. sn=ns_n=n

Exercise:

  1. Give an example of a convergent sequence {sn}\{s_n\} of real numbers such that the set {sn:nN}\{s_n:n\in\mathbb{N}\} has exactly one limit point.
  2. Give an example of a convergent sequence {sn}\{s_n\} of real numbers such that the set {sn:nN}\{s_n:n\in\mathbb{N}\} has no limit point.
  3. Prove: If a sequence {sn}\{s_n\} of real numbers converges, then the set {sn:nN}\{s_n:n\in\mathbb{N}\} has at most one limit point.

Proof.

  1. sn=1/ns_n=1/n
  2. sn=1s_n=1
  3. Suppose limnsn=s\lim_{n\rightarrow\infty}s_n=s. Enough to show that if there is a limit point tt for the set {sn:nN}\{s_n:n\in\mathbb{N}\}, then s=ts=t. Deny. Then pick any positive real number ϵ<st/2\epsilon<|s-t|/2. There is an integer NN such that nNn\geq N implies ssn<ϵ|s-s_n|<\epsilon. Observe that now tt is still the limit point of the set {sn:nN}\{s_n:n\geq N\}, because chopping off finitely many points doesn't change the limit point. However, by the choice of ϵ\epsilon and NN, for all nNn\geq N, tsn>st/2>0|t-s_n|>|s-t|/2>0. A contradiction.

Exercise: Suppose XX\neq\emptyset is equipped with the discrete metric. Characterize all convergent sequences in XX.

Proof.
We say a sequence {sn:nN}\{s_n:n\in\mathbb{N}\} in XX is eventually constant if and only if there are NNN\in\mathbb{N} and s Xs\in\ X such that for all n>Nn>N, sn=ss_n=s.
We claim that all eventually constant sequences are all convergent sequences in XX. Easy to see they are convergent. Left to see that every convergent sequence is eventually constant. Suppose {sn:nN}\{s_n:n\in\mathbb{N}\} convergences to ss. By definition of convergence, there is an integer NN such that n>Nn>N implies that d(sn,s)<1/2d(s_n,s)<1/2, which, indeed, implies sn=ss_n=s because of the discrete metric.

Exercise:

  1. Prove: If a sequence {sn}R\{s_n\}\subset\mathbb{R} converges in R\mathbb{R} then the sequence {sn}\{|s_n|\} converges in R\mathbb{R}.
  2. Give an example of a sequence {sn}R\{s_n\}\subset\mathbb{R} such that {sn}\{|s_n|\} converges but {sn}\{s_n\} does not converge.
  3. For a sequence {sn}R\{s_n\}\subset\mathbb{R} we define the sequences {sn+}\{s_n^+\} and {sn}\{s_n^-\} where sn+:=max{sn,0}s_n^+:=\max\{s_n, 0\} and sn:=min{sn,0}s_n^-:=\min\{s_n,0\} for all nNn\in\mathbb{N}. Prove: the sequence {sn}\{s_n\} converges if and only if {sn+}\{s_n^+\} and {sn}\{s_n^-\} converge.

Proof.

  1. Suppose {sn}\{s_n\} converges to ss. Then for every ϵ>0\epsilon > 0, there is an integer NN such that nNn\geq N implies sns<ϵ|s_n-s| < \epsilon. Notice that snssns<ϵ||s_n|-|s||\leq |s_n-s| < \epsilon. Hence {sn}\{|s_n|\} converges to s|s|.
  2. For example, sn=(1)ns_n=(-1)^n.
  3. Suppose {sn}\{s_n\} converges. Notice that sn+=(sn+sn)/2s_n^+=(s_n+|s_n|)/2 and sn=(snsn)/2s_n^-=(s_n-|s_n|)/2. By (a) and theorem 3.4, we know {sn+}\{s_n^+\} and {sn}\{s_n^-\} also converge. On the other hand, if {sn+}\{s_n^+\} and {sn}\{s_n^-\} converge, {sn}\{s_n\} converges since sn=sn++sns_n=s_n^++s_n^-.