If $M$ is a metric space then the set of Cauchy sequences is closed in $\mathcal{B}(\mathbb{N},M)$ (set of bounded functions $x : \mathbb{N} \rightarrow M$).

My attempt:

Let $A= \lbrace x \in \mathcal{B}(\mathbb{N},M) | \space x_n \space \text{is Cauchy} \rbrace$.

I tried to show that $\mathcal{B}(\mathbb{N},M) - A = \lbrace x:\mathbb{N} \rightarrow M | \space x_n \space \text{is not Cauchy} \rbrace$ is open.

If $x \in \mathcal{B}(\mathbb{N},M) - A$ then $x_n$ is not a Cauchy sequence, so

$ \forall N \in \mathbb{N}, \exists \epsilon>0$ so that $m,n >N \rightarrow d(x_n,x_m)> \epsilon $

Now, fix some $N \in \mathbb{N}$ and consider the ball $B(x,\frac{\epsilon}{4})= \lbrace y \in \mathcal{B}(\mathbb{N},M) | \quad\|x-y\| < \frac{\epsilon}{4} \rbrace$

Let $y \in B(x, \frac{\epsilon}{4}) \Rightarrow \|x-y\|< \frac{\epsilon}{4} \Rightarrow \sup_{n \in \mathbb{N}} d(x_n,y_n) < \frac{\epsilon}{4}$

Then, if $m,n >N$ we have $d(x_n,y_n)<\frac{\epsilon}{4}$

Now, if $m,n>N$ we have that

$d(y_n,y_m) > d(x_n,y_m) - d(x_n,y_n)>d(x_n,x_m)-d(x_m,y_m)-d(x_n,y_n)>\epsilon -\frac{2\epsilon}{4}=\frac{\epsilon}{2}$

Then, $y_n$ is not a Cauchy sequence. So, $B(x,\frac{\epsilon}{4}) \subset \mathcal{B}(\mathbb{N},M)-A$

Then, every point in $\mathcal{B}(\mathbb{N},M)-A$ is a interior point, follows that $\mathcal{B}(\mathbb{N},M)-A$ is open and $A$ is closed.


It's just fine. That's how I would have done it. Two minor suggestions:

  1. You should have stated which distance you're using in $\mathcal{B}(\mathbb{N},M)$.
  2. Plase type $A\setminus B$ and not $A-B$. Sometimes, the later option is ambiguous.

There is just one error. You have not correctly negated the statement "$x$ is a Cauchy sequence". You have written:

If $x \in \mathcal{B}(\mathbb{N},M) - A$ then $x_n$ is not a Cauchy sequence, so

$ \forall N \in \mathbb{N}, \exists \epsilon>0$ so that $m,n >N \rightarrow d(x_n,x_m)> \epsilon $.

It should instead be:

If $x \in \mathcal{B}(\mathbb{N},M) - A$, then $\{ x_n \}_{n \in \mathbb{N}}$ is not a Cauchy sequence, so

$\exists\ \epsilon > 0$ such that $\forall\ N \in \mathbb{N}$, there exist $m,n > N$ such that $d(x_n,x_m) \geq \epsilon$.

So, the part:

Now, if $m,n>N$ we have that ...

needs to be modified slightly, into:

Now, choose $m,n > N$ such that $d(x_n,x_m) \geq \epsilon$. Then, for this value of $m,n \in \mathbb{N}$ we have that ...

Apart from this, the proof is fine.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.