Proving a Certain Set is Closed in the Uniform Topology

Bashyboy
Messages
1,419
Reaction score
5

Homework Statement


Let ##Q = \{(x_1,x_2,...,) \in \mathbb{R}^\omega ~|~ \lim x_n = 0 \}##. I would like to show this set is closed in the uniform topology, which is generated by the metric ##\rho(x,y) = \sup d(x_i,y_i)##, where ##d## is the standard bounded metric on ##\mathbb{R}##.

Homework Equations

The Attempt at a Solution



Let ##x = (x_n) \in \overline{Q} - Q##. Then ##\lim x_n = \ell \neq 0##, and WLOG take ##\ell > 0##. In my attempt, I mistakenly thought that ##\prod (x_n - \ell, x_n + \ell)## was a basis element for the product topology containing ##x##, concluding from this that it must be open in the uniform topology. From there I proved that if ##z = (z_n) \in \prod (x_n -\ell, x_n + \ell)##, then it cannot converge to ##0##, which gives the desired contradiction.

My question is, is ##\prod (x_n -\ell, x_n + \ell)## in fact open in the uniform topology?
 
Physics news on Phys.org
Bashyboy said:
Let ##x = (x_n) \in \overline{Q} - Q##. Then ##\lim x_n = \ell \neq 0##,
Is ##\overline{Q} ## only convergent sequences?
 
Yes. ##\overline{Q}## a subset of ##\mathbb{R}^\omega## which consists of all convergent real sequences.
 
Bashyboy said:
Yes. ##\overline{Q}## a subset of ##\mathbb{R}^\omega## which consists of all convergent real sequences.
But there is nothing in the problem statement that allows you to restrict attention that subset.
 
haruspex said:
But there is nothing in the problem statement that allows you to restrict attention that subset.

I don't quite follow...

I have a topological space ##\mathbb{R}^\omega##, which I defined in a previous post; the space is endowed with the uniform topology, defined in the original post. Now, I have a subset ##Q## which I am trying to show is closed in this topology. Yes, you are right that I am restricting my attention to this particular subset, but I am doing so purposefully.
 
Bashyboy said:
I don't quite follow...

I have a topological space ##\mathbb{R}^\omega##, which I defined in a previous post; the space is endowed with the uniform topology, defined in the original post. Now, I have a subset ##Q## which I am trying to show is closed in this topology. Yes, you are right that I am restricting my attention to this particular subset, but I am doing so purposefully.
My problem is that I so far see no connection between the thing to be proved and neither the ##\overline Q## subset nor the product topology you mention.
At the same time, I see no reference made to the metric nor to any usual definition of closure. If you are appealing to some standard results, please quote them.

I understand that (xn-l) is a sequence, but I have no idea what you mean by ##\prod (x_n - \ell, x_n + \ell)##. ## (x_n - \ell, x_n + \ell) ## is a pair of numbers.## ((x_n - \ell), (x_n + \ell)) ## would be a pair of sequences. For a product topology we need pairs of sets.

Edit: I believe I have a fairly short proof from first principles.
 
Last edited:
haruspex said:
My problem is that I so far see no connection between the thing to be proved and neither the ¯¯¯¯QQ¯\overline Q subset nor the product topology you mention.
At the same time, I see no reference made to the metric nor to any usual definition of closure. If you are appealing to some standard results, please quote them.

I understand that (xn-l) is a sequence, but I have no idea what you mean by ∏(xn−ℓ,xn+ℓ)∏(xn−ℓ,xn+ℓ)\prod (x_n - \ell, x_n + \ell). (xn−ℓ,xn+ℓ)(xn−ℓ,xn+ℓ) (x_n - \ell, x_n + \ell) is a pair of numbers.((xn−ℓ),(xn+ℓ))((xn−ℓ),(xn+ℓ)) ((x_n - \ell), (x_n + \ell)) would be a pair of sequences. For a product topology we need pairs of sets.

Edit: I believe I have a fairly short proof from first principles.

Points in ##\mathbb{R}^\omega## can be regarded as points in a space, or they can be thought of as real convergent sequences in ##\mathbb{R}##. When I write ##(x_n)## or ##(x_n-\ell)##, I am referring to either one of these interpretations. When I write ##\prod (x_n - \ell, x_n + \ell)##, on the other hand, I am writing a subset of the product space ##\mathbb{R}^\omega##. This is in fact a basis element in the box topology. What I am trying to show is that it is open in the uniform topology, but I have been unsuccessful.

Every approach I have tried requires that I construct a basis element that containing no sequences that converge to zero. You say that you have a solution. Would you mind offering a hint?
 
Bashyboy said:
You say that you have a solution. Would you mind offering a hint?
The definition of closure in a metric space with which I am familiar is that the set contains all its limit points. So I took a sequence (of sequences) which, according to the given metric, converges, and showed ithat the sequence it converges to converges to zero.
 
haruspex said:
The definition of closure in a metric space with which I am familiar is that the set contains all its limit points. So I took a sequence (of sequences) which, according to the given metric, converges, and showed ithat the sequence it converges to converges to zero.

Unfortunately, I do not have that characterization of closure in a metric space. I only have the definition, which states that the closure of ##A## is the intersection of all closed sets containing ##A##, and the general characterization that ##x \in \overline{A}## if and only if every open neighborhood of ##x## intersects ##A##.
 
  • #10
Bashyboy said:
Unfortunately, I do not have that characterization of closure in a metric space. I only have the definition, which states that the closure of ##A## is the intersection of all closed sets containing ##A##,
Ok, but that only defines closure in terms of other closed sets, so cannot be fundamental.
See bullet 2 at
http://mathworld.wolfram.com/ClosedSet.html

Bashyboy said:
and the general characterization that ##x \in \overline{A}## if and only if every open neighborhood of ##x## intersects ##A##.
Ok, assuming the overline here means closure of, that matches the definition at bullet 4 of
http://mathworld.wolfram.com/SetClosure.html
In a metric space, if every open neighbourhood of x intersects set S then it follows that there exists an infinite sequence in S that converges to x. Thus x is a limit point of S. Since a closed set is one that is its own closure, the definition then says that a closed set contains all its limit points.
 
Back
Top