Every Closed Set in R has a Countable Dense Subset

  • Context: MHB 
  • Thread starter Thread starter joypav
  • Start date Start date
  • Tags Tags
    Closed Set
Click For Summary

Discussion Overview

The discussion revolves around the problem of demonstrating that every closed set in the real numbers has a countable dense subset. Participants explore various approaches, including the implications of compactness and the use of rational numbers, while considering both bounded and unbounded closed sets.

Discussion Character

  • Exploratory
  • Technical explanation
  • Mathematical reasoning
  • Debate/contested

Main Points Raised

  • One participant suggests starting with the properties of closed sets in R, noting that a countable dense subset is equivalent to being separable.
  • Another participant proposes considering the case when the closed set is bounded, leading to the conclusion that it is compact and thus separable.
  • There is a discussion about using finite subcovers from open covers to construct a countable dense subset, with one participant refining the approach by suggesting the use of centers of balls of decreasing radius.
  • A later reply emphasizes that a finite subcover alone does not guarantee density, proposing a method to create a dense set by considering finite $1/n$-nets.
  • One participant outlines a detailed construction of a countable dense subset for a bounded closed set, arguing that it is countable and dense based on the properties of the chosen centers.
  • Another participant suggests that the explanation could be improved for clarity and conciseness, reiterating the steps involved in proving density.
  • A participant expresses uncertainty about how to extend the argument to unbounded closed sets, feeling that there is insufficient information.
  • In response, another participant indicates that there is enough information by suggesting the decomposition of the unbounded closed set into compact subsets, which are known to be separable.

Areas of Agreement / Disagreement

Participants generally agree on the approach for bounded closed sets and the use of compactness to establish separability. However, there is disagreement and uncertainty regarding the extension of these ideas to unbounded closed sets, with multiple viewpoints on how to proceed.

Contextual Notes

The discussion includes assumptions about the properties of closed sets and compactness, but these assumptions are not universally accepted or resolved. The steps taken to prove density for unbounded sets remain unresolved.

joypav
Messages
149
Reaction score
0
Show that every closed set in R has a countable dense subset.
Let's call the set F.

I've been thinking about this problem for a little bit, and it just doesn't seem like I have enough initial information!
I tried listing some things that I know about closed sets in R:

$\cdot$ Countable dense subset is the same as being separable (I think?)
$\cdot$ F contains all of it's limit points
$\cdot$ Every cauchy sequence in F converges to a point in F
$\cdot$ If F is closed then F is $G_\delta$, a countable intersection of open sets (I thought this may be helpful because the sets are countable?)

I considered somehow utilizing the rational numbers, because they are dense in R, and for each x in F having a sequence of rational numbers that converges to x. However, this was just an idea.

Obviously I haven't done much concrete work on it. But I would appreciate a push in the right direction!
 
Physics news on Phys.org
Hi joypav,

First consider the case when $E$ is bounded. Then $E$ is compact. Can you prove in this case, that $E$ is separable (i.e., has a countable dense subset)?
 
Euge said:
Hi joypav,

First consider the case when $E$ is bounded. Then $E$ is compact. Can you prove in this case, that $E$ is separable (i.e., has a countable dense subset)?

If E is compact, then every open cover has a finite subcover. Then, basically, you can have a union of the center of each "ball" of your finite subcover. (by shrinking the radius) And that would be a dense and countable subset.
 
Not quite. Just one finite subcover will not give density. But for each $n\in \Bbb N$, we may consider finitely many centers of balls of radius $1/n$ covering $E$ (which forms a finite $1/n$-net), and let $A$ be the union of all these centers. Then we can show that $A$ is dense in $E$. Being the countable union of finite sets, $A$ is countable.
 
Here we go...

Let $E \subset R$, E closed.
Assume E is bounded $\implies$ E is compact $\implies$ every open cover of E has a finite subcover.
Consider the open cover,

$E \subset \cup_{x \in E}B(x,1/n)$, for $n \in \Bbb{N}$

Then there exists a finite subcover, for some $N \geq 1$,

$E \subset \cup^{N}_{i=1}B(x_i,1/n) \subset \cup_{x \in E}B(x,1/n) $

Define $C_n = $ the set of all centers of $B(x_i,1/n)$, $1\leq i \leq N = x_1 , x_2 , ... , x_N $

Create this set for every $n \in \Bbb{N}$.
Meaning,
$E \subset \cup^{N_1}_{i=1}B(x_i,1) \subset \cup_{x \in E}B(x,1) $, $C_1 = x_1, x_2, ... , x_{N_1} $
$E \subset \cup^{N_2}_{i=1}B(x_i,1/2) \subset \cup_{x \in E}B(x,1/2) $, $C_2 = x_1, x_2, ... , x_{N_2} $
... and so on for $C_3, C_4, ...$

Let $C = \cup_{n \in \Bbb{N}} C_n$.

Claim: C is our countable dense subset of E.
1. C is countable.
$\forall n \in \Bbb{N}, C_n \subset \cup^{N_n}_{i=1}B(x_i,1/n) $ and a subset of a countable set is countable
$\implies \forall n \in \Bbb{N}, C_n $ is countable.
$C = \cup_{n \in \Bbb{N}} C_n$, a countable union of countable sets $\implies$ C is countable.

2. C is dense in E.
Take $x \in E$. Given $\epsilon > 0$, $\exists n \in \Bbb{N}$ such that $1/n < \epsilon$.
Then, $\exists y \in C_n \subset C$ so that $x \in B(y, 1/n) \implies y \in B(x,1/n) \subset B(x,\epsilon)$
So $y \in B(x,\epsilon) \cap C \implies $ C is dense in E.

How's that look?
 
It looks better, but it could be improved by writing the following:

If $E$ is bounded, then $E$ is compact. Thus, for every $n\in \Bbb N$, there is a finite $1/n$-net $C_n$. If $C = \bigcup C_n$, then $C$ is countable, being the countable union of finite sets. Further, if $\epsilon > 0$ and $x\in E$, we may choose an $n\in \Bbb N$ such that $1/n < \epsilon$, and select $y\in C_n$ such that $d(x,y) < 1/n$. Then $d(x,y) < \epsilon$, showing that $x\in \bar{C}$. Hence, $C$ is dense in $E$.
 
Then if we want to show the same for an unbounded close set, it just seems like I don't have enough info to use?
 
There is enough info. In the general case, write $E$ as the union of sets $E_n := E\cap [-n,n]$ for $n = 1,2,3,\ldots$, and note each $E_n$, being a closed subspace of the compact interval $[-n,n]$, is compact, hence separable. See what you can do from there.
 

Similar threads

  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 2 ·
Replies
2
Views
4K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 2 ·
Replies
2
Views
3K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 44 ·
2
Replies
44
Views
7K
  • · Replies 1 ·
Replies
1
Views
2K