Jump to content

Lebesgue's number lemma: Difference between revisions

From Wikipedia, the free encyclopedia
Content deleted Content added
WYedit (talk | contribs)
Added a proof by contradiction and labelled the previous proof as "Direct Proof".
Citation bot (talk | contribs)
Added publisher. | Use this bot. Report bugs. | Suggested by Dominic3203 | Category:Lemmas | #UCB_Category 25/76
 
(13 intermediate revisions by 8 users not shown)
Line 1: Line 1:
{{Short description|Given a cover of a compact metric space, all small subsets are subset of some cover set}}
{{Short description|Given a cover of a compact metric space, all small subsets are subset of some cover set}}
In [[topology]], '''Lebesgue's number lemma''', named after [[Henri Lebesgue]], is a useful tool in the study of [[compact space|compact]] [[metric space]]s. It states:
In [[topology]], the '''Lebesgue covering lemma''' is a useful tool in the study of [[compact space|compact]] [[metric space]]s.


:If the metric space <math>(X, d)</math> is compact and an [[open cover]] of <math>X</math> is given, then there exists a number <math>\delta > 0</math> such that every [[subset]] of <math>X</math> having [[diameter]] less than <math>\delta</math> is contained in some member of the cover.
Given an [[Cover (topology)#Cover in topology|open cover]] of a compact metric space, a '''Lebesgue's number''' of the cover is a number <math>\delta > 0</math> such that every [[subset]] of <math>X</math> having [[Diameter of a set|diameter]] less than <math>\delta</math> is contained in some member of the cover.


The existence of Lebesgue's numbers for compact metric spaces is given by the Lebesgue's covering lemma:
Such a number <math>\delta</math> is called a '''Lebesgue number''' of this cover. The notion of a Lebesgue number itself is useful in other applications as well.

:If the metric space <math>(X, d)</math> is compact and an open cover of <math>X</math> is given, then the cover admits some Lebesgue's number <math>\delta > 0</math>.

The notion of Lebesgue's numbers itself is useful in other applications as well.


== Proof ==
== Proof ==
Line 11: Line 15:


Let <math>\mathcal U</math> be an open cover of <math>X</math>. Since <math>X</math> is compact we can extract a finite subcover <math>\{A_1, \dots, A_n\} \subseteq \mathcal U</math>.
Let <math>\mathcal U</math> be an open cover of <math>X</math>. Since <math>X</math> is compact we can extract a finite subcover <math>\{A_1, \dots, A_n\} \subseteq \mathcal U</math>.
If any one of the <math>A_i</math>'s equals <math>X</math> then any <math> \delta > 0 </math> will serve as a Lebesgue number.
If any one of the <math>A_i</math>'s equals <math>X</math> then any <math> \delta > 0 </math> will serve as a Lebesgue's number.
Otherwise for each <math>i \in \{1, \dots, n\}</math>, let <math>C_i := X \smallsetminus A_i</math>, note that <math>C_i</math> is not empty, and define a function <math>f : X \rightarrow \mathbb R</math> by
Otherwise for each <math>i \in \{1, \dots, n\}</math>, let <math>C_i := X \smallsetminus A_i</math>, note that <math>C_i</math> is not empty, and define a function <math>f : X \rightarrow \mathbb R</math> by


Line 17: Line 21:


Since <math>f</math> is continuous on a compact set, it attains a minimum <math>\delta</math>.
Since <math>f</math> is continuous on a compact set, it attains a minimum <math>\delta</math>.
The key observation is that, since every <math>x</math> is contained in some <math>A_i</math>, the [[extreme value theorem#Generalization to metric and topological spaces|extreme value theorem]] shows <math>\delta > 0</math>. Now we can verify that this <math>\delta</math> is the desired Lebesgue number.
The key observation is that, since every <math>x</math> is contained in some <math>A_i</math>, the [[extreme value theorem#Generalization to metric and topological spaces|extreme value theorem]] shows <math>\delta > 0</math>. Now we can verify that this <math>\delta</math> is the desired Lebesgue's number.
If <math>Y</math> is a subset of <math>X</math> of diameter less than <math>\delta</math>, then there exists <math>x_0\in X</math> such that <math>Y\subseteq B_\delta(x_0)</math>, where <math>B_\delta(x_0)</math> denotes the ball of radius <math>\delta</math> centered at <math>x_0</math> (namely, one can choose <math>x_0</math> as any point in <math>Y</math>). Since <math>f(x_0)\geq \delta</math> there must exist at least one <math>i</math> such that <math>d(x_0,C_i)\geq \delta</math>. But this means that <math>B_\delta(x_0)\subseteq A_i</math> and so, in particular, <math>Y\subseteq A_i</math>.
If <math>Y</math> is a subset of <math>X</math> of diameter less than <math>\delta</math>, choose <math>x_0</math> as any point in <math>Y</math>, then by definition of [[Diameter of a set|diameter]], <math>Y\subseteq B_\delta(x_0)</math>, where <math>B_\delta(x_0)</math> denotes the ball of radius <math>\delta</math> centered at <math>x_0</math>. Since <math>f(x_0)\geq \delta</math> there must exist at least one <math>i</math> such that <math>d(x_0,C_i)\geq \delta</math>. But this means that <math>B_\delta(x_0)\subseteq A_i</math> and so, in particular, <math>Y\subseteq A_i</math>.


===Proof by Contradiction===
===Proof by Contradiction===


Assume <math>X</math> is sequentially compact, <math>\mathcal{A} = \{U_{\alpha} | \alpha \in J\}</math> is an open covering of <math>X</math> and the Lebesgue number <math>\delta</math> does not exist. So, <math>\forall \delta > 0</math>, <math>\exists A \subset X</math> with <math>diam (A) < \delta</math> such that <math>\neg\exists\beta \in J</math> where <math>A\subset U_{\beta}</math>.
Suppose for contradiction that <math>X</math> is [[sequentially compact space|sequentially compact]], <math>\{ U_{\alpha} \mid \alpha \in J \}</math> is an open cover of <math>X</math>, and the Lebesgue number <math>\delta</math> does not exist. That is: for all <math>\delta > 0</math>, there exists <math>A \subset X</math> with <math>\operatorname{diam} (A) < \delta</math> such that there does not exist <math>\beta \in J</math> with <math>A \subset U_{\beta}</math>.

This allows us to make the following construction:

<center><math>\delta_{1}=1</math>, <math>\exists A_{1} \subset X</math> where <math> (diam (A_{1})<\delta_{1})</math> and <math>\neg\exists \beta (A_{1} \subset U_{\beta})</math></center>

<center><math>\delta_{2}=\frac{1}{2}</math>, <math>\exists A_{2} \subset X</math> where <math>(diam (A_{2})<\delta_{2})</math> and <math>\neg\exists \beta (A_{2} \subset U_{\beta})</math></center>

<center><big>'''⋮'''</big></center>

<center><math>\delta_{k}=\frac{1}{k}</math>, <math>\exists A_{k} \subset X</math> where <math>(diam (A_{k})<\delta_{k})</math> and <math>\neg\exists \beta (A_{k} \subset U_{\beta})</math></center>


This enables us to perform the following construction:
<center><big>'''⋮'''</big></center>


<div class="center"><math>\delta_{1} = 1, \quad \exists A_{1} \subset X \quad \text{where} \quad \operatorname{diam} (A_{1}) < \delta_{1} \quad \text {and} \quad \neg\exists \beta (A_{1} \subset U_{\beta})</math></div>


For all <math> n \in \mathbb{Z}^{+}</math>, <math>A_{n} \neq \emptyset</math> since <math>A_{n} \not\subset U_{\beta}</math>.
<div class="center"><math>\delta_{2} = \frac{1}{2}, \quad \exists A_{2} \subset X \quad \text{where} \quad \operatorname{diam} (A_{2}) < \delta_{2} \quad \text{and} \quad \neg\exists \beta (A_{2} \subset U_{\beta})</math></div>


<div class="center"><math>\vdots</math></div>
It is therefore possible to generate a sequence <math>\{x_{n}\}</math> where <math>x_{n} \in A_{n}</math> by axiom of choice. By sequential compactness, there exists a subsequence <math>\{x_{n_{k}}\}, k \in \mathbb{Z}^{+}</math> that converges to <math>x_{0} \in X</math>.


<div class="center"><math>\delta_{k}=\frac{1}{k}, \quad \exists A_{k} \subset X \quad \text{where} \quad \operatorname{diam} (A_{k}) < \delta_{k} \quad \text{and} \quad \neg\exists \beta (A_{k} \subset U_{\beta})</math></div>
Using the fact that <math>\mathcal{A}</math> is an open covering, <math>\exists \alpha_{0} \in J</math> where <math>x_{0} \in U_{\alpha_{0}}</math>. As <math>U_{\alpha_{0}}</math> is open, <math>\exists r > 0</math> such that <math>B_{d}(x_{0},r) \subset U_{\alpha_{0}}</math>. By definition of convergence, <math> \exists L \in \mathbb{Z}^{+}</math> such that <math>x_{n_{p}} \in B_{d} \left(x_{0},\frac{r}{2}\right)</math> for all <math>
p \geq L</math>.


<div class="center"><math>\vdots</math></div>
Furthermore, <math>\exists M \in \mathbb{Z}^{+}</math> where <math> \delta_{M}=\frac{1}{K}<\frac{r}{2} </math>. So, <math>\forall z \in \mathbb{Z}^{+} z \geq M \Rightarrow diam (A_{M})<\frac{r}{2}</math>.


Finally, let <math>q \in \mathbb{Z}^{+}</math> such that <math>n_{q} \geq M</math> and <math>q \geq L</math>. For all <math>x' \in A_{n_{q}}</math>, notice that:
*<math> d(x_{n_{q}},x')\leq diam (A_{n_{q}})<\frac{r}{2}</math> because <math>n_{q} \geq M</math>.
*<math>d(x_{n_{q}},x_{0})<\frac{r}{2}</math> because <math>q \geq L</math> which means <math>x_{n_{q}} \in B_{d}(x_{0},\frac{r}{2})</math>.


Note that <math>A_{n} \neq \emptyset</math> for all <math> n \in \mathbb{Z}^{+}</math>, since <math>A_{n} \not\subset U_{\beta}</math>. It is therefore possible by the [[axiom of choice]] to construct a sequence <math>(x_{n})</math> in which <math>x_{i} \in A_{i}</math> for each <math>i</math>. Since <math>X</math> is sequentially compact, there exists a subsequence <math>\{x_{n_{k}}\}</math> (with <math>k \in \mathbb{Z}_{> 0}</math>) that converges to <math>x_{0}</math>.
By the triangle inequality, <math>d(x_{0},x')<r</math>, implying that <math>A_{n_{q}} \subset U_{\alpha_{0}}</math> which is a contradiction.


Because <math>\{ U_{\alpha} \}</math> is an open cover, there exists some <math>\alpha_{0} \in J</math> such that <math>x_{0} \in U_{\alpha_{0}}</math>. As <math>U_{\alpha_{0}}</math> is open, there exists <math>r > 0</math> with <math>B_{d}(x_{0},r) \subset U_{\alpha_{0}}</math>. Now we invoke the convergence of the subsequence <math> \{ x_{n_{k}} \} </math>: there exists <math> L \in \mathbb{Z}^{+}</math> such that
<math> L \le k</math> implies <math>x_{n_{k}} \in B_{r/2} (x_{0})</math>.


Furthermore, there exists <math>M \in \mathbb{Z}_{> 0}</math> such that <math> \delta_{M}= \tfrac{1}{M} < \tfrac{r}{2} </math>. Hence for all <math>z \in \mathbb{Z}_{> 0}</math>, we have <math>M \le z</math> implies <math>\operatorname{diam} (A_{M}) < \tfrac{r}{2}</math>.


Finally, define <math>q \in \mathbb{Z}_{> 0}</math> such that <math>n_{q} \geq M</math> and <math>q \geq L</math>. For all <math>x' \in A_{n_{q}}</math>, notice that:
*<math> d(x_{n_{q}},x') \leq \operatorname{diam} (A_{n_{q}})<\frac{r}{2}</math>, because <math>n_{q} \geq M</math>.
*<math>d(x_{n_{q}},x_{0})<\frac{r}{2}</math>, because <math>q \geq L</math> entails <math>x_{n_{q}} \in B_{r/2}\left(x_{0}\right)</math>.


Hence <math>d(x_{0},x')<r</math> by the [[triangle inequality]], which implies that <math>A_{n_{q}} \subset U_{\alpha_{0}}</math>. This yields the desired contradiction.


==References==
==References==
* {{Citation|last=Munkres|first=James R.|title=Topology: A first course|year=1974|isbn=978-0-13-925495-6|page=[https://archive.org/details/topologyfirstcou00munk_0/page/179 179]|url-access=registration|url=https://archive.org/details/topologyfirstcou00munk_0/page/179}}
* {{Citation|last=Munkres|first=James R.|title=Topology: A first course|year=1974|isbn=978-0-13-925495-6|page=[https://archive.org/details/topologyfirstcou00munk_0/page/179 179]|publisher=Prentice-Hall |url-access=registration|url=https://archive.org/details/topologyfirstcou00munk_0/page/179}}


{{DEFAULTSORT:Lebesgue's Number Lemma}}
{{DEFAULTSORT:Lebesgue's Number Lemma}}

Latest revision as of 15:59, 17 October 2024

In topology, the Lebesgue covering lemma is a useful tool in the study of compact metric spaces.

Given an open cover of a compact metric space, a Lebesgue's number of the cover is a number such that every subset of having diameter less than is contained in some member of the cover.

The existence of Lebesgue's numbers for compact metric spaces is given by the Lebesgue's covering lemma:

If the metric space is compact and an open cover of is given, then the cover admits some Lebesgue's number .

The notion of Lebesgue's numbers itself is useful in other applications as well.

Proof

[edit]

Direct Proof

[edit]

Let be an open cover of . Since is compact we can extract a finite subcover . If any one of the 's equals then any will serve as a Lebesgue's number. Otherwise for each , let , note that is not empty, and define a function by

Since is continuous on a compact set, it attains a minimum . The key observation is that, since every is contained in some , the extreme value theorem shows . Now we can verify that this is the desired Lebesgue's number. If is a subset of of diameter less than , choose as any point in , then by definition of diameter, , where denotes the ball of radius centered at . Since there must exist at least one such that . But this means that and so, in particular, .

Proof by Contradiction

[edit]

Suppose for contradiction that is sequentially compact, is an open cover of , and the Lebesgue number does not exist. That is: for all , there exists with such that there does not exist with .

This enables us to perform the following construction:


Note that for all , since . It is therefore possible by the axiom of choice to construct a sequence in which for each . Since is sequentially compact, there exists a subsequence (with ) that converges to .

Because is an open cover, there exists some such that . As is open, there exists with . Now we invoke the convergence of the subsequence : there exists such that implies .

Furthermore, there exists such that . Hence for all , we have implies .

Finally, define such that and . For all , notice that:

  • , because .
  • , because entails .

Hence by the triangle inequality, which implies that . This yields the desired contradiction.

References

[edit]
  • Munkres, James R. (1974), Topology: A first course, Prentice-Hall, p. 179, ISBN 978-0-13-925495-6