Lebesgue's number lemma

From HandWiki
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 metric spaces. It states:

If the metric space (X,d) is compact and an open cover of X is given, then there exists a number δ>0 such that every subset of X having diameter less than δ is contained in some member of the cover.

Such a number δ is called a Lebesgue number of this cover. The notion of a Lebesgue number itself is useful in other applications as well.

Proof

Direct Proof

Let 𝒰 be an open cover of X. Since X is compact we can extract a finite subcover {A1,,An}𝒰. If any one of the Ai's equals X then any δ>0 will serve as a Lebesgue number. Otherwise for each i{1,,n}, let Ci:=XAi, note that Ci is not empty, and define a function f:X by

f(x):=1ni=1nd(x,Ci).

Since f is continuous on a compact set, it attains a minimum δ. The key observation is that, since every x is contained in some Ai, the extreme value theorem shows δ>0. Now we can verify that this δ is the desired Lebesgue number. If Y is a subset of X of diameter less than δ, then there exists x0X such that YBδ(x0), where Bδ(x0) denotes the ball of radius δ centered at x0 (namely, one can choose x0 as any point in Y). Since f(x0)δ there must exist at least one i such that d(x0,Ci)δ. But this means that Bδ(x0)Ai and so, in particular, YAi.

Proof by Contradiction

Assume X is sequentially compact, 𝒜={Uα|αJ} is an open covering of X and the Lebesgue number δ does not exist. So, δ>0, AX with diam(A)<δ such that ¬βJ where AUβ.

This allows us to make the following construction:

δ1=1, A1X where (diam(A1)<δ1) and ¬β(A1Uβ)
δ2=12, A2X where (diam(A2)<δ2) and ¬β(A2Uβ)
δk=1k, AkX where (diam(Ak)<δk) and ¬β(AkUβ)


For all n+, An since An⊄Uβ.

It is therefore possible to generate a sequence {xn} where xnAn by axiom of choice. By sequential compactness, there exists a subsequence {xnk},k+ that converges to x0X.

Using the fact that 𝒜 is an open covering, α0J where x0Uα0. As Uα0 is open, r>0 such that Bd(x0,r)Uα0. By definition of convergence, L+ such that xnpBd(x0,r2) for all pL.

Furthermore, M+ where δM=1K<r2. So, z+zMdiam(AM)<r2.

Finally, let q+ such that nqM and qL. For all xAnq, notice that:

  • d(xnq,x)diam(Anq)<r2 because nqM.
  • d(xnq,x0)<r2 because qL which means xnqBd(x0,r2).

By the triangle inequality, d(x0,x)<r, implying that AnqUα0 which is a contradiction.




References