On a Hausdorff topological space $X$ with a sigma algebra $ Σ$ at least as fine as the Borel sigma algebra,

  • a measure $\mu$ is said to be inner regular, if for every set $A \in Σ$, $\mu(A) = \sup \{ \mu(K) | \text{ compact }K \subseteq A \}$.
  • a measure is said to be tight, if for all $ε > 0$, there is some compact subset $K$ of $X$ such that $μ(X - K) < ε$.

Wikipedia says that a measure is inner regular iff it is tight. I was wondering why? Is it still true when the topological space $X$ is not necessarily Hausdorff? References are also appreciated!

Thanks and regards!


Define $\: X = \omega \:$ and $\: \mathcal{T} = 2^X \:$ and $\: \Sigma = 2^X \:$. $\;\;$ $\langle X,\mathcal{T}\hspace{.01 in}\rangle$ is a Hausdorff topological space.
Define $\: \mu_c : \Sigma \to [0,\scriptsize+\normalsize\infty] \:$ to be the counting measure.

$\operatorname{Borel}(\langle X,\mathcal{T}\hspace{.01 in}\rangle) = \operatorname{Borel}\left(\left\langle X,2^X\right\rangle\right) = 2^X = \Sigma \;\;\;$ and $\mu_c$ is inner regular but not tight.

$\omega = \{0,1,2,3,4,5,...\} \;\;$ (http://en.wikipedia.org/wiki/Ordinal_number)

For all finite subsets $S$ of $\omega$, $S$ is compact, which gives $\mu(S) \leq \operatorname{sup}(\{\mu_c(K) : (K \text{ is compact) and } (K\subseteq S)\}) = \mu(S) \;\;$.

For all compact subsets $K$ of $X$, $K$ is finite, so $\: X-K \:$ is infinite, so $\: 0 < 1 < \scriptsize+\normalsize\infty = \mu_c(X-K) \;\;$.

Yes. $\:$ azarel mentioned one direction, I don't know anything about the other.

  • $\begingroup$ Thanks! (1) What is $\omega$ in $X=\omega$? (2) Why $μ_c$ is inner regular but not tight? (3) So are you saying even when $X$ is Hausdorff, it is not true that a measure is inner regular iff it is tight? Is only one direction true or both directions are false? $\endgroup$ – Tim Jan 28 '12 at 7:04
  • $\begingroup$ If $\mu$ is a finite measure then inner regular clearly implies tight (you don't need the Hausdorff hypothesis) $\endgroup$ – azarel Jan 28 '12 at 7:20
  • $\begingroup$ Thanks! If Wikipedia is wrong on "a measure is inner regular iff it is tight" for a Hausdorff topological space, what would you suggest to modify the conditions on the topological space and/or the measure so that it is true? $\endgroup$ – Tim Jan 28 '12 at 7:49
  • $\begingroup$ @azarel: Thanks!(1) Under what some conditions, will tight implies inner regular? (2) Do you have some idea what Wiki is actually trying to say instead, since it is not correct? $\endgroup$ – Tim Jan 28 '12 at 15:00
  • 1
    $\begingroup$ Well, $\omega$ is a certain ordinal number. Set theorists identify this with a certain set. So saying $X = \omega$ is an obscure way to say $X = \mathbb N$. Since Ricky used no special properties of $\omega$ or of $\mathbb N$, it may have been preferable to say that $X$ is any countably infinite set. $\endgroup$ – GEdgar Aug 17 '12 at 14:48

Wikipedia is wrong because its definition of a tight measure as "for all ε>0, there is some compact subset K of X such that μ(X−K)<ε" fails to note that the measure must be a probability (or totally finite) measure. (Parthasarathy, 1967, Probability Measures on Metric Spaces, p. 26, fn 1.) Otherwise, a measure μ is said to be tight, if for every set A∈Σ, μ(A)=sup{μ(K)| compact K⊆A} and inner regular, if for every set A∈Σ, μ(A)=sup{μ(F)| F closed and F⊆A}.

  • $\begingroup$ ... and in non-Hausdorff spaces, a compact set need not be closed?? $\endgroup$ – GEdgar Aug 17 '12 at 14:51

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.