0
$\begingroup$

Let $\mu$ be a regular "outer" measure on $\mathbb{R}^N$ (for example, the Lebesgue outer measure). By regularity I mean that for all $A\subset \mathbb{R}^N$, there is $E$ measurable with $A\subset E$ and $\mu(A)=\mu(E)$.

I have two questions which are bothering me for a long time.

1 - Let $A\subset \mathbb{R}^N$, $E$ measurable with $A\subset E$ and $\mu(A)=\mu(E)$. Let $Q$ be a cube. Is it true that $$\mu(A\cap Q)=\mu (E\cap Q)?$$

2- In the above conditions, is there any example for which $\mu(E\setminus A)\neq 0$?

Any idea or reference is appreciated.

$\endgroup$
  • $\begingroup$ That's a very weak "regularity" condition. $\;$ $\endgroup$ – user57159 Aug 17 '14 at 1:05
  • $\begingroup$ Are you sure about your definitions? I only know regularity as follows: $\mu$ is regular if for each set $A$ there is a measurable set $E$ such that $\mu(A)=\mu(E)$ AND $A\subset E$. $\endgroup$ – sranthrop Aug 17 '14 at 1:06
  • $\begingroup$ You are right @sranthrop, let me fix it. $\endgroup$ – Tomás Aug 17 '14 at 1:21
  • 3
    $\begingroup$ Does Vitali set of outer-measure exactly 1 answer part 2? For this subset of $E=[0,1]$, the outer measure is $1$. Yet, $\mu(E\setminus A)$ is positive, for else $A$ would be measurable. $\endgroup$ – user147263 Aug 17 '14 at 4:24
  • 2
    $\begingroup$ This question does not meet your own question standard, being a pure problem statement (PSQ), showing absolutely no work, no context, no knowledge-level, etc. Please improve the question. Please show some work. $\endgroup$ – Bill Dubuque Aug 19 '14 at 21:50
1
$\begingroup$

Assuming the following additional conditions I give an answer to your question.

  • For each set $A\subset\mathbb R^N$ there exists a measurable set $E$ such that $A\subset E$ and $\mu(A)=\mu(E)$.
  • Q is measurable.

Since $A\subset E$ we clearly have $\mu(A\cap Q)\leq\mu(E\cap Q)$ and $\mu(A\backslash Q)\leq\mu(E\backslash Q)$. Since $Q$ and $E$ are measurable we have $\mu(A\cap Q)=\mu(A)-\mu(A\backslash Q)=\mu(E)-\mu(A\backslash Q)\geq\mu(E)-\mu(E\backslash Q)=\mu(E\cap Q)$. Consequently, $\mu(A\cap Q)=\mu(E\cap Q)$.

$\endgroup$
  • $\begingroup$ If $Q$ is not measurable, do you think that equality may fail? $\endgroup$ – Tomás Aug 17 '14 at 1:25
  • $\begingroup$ I think one could find a counterexample, but I am not sure. Do you know anything else about your cube? Or about the measure? If $\mu$ were Borel regular for example, and if your cube were open/closed (=> Borel set), then you knew that it is automatically measurable. $\endgroup$ – sranthrop Aug 17 '14 at 1:34
  • $\begingroup$ In fact, in these particular cases you have solved the problem, I alread knew the solution (only for question 1), so I am more interested in the general case. $\endgroup$ – Tomás Aug 17 '14 at 1:55

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.