0
$\begingroup$

Does there exist an open subset $A \subset [0,1]$ such that $m_*(A)\neq m_*(\bar{A})$?

I was thinking we could approximate any set from inside by a closed set . This need not true from outside.

So I was expecting there should be some counterexample to the above statement.

Please help me to construct such an example.

Any help will be appreciated

$\endgroup$
  • $\begingroup$ Is $m_*$ outer or inner Lebesgue measure? $\endgroup$ – Alex Ortiz Feb 17 at 4:05
  • $\begingroup$ Sir it is outer measure. Is answer will change if we put inner measure? As i had only studied outer measure. I was thinking till thar inner measure will work same. $\endgroup$ – SRJ Feb 17 at 4:08
  • $\begingroup$ It is actually inconsequential since both inner and outer measure are subadditive with respect to countable unions; see my answer below. Cheers! $\endgroup$ – Alex Ortiz Feb 17 at 4:19
1
$\begingroup$

Enumerate the rational numbers in $[1/4,3/4]$ as $\{r_n:n\in\mathbf{Z}_{>0}\}$, and let $\epsilon < 1/8$. For each $n$, choose an interval $I_n$ centered at $r_n$ of width $\epsilon/2^n$, and put $A = \bigcup_{n=1}^\infty I_n$. Note that $A$ is an open subset of $\mathbf R$ contained in $[0,1]$ because it is a union of open intervals. Since $A$ contains the rationals in $[1/4,3/4]$, density of $\mathbf Q$ in $\mathbf R$ implies that the closure $\overline A$ contains the closed interval $[1/4,3/4]$.

By countable subadditivity (of Lebesgue outer or inner measure, this is an inconsequential point for the matter at hand), $$ m_*(A) \le \sum_{n=1}^\infty m_*(I_n) = \sum_{n=1}^\infty \frac{\epsilon}{2^n} = \epsilon < 1/8. $$ On the other hand, by monotonicity, $$ m_*\big(\overline A\big) \ge m_*\big([1/4,3/4]\big) = 1/2, $$ so $$ m_*(A) < 1/8 < 1/2 = m_*\big(\overline A\big). $$

$\endgroup$
  • $\begingroup$ Why not just let $A$ be an open subset of $(0,1)$ containing $\mathbb Q \cap (0,1)$ such that $m(A)<1/2?$ $\endgroup$ – zhw. Feb 17 at 4:43
  • $\begingroup$ @zhw.: That is essentially what I have "constructed" here. I was trying to give a more explicit set and only trying to be careful of avoiding the endpoints of the interval with the nitpick-y estimates. Your comment does get to the heart of what my example is trying to illustrate, though. $\endgroup$ – Alex Ortiz Feb 17 at 4:46

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.