0
$\begingroup$

Let $s, \delta \in (0,1)$. Consider the outer measure on $\mathbb{R}$, $\mu^s_{\delta}$, defined by \begin{align*} \mu^s_{\delta}(E):=\inf \left\{\sum_{j}\lvert I_{j}\rvert^s: E \subset \bigcup_{j} I_{j}: I_{j} \text { closed intervals, } \lvert I_j\rvert\leq\delta\right\}. \end{align*} For an interval $I \subset \mathbb{R}$, $\lvert I\rvert$ denotes the length of $I$. I want to prove that if $E$ is an interval and $\delta< \lvert E\rvert$, then \begin{align*} \mu^s_{\delta}(E) \geq \delta^{s-1}\left|E\right|-\delta^s. \end{align*} I think that by definition, $\exists \epsilon>0$ such that $\mu^s_{\delta}(E)+\epsilon\geq \sum_{j}\left|I_{j}\right|^s\geq \left(\sum_{j}\left|I_{j}\right|\right)^s\geq \left|E\right|^s>\delta^s.$ I wonder how to get $\delta^{s-1}$. I feel like I need to show that the delta cover is the smallest one among all the best covers, thankswhich is achieved through infimum. Such a delta cover is a little bit larger than $\delta^{s-1}\left|E\right|-\delta^{s}$.

$\endgroup$
4
  • $\begingroup$ What does "make $\delta$ as many as possible" mean? $\endgroup$
    – LSpice
    Commented Oct 14, 2021 at 21:56
  • $\begingroup$ @LSpice Sorry, it should be as few as possible since I don't want too many summands, just the right amount to cover $E$. $\endgroup$ Commented Oct 14, 2021 at 22:13
  • $\begingroup$ But what does it mean to make $\delta$ as many, or as few, as possible? Isn't $\delta$ fixed? Do you mean as few intervals of length $\delta$ as possible? $\endgroup$
    – LSpice
    Commented Oct 14, 2021 at 22:18
  • 1
    $\begingroup$ @LSpice Yes, that's what I mean, thanks. $\endgroup$ Commented Oct 14, 2021 at 22:19

1 Answer 1

3
$\begingroup$

$\newcommand\de\delta\newcommand\ol\overline$Your goal cannot be attained in general. Indeed, suppose that $\de\in(0,1/2)$. Take any interval $E$ of length $\de_1:=|E|\in(\de,2\de)$. Then for the closure $\ol E$ of $E$ and some closed intervals $I_1$ and $I_2$ of lengths $|I_1|=\de$ and $|I_2|=\de_1-\de\le\de$ we have $E=I_1\cup I_2$, so that $$\mu_\de^s(E)\le|I_1|^s+|I_2|^s =\de^s+(\de_1-\de)^s\underset{\de_1\downarrow\de}\longrightarrow \de^s<\de^{s-1}-\de^s,$$ so that for some interval $E$ with $\de<|E|$ we have $$\mu_\de^s(E)\not\ge\de^{s-1}-\de^s.$$

(It is actually easy to see that for any interval $E$ of length $t$ we have $\mu_\de^s(E)=k\de^s+(t-k\de)^s$, where $k:=\lfloor t/\de\rfloor$.)

$\endgroup$
3
  • $\begingroup$ Do ypu mean $(k\delta)^s$ in the last equality? $\endgroup$ Commented Oct 14, 2021 at 20:25
  • $\begingroup$ Is there any way to show this is the infimum? $\endgroup$ Commented Oct 14, 2021 at 20:32
  • 1
    $\begingroup$ @Analyst_311419 : The term is $k\delta^s$, as written, not $(k\delta)^s$. Since your posted question has been fully answered, please mark it accordingly, to keep things in good order. If you have further questions on the additional statement in the parentheses, you can post them separately, and then I will answer them. $\endgroup$ Commented Oct 14, 2021 at 21:14

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .