2
$\begingroup$

Let $\mathcal F$ be a family of functions on $[0,1]$ each of which is integrable over $[0,1]$ and has $\int_a^b|f|\le b-a$ for all $[a,b] \subseteq [0,1]$. Is $\mathcal F$ uniformly integrable over $[0,1]$?

I think it is U.I. and this is what I have done so far. I wanted to see if this was right.

Let $\epsilon >0$ and let $\delta = \epsilon$. For any measurable set $A \subseteq [0,1]$ with $m(A) < \epsilon $ there exists a finite collection of open intervals, $\{I_k\}_{k=1}^n$, such that $m(I_k) < \frac{\epsilon}{n}$ for each $k$ and $A \subset \cup_{k=1}^nI_k$. This finite covering exists because any bounded subset of $\Bbb R$ is totally bounded. Then since $\int_A |f| \le \int_{\cup_{k=1}^nI_k}|f| \le \sum_{k=1}^n \int_{I_k}|f| \le \sum_{k=1}^n m(I_k) < \sum_{k=1}^n\frac{\epsilon}{n}=\epsilon$ then this implies that $\mathcal F$ is uniformly integrable over $[0,1]$. Is this right?

$\endgroup$
  • 3
    $\begingroup$ The fact that $A$ is totally bounded says $A$ can be covered by finitely many intervals of length $\epsilon/n$. It does not say that $A$ can be covered by $n$ intervals of length $\epsilon/n$. If that were right then every subset of $[0,1]$ would have measure zero. But $m(A)<\epsilon$ does say that $A$ can be covered by countably many intervals $I_k$ with $\sum m(I_k)<\epsilon$. Use that. $\endgroup$ – David C. Ullrich Jun 8 '16 at 20:42
1
$\begingroup$

This finite covering exists because any bounded subset of ℝ is totally bounded.

It is true that a finite covering exists, but you do not control the number of elements of this covering.

In order to solve the problem, we can use the following facts:

  1. If $O$ is an open subset of $[0,1]$, then $O$ can be written as a countable disjoint union of open intervals.
  2. For each Borel-measurable subset $B$ of $[0,1]$ and each positive $\delta$, there exists a closed set $F$ and an open set $O$ such that $F\subset B\subset O$ and $\lambda\left(O\setminus F\right)\lt\delta$.

Use fact 1. in order to prove that for each open set $O$, and each $f\in\mathcal F$, $ \int_O\left|f\right|\mathrm d\lambda\leqslant \lambda\left(O\right)$. Then use fact 2. to control $\int_B\left|f\right|\mathrm d\lambda$.

$\endgroup$

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.