1
$\begingroup$

Let $f$ be a strictly positive (almost everywhere) measurable function which is also integrable. Let $E_n$ be a sequence of measurable sets such that $\int_{E_n}f\to 0$ as $n\to\infty.$ Is it true that $\mu(E_n)\to 0$ where $\mu$ is a positive measure with respect to $f$ is integrable. The measure space is finite.

$\endgroup$
3
  • 2
    $\begingroup$ @op obviously not. take lebesgue measure and $f(x) = \frac{1}{x^2}$ on $(1,\infty)$ and $E_n = [n,n+1]$. $\endgroup$ Commented Aug 13, 2019 at 11:46
  • $\begingroup$ In you example the measure space is not finite. $\endgroup$ Commented Aug 13, 2019 at 11:48
  • 1
    $\begingroup$ See also Chebyshev's Inequality en.wikipedia.org/wiki/Chebyshev%27s_inequality $\endgroup$
    – GEdgar
    Commented Aug 13, 2019 at 11:58

1 Answer 1

2
$\begingroup$

Yes. Suppose otherwise. Then, by passing to a subsequence, there is some $\epsilon > 0$ with $\int_{E_n} f \to 0$ but $\mu(E_n) \ge \epsilon$ for each $n$. For $k \ge 1$, let $F_k = \{x \in X : |f(x)| < \frac{1}{k}\}$. Since $\emptyset = \cap_{k \ge 1} F_k$ and $(X,\mu)$ is finite, there is some $k$ with $\mu(F_k) < \frac{\epsilon}{2}$. Then $\mu(F_k^c) \ge \mu(X)-\frac{\epsilon}{2}$ implies that for any $n \ge 1$, $\mu(E_n \cap F_k^c) \ge \frac{\epsilon}{2}$ and so $\int_{E_n} f \ge \int_{E_n \cap F_k^c} f \ge \int_{E_n \cap F_k^c} \frac{1}{k} \ge \frac{\epsilon}{2}\frac{1}{k}$. Contradiction.

$\endgroup$
3
  • $\begingroup$ Thank you good job. $\endgroup$ Commented Aug 13, 2019 at 11:58
  • $\begingroup$ @DavidMitra I mean, you can always create counterexamples by ignoring context and generalizing in the way you please. I of course was using the fact that $F_{k+1} \subseteq F_k$ for each $k$. But also, if you just want to give an example of arbitrary sets of large measure with empty intersection, just do $F_1 = (0,\frac{1}{2})$ and $F_n = (\frac{1}{2},1)$ for $n \ge 2$. $\endgroup$ Commented Aug 13, 2019 at 13:27
  • $\begingroup$ What do you mean by your last comment? I guess for finite measure space you already have given a concrete argument. $\endgroup$ Commented Aug 14, 2019 at 8:39

You must log in to answer this question.

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