0
$\begingroup$

Let $(X, \mathcal{F}, \mu)$ be a measurable space and $(f_n)$ a sequence of nonnegative measurable functions.

Prove that when $$ \lim_{n \to \infty} \int f_n d \mu \to 0 $$ then $f_n$ converges with measure $\mu$ to 0

My attempt:

Since $\lim_{n \to \infty} \int f_n d \mu \to 0$ is stronger than Lebesgue's Dominated Convergence Theorem we can conclude that $\lim_{n \to \infty}f_n = f \equiv 0$.

We want to show that for every $\epsilon > 0$: $$ \lim_{n \to \infty} \mu (\{x: |f_n(x)| > \epsilon \}) = 0 $$ Is it then: $$ \lim_{n \to \infty} \mu (\{x: |f_n(x)| > \epsilon \}) = \mu (\{x: 0 > \epsilon \}) = 0 $$

If not, what is the correct proof?

$\endgroup$
3
$\begingroup$

The first sentence in your attempt has lead you to a wrong direction. Think about it: if we have $f_n\to0$, why the queation asks us to prove the much weaker convergence of $f_n$, which is the convergence in measure?

Note that $$\mu (\{x: |f_n(x)| > \epsilon \})\leq\frac{\int f_n\,d\mu}\epsilon\to0,\ \ n\to\infty.$$

The inequality I used is called Chebyshev's inequality or Markov's inequality. Thanks to @Yanko for pointing out in the comments.

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ Could you elaborate why is the inequality true? $\endgroup$ – Никита Васильев Jan 12 at 11:51
  • 2
    $\begingroup$ @math_beginner you can find it here en.wikipedia.org/wiki/Markov%27s_inequality $\endgroup$ – Yanko Jan 12 at 11:51
  • $\begingroup$ @Yanko Thank you for the link! In my country I have no access to wiki. Your link will make my answer more readable to the OP. $\endgroup$ – Feng Shao Jan 12 at 11:55
1
$\begingroup$

You already have an answer so let me just note that your conclusion that $f_n(x)\rightarrow 0$ for all $x\in X$ is wrong.

To see this, let $f_n:[-1,1]\rightarrow\mathbb{R}$ be the function whose graph is a triangle which begins at $(-\frac{1}{n^2},0)$ increase linearily to $(0,1)$ and then decrease linearily to $(0,\frac{1}{n^2})$ (Draw).

Therefore the integral of $f_n$ is the area of the triangle which is $\frac{1}{n^2}$, hence goes to zero as $n$ goes to infinitey. However $f_n(0)=1$ for all $n$ and so $f_n\not\rightarrow 0$.

| cite | improve this answer | |
$\endgroup$

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.