0
$\begingroup$

I know that convergence almost everywhere implies convergences in measure(sometime it said locally). I also know that convergence in measure implies the existence of subsequence which is convergent almost everywhere.

However I wonder if there are some additional conditions from which the convergence in measure will imply convergence almost everywhere (for a whole sequence, not just subsequence).

Please don't mention discreteness. Let's just take real line for example, and a sequence of functions on it.

Thank you, for any help.

$\endgroup$
  • $\begingroup$ monotone convergence ? $\endgroup$ – Hua Feb 20 '17 at 6:05
  • $\begingroup$ Convergence a.e do not implies in general convergence in measure. This happens, for example, when the space is finite. And I never seen such sufficient condition for what you want. What you purpose with this? $\endgroup$ – Filburt Feb 20 '17 at 13:36
2
$\begingroup$

If the convergence in measure is rapid enough, then a.e. convergence will follow. Suppose $(f_n)$ converges in measure on the measure space $(E,\mathcal E,\mu)$. If $\mu(\{x\in E: |f_n(x)-f(x)|>\epsilon\})$ converges to $0$ so rapidly that $\sum_n\mu(\{x\in E: |f_n(x)-f(x)|>\epsilon\})<\infty$ for each $\epsilon>0$ then $\mu(\{x\in E: \limsup_n|f_n(x)-f(x)|>\epsilon\})=0$ for each $\epsilon>0$ (Borel-Cantelli lemma), so $f_n\to f$ $\mu$-a.e.

$\endgroup$

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.