0
$\begingroup$

Hi in the proof of f integrable, $g(x)= \int_{[0,x]} f dt$ is uniformly continuous, i saw a proof that uses the following argument. Let $\epsilon > 0$ be given, take $δ := \epsilon^-1 · (\int |f|dt) >0$ since f is integrable. Let $x

My question is, how do we get that $\int_{[x,y]} |f(t)|dt ≤ |x − y|\int_{R} |f(t)|dt$? I have tried to bound it and only able to conclude that $\int_{[x,y]} |f(t)|dt\leq |x − y| sup(|f|)$ ?

Thanks in advance

  • 2
    The proposed inequality is false. That would imply that $g$ is Lipschitz on $\Bbb{R}$, which then implies that $f$ is bounded. Of course, there are integrable functions which are unbounded on any open sets. This necessitates the control over such exceptional 'peaks'.2017-02-08
  • 0
    I see, so if f is bounded and integrable, then we can do this?2017-02-08
  • 0
    @Socchi: No, even then it is false. Take $f = 2 \cdot 1_{[0,1/2]}$.2017-02-08

1 Answers 1

0

You can't, the proposed proof is wrong. A simple counterexample is $f(x) = |x|^{-1/2}$ for $-1 \le x \le 1$, $f(x)=0$ otherwise. Note $g(x) = 2 \sqrt{|x|}$ for $-1 \le x \le 1$ which is not Lipschitz.

  • 0
    But your f(x) is not integrable?2017-02-08
  • 0
    I was worried about the region, thanks. I wonder whether a part of the proof is missing?2017-02-08
  • 0
    Oh, at infinity, you're right. Fixed. What's relevant is what happens near 0.2017-02-08
  • 0
    Thanks, I guess that claim is wrong? Maybe some part of the proof is missing to get to that conclusion?2017-02-08
  • 0
    I think it's just wrong. I can't think of any useful class of functions $f$ for which it is true.2017-02-08