1

There was a similar question on this matter, but "continuous differentiable" is too strong an assumption, so I was wondering if this can be relaxed to "differentiable".

Ryan
  • 303

1 Answers1

4

I got an answer referencing Rudin'n Real and Complex Analysis, Thm 7.21, which the author deleted later because of a mistake. I am grateful that he led me to this theorem.

Assume $f\in BV([0,1])$, then $f=f_1-f_2$, where $f_i$ are bounded increasing, $f'_i$ exists a.e., and $f'=f'_1-f'_2$ a.e.. Since $\int_0^1 f'_i\leq f_i(1)-f_i(0)$, we know $f'_i\in L^1([0,1])$ and thus $f\in L^1([0,1])$. Then we can apply Rudin 7.21 (i.e. If $f$ is differentiable, $f'\in L^1([0,1])$, then $f$ is absolutely continuous)

Mittens
  • 46,352
Ryan
  • 303