Which of the following condition below imply that the $f:[0,1]\to\mathbb{R}$ is necessarily Bounded Variation?

  1. monotone;
  2. continuous and monotone;
  3. has derivative on $(0,1)$;
  4. bounded derivative on $(0,1)$.

I studied one result long ago that a Monotone function is of bounded variation, and The function $\cos \frac{\pi}{x}$ is continous but not a function of founded variation I can not recall the proof. for $4$ by MVT we get $|f(x)-f(y)|<K|x-y|$ where $|f'(x)|<K$ Then $\sum_{i=1}^{n}|f(x_i)-f(x_{i-1})|$ is definitely bounded and so $4$ is also true, could any one tell me what else can be said?

  • $\begingroup$ If $f$ is monotone, how can you write $|f(x_i)-f(x_{i-1})|$ differently? - For 3.: what about the function you remembered was not of bounded variation? $\endgroup$ – Thomas Apr 18 '13 at 8:44
  1. implies BV because due to monotonicity the sum $\sum |f(x_i)-f(x_{i-1})|$ becomes either $\sum f(x_i)-f(x_{i-1})$ or $-\sum f(x_i)-f(x_{i-1})$ and subsequently telescopes.

  2. See 1.

  3. Indeed, $\cos(\pi/x)$ (defined arbitrarily at $0$) is differentiable on $(0,1)$ and does not have bounded variation, as you can check by considering partition points $\{x_1,\dots,x_n\} = \{n^{-1},(n-1)^{-1},\dots, 1/2\}$ where $n$ can be as large as you want.

  4. implies $f$ is Lipschitz, and a Lipschitz function is BV because $\sum |f(x_i)-f(x_{i-1})|\le L\sum (x_i-x_{i-1}) $ where the latter sum telescopes.


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.