10

As we know, the Weierstrass function defined by $$ f(x) = \sum_{n=0}^{\infty} a^n \cos(b^n \pi x), 0<a<1,b \in 2\mathbb N+1,ab > 1+\frac{3\pi}{2} $$ is differentiable nowhere (see here for a proof of the same). As a consequence, $f$ is not of bounded variation on any finite interval, ($f$ is said to be of bounded variation on $[a,b]$ if for any partition $a=x_0<x_1<x_2<\cdots<x_{n-1}<x_n=b$, $\sum_{k=1}^n|f(x_k)-f(x_{k-1})|\leq M$ for a constant $M$ independent of the choice of partition) since functions of bounded variation are differentiable a.e. (as a corollary of this and the Lebesgue theorem). In particular, the proof that $f$ is not of bounded variation is not shown explicitly via the definition of functions of bounded variation.

Question: Can we construct a sequence of partitions on any interval $[a,b]$ for which the sequence of sums $\sum_{k=1}^n|f(x_k)-f(x_{k-1})|$ increases to positive infinite, showing that $f$ is not of bounded variation on $[a,b]$?

Arctic Char
  • 16,007
brushmonk
  • 144
  • A function of bounded variation is differentiable a.e :) – Sourav Ghosh Apr 08 '23 at 16:42
  • 5
    @SouravGhosh The OP probably knows, but clearly asked for a direct proof that Weierstrass function is not of bounded variation. – Anne Bauval Apr 08 '23 at 17:47
  • 1
    I added some links to reflect the definitions and results used to motivate the above question. In particular, the main reason, namely that the "standard" proof that $f$ is not BV is not explicit enough, has been made much clearer than previous, without adding any further information. – Sarvesh Ravichandran Iyer May 15 '23 at 10:04
  • 4
    This is clearly an interesting and difficult analysis problem that the Op didn't really know how to start approaching themselves. In situations such as this it is of benefit to the community to not close such questions, I feel; after all, we care about building a repository of knowledge and examples of techniques. Note a question like this would probably be well-received on MO, so why not on MSE? – FShrike May 15 '23 at 10:06
  • @FShrike The close votes were all casted before the edit. If you read the original post, it has a wrong/unclear definition of bounded variations, and from the wording, it's not clear if the OP know or not know that the function is of bounded variation. I think the closure is suitable. – Arctic Char May 15 '23 at 12:56
  • @ArcticChar That’s fair. The initial revision was low quality – FShrike May 15 '23 at 13:10

0 Answers0