1

Combining the identity $(1)$ from [1], I am saying the specialization $a_n=\sqrt{2n^3}$ and $b_n=i\sqrt{n^3}$ (here $i$ denotes the imaginary unit, thus $i^2=-1$) for integers $n\geq 1$, and the explanation of the criterion of Fubini's theorem, see [2] if you need it, I can prove that $$\frac{1}{\zeta(3)}=\int_0^\infty\sum_{n=1}^\infty\frac{\mu(n)\sin(\sqrt{2n^3}x)}{\sqrt{2n^3}}e^{-i\sqrt{n^3}x}dx,\tag{A}$$ where $\mu(n)$ is the Möbius function.

Question. I was wondering about questions involving this function $$f(x):=\sum_{n=1}^\infty\frac{\mu(n)\sin(\sqrt{2n^3}x)}{\sqrt{2n^3}}e^{-i\sqrt{n^3}x}$$ defined on $[0,\infty)$ that I know how solve or I don't know how solve those.

I know that the function $f(x)$ is uniformly and absolutely convergent on $[0,\infty)$, but what about the differentiability? I am asking what relevant facts we can deduce about the differentiability of our function $f(x)$ for real numbers $0\leq x<\infty$.

Many thanks.

Feel free, if you prefer, add hints for some of previous question, instead of a full answer.

References, both from this Mathematics Stack Exchange:

[1] See the answer by D'Aurizio for Cantarini's lemma, identity $(1)$ from: Find the closed form for $\int_{0}^{\infty}\cos{x}\ln\left({1+e^{-x}\over 1-e^{-x}}\right)dx=\sum_{n=0}^{\infty}{1\over n^2+(n+1)^2}$.

[2] See the second paragraph of the answer by Eldredge: When can a sum and integral be interchanged?

  • 1
    Leibnez's Integral Rule for differentiability of $g(x)$. Take your integral over measure $\mathbb{N}$ – Brevan Ellefsen May 17 '17 at 19:55
  • Thanks for your hint, it is appreciated @BrevanEllefsen I try think about your hint. Perfect and thanks one more time. –  May 17 '17 at 19:57
  • 1
    No problem. For a less involved method, see this post – Brevan Ellefsen May 17 '17 at 19:59
  • I would recommend splitting your question here... It's generally not good policy to have two questions on one post, especially when they are only somewhat related questions. Leave this question on the differentiability of $g(x) $ and make your limit a separate question. You'll be more likely to get answers that way! (and it means double the reputation, and thats always nice) – Brevan Ellefsen May 18 '17 at 15:35
  • You are right. I do it in next few hours. Many thanks @BrevanEllefsen –  May 18 '17 at 15:42

0 Answers0