2
$\begingroup$

Define \begin{equation} F(\sigma) = \Re \sum_{h=1}^{\infty} \sum_{n=1}^{\infty} \frac{b_{n, h}}{n^{2\sigma}}(1+h/n)^{-\sigma}\Bigg( \frac{e^{i\log(1+h/n)}-1}{i\log(1+h/n)} \Bigg) \end{equation} where $\sigma \in \mathbb{R}$ and $|b_{n, h}|\ll \log n$. What is the minimal real number $c$ such that $F(\sigma)$ has a holomorphic continuation in $\sigma>c$ ?

Heuristically, it seems to me that $c \leq 1/2$, considering the fact that the inner sum is dominated by terms with small $h/n$, so that convergence of the whole series is assured for $2\sigma>1$.

$\endgroup$
2
$\begingroup$

Observe first that $$\lim_{x\to 0}\frac{e^{i\log(1+x)}-1}{i\log(1+x)}=1,$$ hence there exists an absolute constant $C>0$ such that $$\Re\frac{e^{i\log(1+h/n)}-1}{i\log(1+h/n)}>\frac{1}{2}\qquad\text{for}\qquad n>Ch.$$ Now assume that $b_{n,h}$ is the indicator function of $n\in(Ch,2Ch)$. Then for $\sigma\in[1,2]$ we have $$F(\sigma)>\frac{1}{2}\left(1+\frac{1}{C}\right)^{-2}\sum_{h=1}^\infty\,\sum_{n\in(Ch,2Ch)}\frac{1}{n^{2\sigma}}\gg\sum_{h=1}^\infty\frac{1}{h^{2\sigma-1}}.$$ It follows that $F(1)$ diverges, and $F(\sigma)\to\infty$ under $\sigma\to 1+$. The second statement shows that $F(\sigma)$ has no continuous extension to $\sigma\geq 1$.

$\endgroup$
1
  • 1
    $\begingroup$ Thanks ! I i guess will be interesting to see what happends modulo $\sigma=1$...That is, whether or not $\sigma=1$ is an isolated singulaity... $\endgroup$ – Pres10 Aug 9 '20 at 7:28

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.