0
$\begingroup$

Let $$B(s)=\sum_{n\le N}b(n)n^{-s}$$ with $b(n)\ll n^{\varepsilon}.$ I want to study $$\frac{1}{2\pi i}\int_C\frac{\zeta'}{\zeta}(s)\zeta'(s)\zeta'(1-s)B(s)B(1-s)\;ds,$$ with $C$ being the contour with vertices $1+1/\log T+i, 1+1/\log T+iT, -1/\log T+iT, -1/\log T+i.$ Something of this form was studied in the 90s (see $S_2$ here) but they formulate the polynomial $B$ slightly differently, and I want to find an asymptotic which can later be specialised to the 'correct' answer simply by substituting in the choices of coefficients $b(k)$.

The reason I ask this is if, for example, I wanted $B(s)$ to resemble something like $\zeta'(s)^2$ I would choose $b(n)=(\log*\log)(n)$ (and then $B(s)$ would just be the truncated Dirichlet series of $\zeta'(s)^2$), but it doesn't seem that one can make such a choice in the paper? Any insights?

$\endgroup$
1
  • $\begingroup$ What exactly are you asking? The question isn't clear to me. $\endgroup$ Commented Oct 7 at 18:31

0

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.