The CFZ–CS ratios conjecture for averages of zeta-function ratios

Let AA and BB be sets of complex numbers with real parts smaller than 1/41/4, and let CC and DD be sets of complex numbers with positive real parts smaller than 1/41/4. Set s=1/2+its=1/2+it and define

RA,B,C,D(T):=0ψ(tT)αAζ(s+α)βBζ(1s+β)γCζ(s+γ)δDζ(1s+δ)dt.\mathcal R_{A,B,C,D}(T):=\int_0^\infty \psi\left(\frac t T\right) \frac{\prod_{\alpha\in A}\zeta(s+\alpha)\prod_{\beta\in B}\zeta(1-s+\beta)}{\prod_{\gamma\in C}\zeta(s+\gamma)\prod_{\delta\in D}\zeta(1-s+\delta)}\,dt.

Let BA,B,C,D(s):=n=1IA,C(n)IB,D(n)ns\mathcal B_{A,B,C,D}(s):=\sum_{n=1}^\infty I_{A,C}(n)I_{B,D}(n)n^{-s}, where IA,CI_{A,C} is defined by

n=1IA,C(n)ns=αAζ(s+α)γCζ(s+γ).\sum_{n=1}^\infty \frac{I_{A,C}(n)}{n^s}=\frac{\prod_{\alpha\in A}\zeta(s+\alpha)}{\prod_{\gamma\in C}\zeta(s+\gamma)}.

The CFZ–CS ratios conjecture. Suppose the sets A,B,C,DA,B,C,D are as above and the imaginary parts of all parameters are O(T1ξ)O(T^{1-\xi}) for some ξ>0\xi>0. Then, for some η>0\eta>0,

RA,B,C,D(T)=0ψ(tT)UA,VBU=V(t2π)α^Uα^β^Vβ^BAU+V,BV+U,C,D(1)dt+O(T1η).\mathcal R_{A,B,C,D}(T)=\int_0^\infty \psi\left(\frac t T\right)\sum_{U\subset A,\,V\subset B\atop |U|=|V|}\left(\frac{t}{2\pi}\right)^{-\sum_{\hat\alpha\in U}\hat\alpha-\sum_{\hat\beta\in V}\hat\beta}\mathcal B_{A-U+V^-,B-V+U^-,C,D}(1)\,dt+O(T^{1-\eta}).

This is the ratios conjecture originally formulated by Conrey, Farmer and Zirnbauer and studied by Conrey and Snaith; the paper relates it to arithmetic correlations and proves that both conjectures imply the same result. Its general status is not established.

Sources & referencesView supporting material

Primary source

Brian Conrey and Jonathan P. Keating, “Averages of ratios of the Riemann zeta-function and correlations of divisor sums”, arXiv:1611.09198 (2016).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.