Granville–Soundararajan conjecture on character sums and smooth numbers

Let ρ(u)\rho(u) denote the Dickman function, let 1y-smooth(n)\boldsymbol{1}_{y\text{-smooth}}(n) indicate that every prime factor of nn is at most yy, and define

otagΨ(x,y;χ)=nxn is y-smoothχ(n),Ψ(x,y;χ0)=nxn is y-smoothχ0(n).otag \textstyle \Psi(x,y;\chi)=\sum_{\substack{n\leq x\\ n\text{ is }y\text{-smooth}}}\chi(n),\qquad \Psi(x,y;\chi_0)=\sum_{\substack{n\leq x\\ n\text{ is }y\text{-smooth}}}\chi_0(n).

Here χ\chi is a non-principal character modulo qq, and χ0\chi_0 is the principal character modulo qq.

Granville–Soundararajan conjecture. There exists a constant A>0A>0 such that, for every non-principal character χ\chi modulo qq and every 1xq1\leq x\leq q, uniformly,

nxχ(n)=Ψ(x,y;χ)+o(Ψ(x,y;χ0)),\sum_{n\leq x}\chi(n)=\Psi(x,y;\chi)+o\bigl(\Psi(x,y;\chi_0)\bigr),

where

y=(logq+log2x)(log2q)A.y=(\log q+\log^2 x)(\log_2 q)^A.

This conjecture asserts that character sums are asymptotically governed by their contributions from smooth numbers, strengthening the conditional estimates discussed immediately beforehand; its resolution status is not specified in the supplied text.

Sources & referencesView supporting material

Primary source

Zikang Dong, Yutong Song, Weijia Wang and Hao Zhang, “On derivatives of zeta and L-functions near the 1-line”, arXiv:2312.12199 (2023).

Additional references

4 papers in this index state this conjecture (2011–2023). The statement above is taken from the most recent of them; the others are arXiv:2204.13826, arXiv:1503.07196, arXiv:1109.1786.

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.