Smooth-cutoff variance conjecture for the k-fold divisor function in arithmetic progressions

About 4 years old · traced to

Fix k≥2k\ge2. Let w(y)w(y) be a smooth function supported in [1,2][1,2] such that

∫w(y)2 dy=1,\int w(y)^2\,dy=1,

and whose Mellin transform M[w]\mathcal{M}[w] satisfies

M[w](σ+it)≪ℓ11+∣t∣ℓ\mathcal{M}[w](\sigma+it)\ll_\ell\frac{1}{1+|t|^\ell}

uniformly for ∣σ∣≤A|\sigma|\le A, for every fixed A>0A>0 and every positive integer ℓ\ell. Let Δw(τk;X,d,a)\Delta_w(\tau_k;X,d,a) denote the smoothed divisor-function error term in an arithmetic progression, and let ak(d)a_k(d) and γk(c)\gamma_k(c) be the arithmetic constant and piecewise polynomial defined above. Smooth-cutoff variance conjecture. If X,d→∞X,d\to\infty with log⁡X/log⁡d→c∈(0,k)\log X/\log d\to c\in(0,k), then

∑1≤a≤d(a,d)=1∣Δw(τk;X,d,a)∣2∼ak(d)γk(c)X(log⁡d)k2−1.\sum_{\substack{1\le a\le d\\ (a,d)=1}}\left|\Delta_w(\tau_k;X,d,a)\right|^2\sim a_k(d)\gamma_k(c)X(\log d)^{k^2-1}.

This conjecture extends the known smoothed variance asymptotic in the range handled by the paper to all c∈(0,k)c\in(0,k). The source indicates that the formula is known in the range 0<c<10<c<1 through earlier work, while the full range remains open; endpoint and off-diagonal difficulties obstruct the current method.

References

Primary source

David T. Nguyen, “Variance of the k-fold divisor function in arithmetic progressions for individual modulus”, arXiv:2205.02354 (2023).

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.