The conjecture for the variance of squarefull numbers in short intervals

About 3 years old · traced to

Let Q(t)Q(t) denote the counting function of squarefull numbers up to tt. Given ϵ>0\epsilon>0, X>1X>1, and Xϵ≤H≤X1/4−ϵX^\epsilon\leq H\leq X^{1/4-\epsilon}, set

y=2xH+H2.y=2\sqrt{x}H+H^2.

Variance conjecture.

1X∫X2X∣Q(x+y)−Q(x)−ζ(3/2)ζ(3)H∣2 dx∼4ζ(4/3)3ζ(2)∫0∞(sin⁡πyπy)2y1/3 dy⋅H2/3.\frac{1}{X}\int_X^{2X}\left|Q(x+y)-Q(x)-\frac{\zeta(3/2)}{\zeta(3)}H\right|^2\,dx\sim \frac{4\zeta(4/3)}{3\zeta(2)}\int_0^\infty\left(\frac{\sin \pi y}{\pi y}\right)^2y^{1/3}\,dy\cdot H^{2/3}.

This predicts the asymptotic size and leading constant of the mean-square fluctuation in the count of squarefull numbers over intervals whose endpoint difference satisfies x+y−x=H\sqrt{x+y}-\sqrt{x}=H. The preceding result gives an upper bound in a shorter-range variance problem, while this sharper asymptotic remains conjectural in the stated range.

References

Primary source

Tsz Ho Chan, “Variance of squarefull numbers in short intervals II”, arXiv:2311.13463 (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.