Let $d(n)$ denote the number of positive divisors of the positive integer $n$. Pick some positive $X,h \in \mathbb{R}$ and consider the sum

$$ S(X,h) := \sum_{X \leq n \leq X+ h} d(n).$$

In view of Huxley's results, we can give the correct asymptotic estimate (as $X \to \infty$) for $S(X,h)$ once $h \gg X^{0.315}$. If I recall correctly, Shiu proved the correct upper bound for $S(x,h)$ even when $h \sim x^\epsilon$.

For small $h$, say $h \sim X^{0.2}$, can we obtain the correct asymptotic for $S(X,h)$?

It seems to me that the methods (Poisson summation, etc.) used to estimate the full sum should be applicable here, so I would like to know whether anyone has already carried this out carefully.

I am less interested in bounds that are valid for "most" values of $X$.


Perhaps it will be useful. https://core.ac.uk/download/pdf/11423881.pdf I did not go into details, my hypothesis is if we take a constant $c< \frac{131}{416}$, then the weakened condition in this article will still be preserved, and moreover $k$ can be limited, i.e. condition $\int_{X}^{X+Y}(\Delta (x+h(x))-\Delta (x))^{k}dx$ is valid for $k< C $ ($C$ - some constant).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.