A variant of the prime number theorem (Q2076515): Difference between revisions

The paper under review deals about sums of the shape \[\sum_{n \leqslant x} f \left( \lfloor x/n \rfloor \right)\] which were first investigated in [the reviewer et al., J. Number Theory 202, 278--297 (2019; Zbl 1454.11009)]. Since this paper was published, a large number of researchers have improved the results in various directions. One of the most important of these is to try to break the \(\frac{1}{2}\)-barrier when \(f\) is a classical arithmetic function. In this work, the authors consider the case \(f = \Lambda\) the von Mangoldt function, a particularly interesting case in view of the very well-known equivalence \[\sum_{n \leqslant x} \Lambda (n) = x + O_\varepsilon \left( x^{1/2+\varepsilon} \right) \iff \textrm{Riemann Hypothesis}.\] In a forthcoming article to be published in the International Journal of Number Theory, the reviewer establishes among other things the estimate \[\sum_{n \leqslant x} \Lambda \left( \left \lfloor \frac{x}{n} \right \rfloor \right) = x \sum_{d=1}^\infty \frac{\Lambda(d)}{d(d+1)} + O_\varepsilon \left(x^{97/203 + \varepsilon} \right).\] The authors sharpen here the exponent \(\frac{97}{203} \approx 0.4778\) to \(\frac{9}{19} \approx 0.4736\). The main novelty of this paper is the use of a new triple exponential sum of type II, instead of a double exponential sum of type II used by the reviewer.