arXiv subjects
Publications and source records attributed to Sary Drappeau.
We show that smooth-supported multiplicative functions $f$ are well-distributed in arithmetic progressions $a_1a_2^{-1} \pmod q$ on average over moduli $q\leq x^{3/5-\varepsilon}$ with $(q,a_1a_2)=1$.
We prove a bound for quintilinear sums of Kloosterman sums, with congruence conditions on the "smooth" summation variables. This generalizes classical work of Deshouillers and Iwaniec, and is key to obtaining power-saving error terms in applications, notably the dispersion method. As a consequence, assuming the Riemann hypothesis for Dirichlet $L$-functions, we prove a power-saving error term in the Titchmarsh divisor problem of estimating $\sum_{p\leq x}τ(p-1)$. Unconditionally, we isolate the possible contribution of Siegel zeroes, showing it is always negative. Extending work of Fouvry and Tenenbaum, we obtain power-saving in the asymptotic formula for $\sum_{n\leq x}τ_k(n)τ(n+1)$, reproving a result announced by Bykovskiĭ and Vinogradov by a different method. The gain in the exponent is shown to be independent of $k$ if a generalized Lindelöf hypothesis is assumed.
According to a general probabilistic principle, the natural divisors of friable integers (i.e.~free of large prime factors) should normally present a Gaussian distribution. We show that this indeed is the case with conditional density tending to 1 as soon as the standard necessary conditions are met. Furthermore, we provide explicit, essentially optimal estimates for the decay of the involved error terms. The size of the exceptional set is sufficiently small to enable recovery of the average behaviour in the same optimal range. Our argument combines the saddle-point method with new large deviations estimates for the distribution of certain additive functions.
In this paper we study Weyl sums over friable integers (more precisely $y$-friable integers up to $x$ when $y = (\log x)^C$ for a large constant $C$). In particular, we obtain an asymptotic formula for such Weyl sums in major arcs, nontrivial upper bounds for them in minor arcs, and moreover a mean value estimate for friable Weyl sums with exponent essentially the same as in the classical case. As an application, we study Waring's problem with friable numbers, with the number of summands essentially the same as in the classical case.
A number is said to be $y$-friable if it has no prime factor greater than $y$. In this paper, we prove a central limit theorem on average for the distribution of divisors of $y$-friable numbers less than $x$, for all $(x, y)$ satisfying $2\leq y \leq {\rm e}^{(\log x)/(\log\log x)^{1+\varepsilon}}$. This was previously known under the additional constraint $y\geq {\rm e}^{(\log\log x)^{5/3+\varepsilon}}$, by work of Basquin. Our argument relies on the two-variable saddle-point method.
An integer n is said to be y-friable if its largest prime factor $P^+(n)$ is less than y. In this paper, it is shown that the y-friable integers less than x have a weak exponent of distribution at least $3/5-\varepsilon$ when $(\log x)^c\leq y\leq x^{1/c}$ for some $c=c(\varepsilon)\geq 1$, that is, they are well distributed in the residue classes of a fixed integer $a$, on average over moduli $\leq x^{3/5-\varepsilon}$ for each fixed $a\neq 0$ and $\varepsilon>0$. We present an application to the estimation of the sum $\sum_{2\leq n\leq x, P^+(n)\leq y}τ(n-1)$ when $(\log x)^c\leq y$. This follows and improves on previous work of Fouvry and Tenenbaum. Our proof combines the dispersion method of Linnik in the setting of Bombieri, Fouvry, Friedlander and Iwaniec, and recent work of Harper on friable integers in arithmetic progressions.
An integer is said to be $y$-friable if its greatest prime factor P(n) is less than $y$. In this paper, we study numbers of the shape $n-1$ when $P(n)\leq y$ and $n\leq x$. One expects that, statistically, their multiplicative behaviour resembles that of all integers less than $x$. Extending a result of Basquin, we estimate the mean value over shifted friable numbers of certain arithmetic functions when $(\log x)^c \leq y$ for some positive $c$, showing a change in behaviour according to whether $\log y / \log\log x$ tends to infinity or not. In the same range in $(x, y)$, we prove an Erdös-Kac-type theorem for shifted friable numbers, improving a result of Fouvry and Tenenbaum. The results presented here are obtained using recent work of Harper on the statistical distribution of friable numbers in arithmetic progressions.
An integer is said to be $y$-friable if its greatest prime factor is less than $y$. In this paper, we obtain estimates for exponential sums over $y$-friable numbers up to $x$ which are non-trivial when $y \geq \exp\{c \sqrt{\log x} \log \log x\}$. As a consequence, we obtain an asymptotic formula for the number of $y$-friable solutions to the equation $a+b=c$ which is valid unconditionnally under the same assumption. We use a contour integration argument based on the saddle point method, as developped in the context of friable numbers by Hildebrand & Tenenbaum, and used by Lagarias, Soundararajan and Harper to study exponential and character sums over friable numbers.
Dans un récent article, Lagarias et Soundararajan étudient les solutions friables à l'équation a+b=c. Sous l'hypothèse de Riemann généralisées aux fonctions L de Dirichlet, ils obtiennent une estimation pour le nombre de solutions pondérées par un poids lisse et une minoration pour le nombre de solutions non pondérées. Le but de cet article est de présenter des arguments qui permettent d'une part de préciser les termes d'erreur et d'étendre les domaines de validité de ces estimations afin de faire le lien avec un travail de la Bretèche et Granville, d'autre part d'obtenir le comportement asymptotique exact du nombre de solutions non pondérées. In a recent paper, Lagarias and Soundararajan study the y-smooth solutions to the equation a+b=c. Under the Generalised Riemann Hypothesis, they obtain an estimate for the number of those solutions weighted by a compactly supported smooth function, as well as a lower bound for the number of bounded unweighted solutions. In this paper, we aim to prove a more precise estimate for the number of weighted solutions that is valid when y is relatively large with respect to x, so as to connect our estimate with the one obtained by La Bretèche and Granville in a recent work. We also prove the conjectured upper bound for the number of bounded unweighted solutions, thus obtaining its exact asymptotic behaviour.