arXiv subjects

Dimitris Koukoulopoulos

Publications and source records attributed to Dimitris Koukoulopoulos.

At least 19 recordsLinked to original sources

arXiv2026

A logarithmic structure theorem for multiplicative functions with small partial sums

Let $D\in\mathbb{N}$, let $A>D+1$, and let $Q\geqslant3$. Consider the class of multiplicative functions $f:\mathbb{N}\to\mathbb{C}$ such that $|\sum_{n\leqslant x}f(n)|\le x(\log Q)^{A-D-1}/(\log x)^A$ for all $x\geqslant Q$, and such that $|\Lambda_f|\leqslant D\Lambda$, where $\Lambda_f$ is defined via the Dirichlet convolution identity $f\log=\Lambda_f*f$ and $\Lambda$ denotes von Mangoldt's function. We prove there exist parameters $m\in\{0,1,\dots,D\}$ and $Q=Q_D\leqslant Q_{D-1}\le \cdots\leqslant Q_m<Q_{m+1}=\infty$ such that $\sum_{p\in I} \mathrm{Re}(f(p)+j)/p=O_{A,D}(1)$ for all $j=m,m+1,\dots,D$ and all compact intervals $I\subset[Q_{j+1},Q_j)$. Moreover, when $|\sum_{n\leqslant x}f(n)|\le x^{1-1/\log Q}/(\log x)^{D+1}$ for all $x\geqslant Q$, we relate the parameters $m$ and $Q_j$ to the location of zeroes of the Dirichlet series $\sum_{n\geqslant1} f(n)/n^s$ in the ball $B(1,1/\log Q)$. These results generalize work of the author when $D=1$. Their proof builds on earlier work of the author with Soundararajan, and of Sachpazis.

math.NT↗

arXiv2025

Irreducibility and Galois groups of random reciprocal polynomials of large degree

Let $A = a_0T^m + \sum_{j=1}^{m-1} a_j (T^{m-j}+T^{m+j}) + T^{2m}+1 \in \mathbf{Z}[T]$ be a monic reciprocal polynomial of degree $2m$ sampled randomly by selecting its coefficients $a_0,a_1,\dots,a_{m-1}$ independently according to a given probability measure $\mu$ on $\mathbf{Z}$. For a wide range of measures $\mu$, we prove that $A$ is irreducible with probability $\ge 1-Cm^{-c}$ for some absolute constants $c,C>0$. In addition, we prove that with the same probability the Galois group of $A$ is either the full hyperoctahedral group $\mathcal{C}_2 \wr \mathcal{S}_m$ or one of two of its index-$2$ subgroups. The main condition that $\mu$ must satisfy is of Fourier-theoretic nature, and holds for example when $\mu$ is the uniform measure on a set of at least $35$ consecutive integers, or on an arbitrary, sufficiently large subset of an interval $[-H,H]$, with $H$ larger than some absolute constant. Our most general result allows for each $a_j$ to be sampled by its own probability measure $\mu_j$. Our approach builds on earlier work of Bary-Soroker, Kozma and the second author, who proved for essentially the same $\mu_j$ that the 'standard' monic polynomial $a_0 + \cdots + a_{m-1}T^{m-1} + T^m$ is irreducible and has as Galois group either the symmetric group $\mathcal{S}_m$ or the alternating group $\mathcal{A}_m$ with high probability, conditioning on $a_0 \neq 0$. In our setting of reciprocal polynomials, we can rule out (all subgroups of) the maximal alternating subgroup $(\mathcal{C}_2 \wr \mathcal{S}_m) \cap \mathcal{A}_{2m}$ of the hyperoctahedral group as likely Galois group of $A$ by analyzing its discriminant.

math.NT↗

arXiv2025

Erd\H{o}s's integer dilation approximation problem and GCD graphs

Let $\mathcal{A}\subset\mathbb{R}_{\geqslant1}$ be a countable set such that $\limsup_{x\to\infty}\frac{1}{\log x}\sum_{\alpha\in\mathcal{A}\cap[1,x]}\frac{1}{\alpha}>0$. We prove that, for every $\varepsilon>0$, there exist infinitely many pairs $(\alpha, \beta)\in \mathcal{A}^2$ such that $\alpha\neq \beta$ and $|n\alpha-\beta| <\varepsilon$ for some positive integer $n$. This resolves a problem of Erd\H{o}s from 1948. A critical role in the proof is played by the machinery of GCD graphs, which were introduced by the first author and by James Maynard in their work on the Duffin--Schaeffer conjecture in Diophantine approximation.

math.NT↗

arXiv2024

On Arratia's coupling and the Dirichlet law for the factors of a random integer

Let $x \ge 2$, let $N_x$ be an integer chosen uniformly at random from the set $\mathbb Z \cap [1, x]$, and let $(V_1, V_2, \ldots)$ be a Poisson--Dirichlet process of parameter $1$. We prove that there exists a coupling of these two random objects such that $$ \mathbb E \, \sum_{i \ge 1} |\log P_i- V_i\log x| \asymp 1, $$ where the implied constants are absolute and $N_x = P_1P_2 \cdots$ is the unique factorization of $N_x$ into primes or ones with the $P_i$'s being non-increasing. This establishes a 2002 conjecture of Arratia arXiv:1305.0941 who constructed a coupling for which the left-hand side in the above estimate is $\ll \log\!\log x$, and who also proved that the left-hand side is $\ge 1-o(1)$ for all couplings. In addition, we use our refined coupling to give a probabilistic proof of the Dirichlet law for the average distribution of the integer factorization into $k$ parts proved in 2023 by Leung arXiv:2206.14728 and we improve on its error term.

math.NT↗

arXiv2024

An almost sharp quantitative version of the Duffin-Schaeffer conjecture

We prove a quantitative version of the Duffin-Schaeffer conjecture with an almost sharp error term. Precisely, let $\psi:\mathbb{N}\to[0,1/2]$ be a function such that the series $\sum_{q=1}^\infty \varphi(q)\psi(q)/q$ diverges. In addition, given $\alpha\in\mathbb{R}$ and $Q\geqslant1$, let $N(\alpha;Q)$ be the number of coprime pairs $(a,q)\in\mathbb{Z}\times\mathbb{N}$ with $q\leqslant Q$ and $|\alpha-a/q|<\psi(q)/q$. Lastly, let $\Psi(Q)=\sum_{q\leqslant Q}2\varphi(q)\psi(q)/q$, which is the expected value of $N(\alpha;Q)$ when $\alpha$ is uniformly chosen from $[0, 1]$. We prove that $N(\alpha;Q)=\Psi(Q)+O_{\alpha,\varepsilon}(\Psi(Q)^{1/2+\varepsilon})$ for almost all $\alpha$ (in the Lebesgue sense) and for every fixed $\varepsilon>0$. This improves upon results of Koukoulopoulos-Maynard and of Aistleitner-Borda-Hauke.

math.NT↗

arXiv2023

A lower bound on the mean value of the Erd\H{o}s-Hooley Delta function

We give an improved lower bound for the average of the Erd\H{o}s-Hooley function $\Delta(n)$, namely $\sum_{n\le x} \Delta(n) \gg_\varepsilon x(\log\log x)^{1+\eta-\varepsilon}$ for all $x\geqslant100$ and any fixed $\varepsilon$, where $\eta = 0.3533227\dots$ is an exponent previously appearing in work of Green and the first two authors. This improves on a previous lower bound of $\gg x \log\log x$ of Hall and Tenenbaum, and can be compared to the recent upper bound of $x (\log\log x)^{11/4}$ of the second and third authors.

math.NT↗

arXiv2023

An upper bound on the mean value of the Erd\H{o}s-Hooley Delta function

The Erd\H{o}s-Hooley Delta function is defined for $n\in\mathbb{N}$ as $\Delta(n)=\sup_{u\in\mathbb{R}} \#\{d|n : e^u<d\le e^{u+1}\}$. We prove that $\sum_{n\le x} \Delta(n) \ll x(\log\log x)^{11/4}$ for all $x\ge100$. This improves on earlier work of Hooley, Hall--Tenenbaum and La Bret\`eche-Tenenbaum.

math.NT↗

arXiv2022

On the $j$-th smallest modulus of a covering system with distinct moduli

Covering systems were introduced by Erd\H{o}s in 1950. In the same article where he introduced them, he asked if the minimum modulus of a covering system with distinct moduli is bounded. In 2015, Hough answered affirmatively this long standing question. In 2022, Balister, Bollob\'as, Morris, Sahasrabudhe and Tiba gave a simpler and more versatile proof of Hough's result. Building upon their work, we show that there exists some absolute constant $c>0$ such that the $j$-th smallest modulus of a minimal covering system with distinct moduli is $\le \exp(cj^2/\log(j+1))$.

math.NT↗

arXiv2022

Rational approximations of irrational numbers

Given quantities $Δ_1,Δ_2,\dots\geqslant 0$, a fundamental problem in Diophantine approximation is to understand which irrational numbers $x$ have infinitely many reduced rational approximations $a/q$ such that $|x-a/q|<Δ_q$. Depending on the choice of $Δ_q$ and of $x$, this question may be very hard. However, Duffin and Schaeffer conjectured in 1941 that if we assume a "metric" point of view, the question is governed by a simple zero--one law: writing $φ$ for Euler's totient function, we either have $\sum_{q=1}^\infty φ(q)Δ_q=\infty$ and then almost all irrational numbers (in the Lebesgue sense) are approximable, or $\sum_{q=1}^\inftyφ(q)Δ_q<\infty$ and almost no irrationals are approximable. We present the history of the Duffin--Schaeffer conjecture and the main ideas behind the recent work of Koukoulopoulos--Maynard that settled it.

math.NT↗

arXiv2020

Irreducibility of random polynomials: general measures

Let $\mu$ be a probability measure on $\mathbb{Z}$ that is not a Dirac mass and that has finite support. We prove that if the coefficients of a monic polynomial $f(x)\in\mathbb{Z}[x]$ of degree $n$ are chosen independently at random according to $\mu$ while ensuring that $f(0)\neq0$, then there is a positive constant $\theta=\theta(\mu)$ such that $f(x)$ has no divisors of degree $\le \theta n$ with probability that tends to 1 as $n\to\infty$. Furthermore, in certain cases, we show that a random polynomial $f(x)$ with $f(0)\neq0$ is irreducible with probability tending to 1 as $n\to\infty$. In particular, this is the case if $\mu$ is the uniform measure on a set of at least 35 consecutive integers, or on a subset of $[-H,H]\cap\mathbb{Z}$ of cardinality $\ge H^{4/5}(\log H)^2$ with $H$ sufficiently large. In addition, in all of these settings, we show that the Galois group of $f(x)$ is either $\mathcal{A}_n$ or $\mathcal{S}_n$ with high probability. Finally, when $\mu$ is the uniform measure on a finite arithmetic progression of at least two elements, we prove a random polynomial $f(x)$ as above is irreducible with probability $\ge\delta$ for some constant $\delta=\delta(\mu)>0$. In fact, if the arithmetic progression has step 1, we prove the stronger result that the Galois group of $f(x)$ is $\mathcal{A}_n$ or $\mathcal{S}_n$ with probability $\ge\delta$.

math.NT↗

arXiv2020

A note on the natural density of product sets

Given two sets of natural numbers $\mathcal{A}$ and $\mathcal{B}$ of natural density $1$ we prove that their product set $\mathcal{A}\cdot \mathcal{B}:=\{ab:a\in\mathcal{A},\,b\in\mathcal{B}\}$ also has natural density $1$. On the other hand, for any $\varepsilon>0$, we show there are sets $\mathcal{A}$ of density $>1-\varepsilon$ for which the product set $\mathcal{A}\cdot\mathcal{A}$ has density $<\varepsilon$. This answers two questions of Hegyvári, Hennecart and Pach.

math.NT↗

arXiv2020

The structure of multiplicative functions with small partial sums

The Landau-Selberg-Delange method provides an asymptotic formula for the partial sums of a multiplicative function whose average value on primes is a fixed complex number $v$. The shape of this asymptotic implies that $f$ can get very small on average only if $v=0,-1,-2,\dots$. Moreover, if $v<0$, then the Dirichlet series associated to $f$ must have a zero of multiplicity $-v$ at $s=1$. In this paper, we prove a converse result that shows that if $f$ is a multiplicative function that is bounded by a suitable divisor function, and $f$ has very small partial sums, then there must be finitely many real numbers $γ_1$, $\dots$, $γ_m$ such that $f(p)\approx -p^{iγ_1}-\cdots-p^{-iγ_m}$ on average. The numbers $γ_j$ correspond to ordinates of zeroes of the Dirichlet series associated to $f$, counted with multiplicity. This generalizes a result of the first author, who handled the case when $|f|\le 1$ in previous work.

math.NT↗

arXiv2020

On the Duffin-Schaeffer conjecture

Let $ψ:\mathbb{N}\to\mathbb{R}_{\ge0}$ be an arbitrary function from the positive integers to the non-negative reals. Consider the set $\mathcal{A}$ of real numbers $α$ for which there are infinitely many reduced fractions $a/q$ such that $|α-a/q|\le ψ(q)/q$. If $\sum_{q=1}^\infty ψ(q)ϕ(q)/q=\infty$, we show that $\mathcal{A}$ has full Lebesgue measure. This answers a question of Duffin and Schaeffer. As a corollary, we also establish a conjecture due to Catlin regarding non-reduced solutions to the inequality $|α- a/q|\le ψ(q)/q$, giving a refinement of Khinchin's Theorem.

math.NT↗

arXiv2020

Sieve weights and their smoothings

We obtain asymptotic formulas for the $2k$th moments of partially smoothed divisor sums of the Möbius function. When $2k$ is small compared with $A$, the level of smoothing, then the main contribution to the moments come from integers with only large prime factors, as one would hope for in sieve weights. However if $2k$ is any larger, compared with $A$, then the main contribution to the moments come from integers with quite a few prime factors, which is not the intention when designing sieve weights. The threshold for "small" occurs when $A=\frac 1{2k} \binom{2k}{k}-1$. One can ask analogous questions for polynomials over finite fields and for permutations, and in these cases the moments behave rather differently, with even less cancellation in the divisor sums. We give, we hope, a plausible explanation for this phenomenon, by studying the analogous sums for Dirichlet characters, and obtaining each type of behaviour depending on whether or not the character is "exceptional".

math.NT↗

arXiv2019

Equal sums in random sets and the concentration of divisors

We study the extent to which divisors of a typical integer $n$ are concentrated. In particular, defining the Erd\H{o}s-Hooley $\Delta$-function by $\Delta(n) := \max_t \# \{d | n, \log d \in [t,t+1]\}$, we show that $\Delta(n) \geq (\log \log n)^{0.35332277\dots}$ for almost all $n$, a bound we believe to be sharp. This disproves a conjecture of Maier and Tenenbaum. We also prove analogs for the concentration of divisors of a random permutation and of a random polynomial over a finite field. Most of the paper is devoted to a study of the following much more combinatorial problem of independent interest. Pick a random set $A \subset \mathbb{N}$ by selecting $i$ to lie in $A$ with probability $1/i$. What is the supremum of all exponents $\beta_k$ such that, almost surely as $D \rightarrow \infty$, some integer is the sum of elements of $A \cap [D^{\beta_k}, D]$ in $k$ different ways? We characterise $\beta_k$ as the solution to a certain optimisation problem over measures on the discrete cube $\{0,1\}^k$, and obtain lower bounds for $\beta_k$ which we believe to be asymptotically sharp.

math.NT↗

arXiv2018

Beyond the LSD method for the partial sums of multiplicative functions

The Landau-Selberg-Delange (LSD) method gives an asymptotic formula for the partial sums of a multiplicative function $f$ whose prime values are $α$ on average. In the literature, the average is usually taken to be $α$ with a very strong error term, leading to an asymptotic formula for the partial sums with a very strong error term. In practice, the average at the prime values may only be known with a fairly weak error term, and so we explore here how good an estimate this will imply for the partial sums of $f$, developing new techniques to do so.

math.NT↗

arXiv2017

The frequency and the structure of large character sums

Let $M(χ)$ denote the maximum of $|\sum_{n\le N}χ(n)|$ for a given non-principal Dirichlet character $χ\pmod q$, and let $N_χ$ denote a point at which the maximum is attained. In this article we study the distribution of $M(χ)/\sqrt{q}$ as one varies over characters $\pmod q$, where $q$ is prime, and investigate the location of $N_χ$. We show that the distribution of $M(χ)/\sqrt{q}$ converges weakly to a universal distribution $Φ$, uniformly throughout most of the possible range, and get (doubly exponential decay) estimates for $Φ$'s tail. Almost all $χ$ for which $M(χ)$ is large are odd characters that are $1$-pretentious. Now, $M(χ)\ge |\sum_{n\le q/2}χ(n)| = \frac{|2-χ(2)|}π\sqrt{q} |L(1,χ)|$, and one knows how often the latter expression is large, which has been how earlier lower bounds on $Φ$ were mostly proved. We show, though, that for most $χ$ with $M(χ)$ large, $N_χ$ is bounded away from $q/2$, and the value of $M(χ)$ is little bit larger than $\frac{\sqrt{q}}π |L(1,χ)|$.

math.NT↗

arXiv2017

Permutations contained in transitive subgroups

In the first paper in this series we estimated the probability that a random permutation $π\in\mathcal{S}_n$ has a fixed set of a given size. In this paper, we elaborate on the same method to estimate the probability that $π$ has $m$ disjoint fixed sets of prescribed sizes $k_1,\dots,k_m$, where $k_1+\cdots+k_m=n$. We deduce an estimate for the proportion of permutations contained in a transitive subgroup other than $\mathcal{S}_n$ or $\mathcal{A}_n$. This theorem consists of two parts: an estimate for the proportion of permutations contained in an imprimitive transitive subgroup, and an estimate for the proportion of permutations contained in a primitive subgroup other than $\mathcal{S}_n$ or $\mathcal{A}_n$.

math.GR↗