arXiv subjects

Shalom Eliahou

Publications and source records attributed to Shalom Eliahou.

At least 19 recordsLinked to original sources

arXiv2025

A Normality Conjecture on Rational Base Number Systems

The rational base number system, introduced by Akiyama, Frougny, and Sakarovitch in 2008, is a generalization of the classical integer base number system. Within this framework two interesting families of infinite words emerge, called minimal and maximal words. We conjecture that every minimal and maximal word is normal over an appropriate subalphabet. To support this conjecture, we present extensive numerical experiments that examine the richness threshold and the deviation from normality of these words. We also discuss the implications that the validity of our conjecture would have for several long-standing open problems, including the existence of $Z$-numbers (Mahler, 1968) and $Z_{p/q}$-numbers (Flatto, 1992), the existence of triple expansions in rational base $p/q$ (Akiyama, 2008), and the Collatz-inspired `4/3 problem' (Dubickas and Mossinghoff, 2009).

math.NT↗

arXiv2025

The number system in rational base $3/2$ and the $3x+1$ problem

The representation of numbers in rational base $p/q$ was introduced in 2008 by Akiyama, Frougny & Sakarovitch, with a special focus on the case $p/q=3/2$. Unnoticed since then, natural questions related to representations in that specific base turn out to intimately involve the Collatz $3x+1$ function. Our purpose in this note is to expose these links and motivate further research into them.

math.NT↗

arXiv2024

On the Gotzmann threshold of monomials

Let $R_n=K[x_1,\dots,x_n]$ be the $n$-variable polynomial ring over a field $K$. Let $S_n$ denote the set of monomials in $R_n$. A monomial $u \in S_n$ is a \textit{Gotzmann monomial} if the Borel-stable monomial ideal $\langle u \rangle$ it generates in $R_n$ is a Gotzmann ideal. A longstanding open problem is to determine all Gotzmann monomials in $R_n$. Given $u_0 \in S_{n-1}$, its \textit{Gotzmann threshold} is the unique nonnegative integer $t_0=τ_n(u_0)$ such that $u_0x_n^t$ is a Gotzmann monomial in $R_n$ if and only if $t \ge t_0$. Currently, the function $τ_n$ is exactly known for $n \le 4$ only. We present here an efficient procedure to determine $τ_n(u_0)$ for all $n$ and all $u_0 \in S_{n-1}$. As an application, in the critical case $u_0=x_2^d$, we determine $τ_5(x_2^d)$ for all $d$ and we conjecture that for $n \ge 6$, $τ_n(x_2^d)$ is a polynomial in $d$ of degree $2^{n-2}$ and dominant term equal to that of the $(n-2)$-iterated binomial coefficient $$ \binom {\binom {\binom d2}2}{\stackrel{\cdots}2}. $$

math.AC↗

arXiv2023

Optimal Bounds on the Growth of Iterated Sumsets in Abelian Semigroups

We provide optimal upper bounds on the growth of iterated sumsets $hA=A+\dots+A$ for finite subsets $A$ of abelian semigroups. More precisely, we show that the new upper bounds recently derived from Macaulay's theorem in commutative algebra are best possible, i.e., are actually reached by suitable subsets of suitable abelian semigroups. Our constructions, in a multiplicative setting, are based on certain specific monomial ideals in polynomial algebras and on their deformation into appropriate binomial ideals via Gröbner bases.

math.AC↗

arXiv2023

A verification of Wilf's conjecture up to genus 100

For a numerical semigroup $S \subseteq \mathbb{N}$, let $m,e,c,g$ denote its multiplicity, embedding dimension, conductor and genus, respectively. Wilf's conjecture (1978) states that $e(c-g) \ge c$. As of 2023, Wilf's conjecture has been verified by computer up to genus $g \le 66$. In this paper, we extend the verification of Wilf's conjecture up to genus $g \le 100$. This is achieved by combining three main ingredients: (1) a theorem in 2020 settling Wilf's conjecture in the case $e \ge m/3$, (2) an efficient trimming of the tree $\mathcal{T}$ of numerical groups identifying and cutting out irrelevant subtrees, and (3) the implementation of a fast parallelized algorithm to construct the tree $\mathcal{T}$ up to a given genus. We further push the verification of Wilf's conjecture up to genus $120$ in the particular case where $m$ divides $c$. Finally, we unlock three previously unknown values of the number $n_g$ of numerical semigroups of genus $g$, namely for $g=73,74,75$.

math.GR↗

arXiv2021

Is the Syracuse falling time bounded by 12?

Let $T \colon \mathbb{N} \to \mathbb{N}$ denote the $3x+1$ function, where $T(n)=n/2$ if $n$ is even, $T(n)=(3n+1)/2$ if $n$ is odd. As an accelerated version of $T$, we define a jump at $n \ge 1$ by jp$(n) = T^{(\ell)}(n)$, where $\ell$ is the number of digits of $n$ in base 2. We present computational and heuristic evidence leading to surprising conjectures. The boldest one, inspired by the study of $2^{\ell}-1$ for $\ell \le 500000$, states that for any $n \ge 2^{500}$, at most four jumps starting from $n$ are needed to fall below $n$, a strong form of the Collatz conjecture.

math.NT↗

arXiv2020

Iterated sumsets and Hilbert functions

Let A be a finite subset of an abelian group (G, +). Let h $\ge$ 2 be an integer. If |A| $\ge$ 2 and the cardinality |hA| of the h-fold iterated sumset hA = A + $\times$ $\times$ $\times$ + A is known, what can one say about |(h -- 1)A| and |(h + 1)A|? It is known that |(h -- 1)A| $\ge$ |hA| (h--1)/h , a consequence of Pl{ü}nnecke's inequality. Here we improve this bound with a new approach. Namely, we model the sequence |hA| h$\ge$0 with the Hilbert function of a standard graded algebra. We then apply Macaulay's 1927 theorem on the growth of Hilbert functions, and more specifically a recent condensed version of it. Our bound implies |(h -- 1)A| $\ge$ $θ$(x, h) |hA| (h--1)/h for some factor $θ$(x, h) > 1, where x is a real number closely linked to |hA|. Moreover, we show that $θ$(x, h) asymptotically tends to e $\approx$ 2.718 as |A| grows and h lies in a suitable range varying with |A|.

math.AC↗

arXiv2020

The Schur degree of additive sets

Let (G, +) be an abelian group. A subset of G is sumfree if it contains no elements x, y, z such that x +y = z. We extend this concept by introducing the Schur degree of a subset of G, where Schur degree 1 corresponds to sumfree. The classical inequality S(n) $\le$ R n (3) -- 2, between the Schur number S(n) and the Ramsey number R n (3) = R(3,. .. , 3), is shown to remain valid in a wider context, involving the Schur degree of certain subsets of G. Recursive upper bounds are known for R n (3) but not for S(n) so far. We formulate a conjecture which, if true, would fill this gap. Indeed, our study of the Schur degree leads us to conjecture S(n) $\le$ n(S(n -- 1) + 1) for all n $\ge$ 2. If true, it would yield substantially better upper bounds on the Schur numbers, e.g. S(6) $\le$ 966 conjecturally, whereas all is known so far is 536 $\le$ S(6) $\le$ 1836.

math.CO↗

arXiv2020

On numerical semigroups with at most 12 left elements

For a numerical semigroup S $\subseteq$ N with embedding dimension e, conductor c and left part L = S $\cap$ [0, c -- 1], set W (S) = e|L| -- c. In 1978 Wilf asked, in equivalent terms, whether W (S) $\ge$ 0 always holds, a question known since as Wilf's conjecture. Using a closely related lower bound W 0 (S) $\le$ W (S), we show that if |L| $\le$ 12 then W 0 (S) $\ge$ 0, thereby settling Wilf's conjecture in this case. This is best possible, since cases are known where |L| = 13 and W 0 (S) = --1. Wilf's conjecture remains open for |L| $\ge$ 13.

math.CO↗

arXiv2019

An adaptive upper bound on the Ramsey numbers $R(3,\dots,3)$

Since 2002, the best known upper bound on the Ramsey numbers R n (3) = R(3,. .. , 3) is R n (3) $\le$ n!(e -- 1/6) + 1 for all n $\ge$ 4. It is based on the current estimate R 4 (3) $\le$ 62. We show here how any closing-in on R 4 (3) yields an improved upper bound on R n (3) for all n $\ge$ 4. For instance, with our present adaptive bound, the conjectured value R 4 (3) = 51 implies R n (3) $\le$ n!(e -- 5/8) + 1 for all n $\ge$ 4.

math.CO↗

arXiv2019

A graph-theoretic approach to Wilf's conjecture

Let S $\subseteq$ N be a numerical semigroup with multiplicity m = min(S \ {0}) and conductor c = max(N \ S) + 1. Let P be the set of primitive elements of S, and let L be the set of elements of S which are smaller than c. A longstand-ing open question by Wilf in 1978 asks whether the inequality |P||L| $\ge$ c always holds. Among many partial results, Wilf's conjecture has been shown to hold in case |P| $\ge$ m/2 by Sammartano in 2012. Using graph theory in an essential way, we extend the verification of Wilf's conjecture to the case |P| $\ge$ m/3. This case covers more than 99.999% of numerical semigroups of genus g $\le$ 45.

math.CO↗

arXiv2019

Gotzmann Monomials In Four Variables

It is a widely open problem to determine which monomials in the n-variable polynomial ring $K[x_1,...,x_n]$ over a field $K$ have the Gotzmann property, i.e. induce a Borel-stable Gotzmann monomial ideal. Since 2007, only the case $n \le 3$ was known. Here we solve the problem for the case $n = 4$. The solution involves a surprisingly intricate characterization.

math.AC↗

arXiv2019

Some Results On The Flynn-Poonen-Schaefer Conjecture

For $c \in \mathbb{Q}$, consider the quadratic polynomial map $φ_c(x)=x^2-c$. Flynn, Poonen and Schaefer conjectured in 1997 that no rational cycle of $φ_c$ under iteration has length more than $3$. Here we discuss this conjecture using arithmetic and combinatorial means, leading to three main results. First, we show that if $φ_c$ admits a rational cycle of length $n \ge 3$, then the denominator of $c$ must be divisible by $16$. We then provide an upper bound on the number of periodic rational points of $φ_c$ in terms of the number of distinct prime factors of the denominator of $c$. Finally, we show that the Flynn-Poonen-Schaefer conjecture holds for $φ_c$ if that denominator has at most two distinct prime factors.

math.CO↗

arXiv2019

Gapsets of small multiplicity

A gapset is the complement of a numerical semigroup in N. In this paper, we characterize all gapsets of multiplicity m $\le$ 4. As a corollary, we provide a new simpler proof that the number of gapsets of genus g and fixed multiplicity m $\le$ 4 is a nondecreasing function of g.

math.GR↗

arXiv2018

Gapsets and numerical semigroups

For g $\ge$ 0, let n g denote the number of numerical semi-groups of genus g. A conjecture by Maria Bras-Amorós in 2008 states that the inequality n g $\ge$ n g--1 + n g--2 should hold for all g $\ge$ 2. Here we show that such an inequality holds for the very large subtree of numerical semigroups satisfying c $\le$ 3m, where c and m are the conductor and multiplicity, respectively. Our proof is given in the more flexible setting of gapsets, i.e. complements in N of numerical semigroups.

math.CO↗

arXiv2018

Monomial ideals with tiny squares

Let $I \subset K[x,y]$ be a monomial ideal. How small can $μ(I^2)$ be in terms of $μ(I)$? It has been expected that the inequality $μ(I^2) > μ(I)$ should hold whenever $μ(I) \ge 2$. Here we disprove this expectation and provide a somewhat surprising answer to the above question.

math.AC↗

arXiv2017

Near-misses in Wilf's conjecture

Let S $\subseteq$ N be a numerical semigroup with multiplicity m, conductor c and minimal generating set P. Let L = S $\cap$ [0, c -- 1] and W(S) = |P||L| -- c. In 1978, Herbert Wilf asked whether W(S) $\ge$ 0 always holds, a question known as Wilf's conjecture and open since then. A related number W0(S), satisfying W0(S) $\le$ W(S), has recently been introduced. We say that S is a near-miss in Wilf's conjecture if W0(S) < 0. Near-misses are very rare. Here we construct infinite families of them, with c = 4m and W0(S) arbitrarily small, and we show that the members of these families still satisfy Wilf's conjecture.

math.CO↗

arXiv2017

A remarkable 20-crossing tangle

For any positive integer r, we exhibit a knot Kr with (20 $\times$ 2 r--1 + 1) crossings whose Jones polynomial V (Kr) is equal to 1 mod-ulo 2 r. Our construction rests on a certain 20-crossing tangle T 20 which is undetectable by the Kauffman bracket polynomial pair mod 2.

math.GT↗