SearcharxivSearch

arXiv subjects

O. Serra

Publications and source records attributed to O. Serra.

6 recordsLinked to original sources

On Doubling and Volume: Chains

The well--known Freiman--Ruzsa Theorem provides a structural description of a set $A$ of integers with $|2A|\le c|A|$ as a subset of a $d$--dimensional arithmetic progression $P$ with $|P|\le c'|A|$, where $d$ and $c'$ depend only on $c$. The estimation of the constants $d$ and $c'$ involved in the statement has been the object of intense research. Freiman conjectured in 2008 a formula for the largest volume of such a set. In this paper we prove the conjecture for a general class of sets called chains.

math.NT

Inverse Additive Problems for Minkowski Sumsets I

We give the structure of discrete two-dimensional finite sets $A,\,B\subseteq \R^2$ which are extremal for the recently obtained inequality $|A+B|\ge (\frac{|A|}{m}+\frac{|B|}{n}-1)(m+n-1)$, where $m$ and $n$ are the minimum number of parallel lines covering $A$ and $B$ respectively. Via compression techniques, the above bound also holds when $m$ is the maximal number of points of $A$ contained in one of the parallel lines covering $A$ and $n$ is the maximal number of points of $B$ contained in one of the parallel lines covering $B$. When $m,\,n\geq 2$, we are able to characterize the case of equality in this bound as well. We also give the structure of extremal sets in the plane for the projection version of Bonnesen's sharpening of the Brunn-Minkowski inequality: $μ(A+B)\ge (μ(A)/m+μ(B)/n)(m+n)$, where $m$ and $n$ are the lengths of the projections of $A$ and $B$ onto a line.

math.NT

Inverse Additive Problems for Minkowski Sumsets II

The Brunn-Minkowski Theorem asserts that $μ_d(A+B)^{1/d}\geq μ_d(A)^{1/d}+μ_d(B)^{1/d}$ for convex bodies $A,\,B\subseteq \R^d$, where $μ_d$ denotes the $d$-dimensional Lebesgue measure. It is well-known that equality holds if and only if $A$ and $B$ are homothetic, but few characterizations of equality in other related bounds are known. Let $H$ be a hyperplane. Bonnesen later strengthened this bound by showing $$μ_d(A+B)\geq (M^{1/(d-1)}+N^{1/(d-1)})^{d-1}(\frac{μ_d(A)}{M}+\frac{μ_d(B)}{N}),$$ where $M=\sup\{μ_{d-1}((\mathbf x+H)\cap A)\mid \mathbf x\in \R^d\}$ and $N=\sup\{μ_{d-1}((\mathbf y+H)\cap B)\mid \mathbf y\in \R^d\}$. Standard compression arguments show that the above bound also holds when $M=μ_{d-1}(π(A))$ and $N=μ_{d-1}(π(B))$, where $π$ denotes a projection of $\mathbb R^d$ onto $H$, which gives an alternative generalization of the Brunn-Minkowski bound. In this paper, we characterize the cases of equality in this later bound, showing that equality holds if and only if $A$ and $B$ are obtained from a pair of homothetic convex bodies by `stretching' along the direction of the projection, which is made formal in the paper. When $d=2$, we characterize the case of equality in the former bound as well.

math.NT

A note on Pollard's Theorem

Let $A,B$ be nonempty subsets of a an abelian group $G$. Let $N_i(A,B)$ denote the set of elements of $G$ having $i$ distinct decompositions as a product of an element of $A$ and an element of $B$. We prove that $$ \sum _{1\le i \le t} |N_i (A,B)|\ge t(|A|+|B|- t-α+1+w)-w, $$ where $α$ is the largest size of a coset contained in $AB$ and $w=\min (α-1,1)$, with a strict inequality if $α\ge 3$ and $t\ge 2$, or if $α\ge 2$ and $t= 2$. This result is a local extension of results by Pollard and Green--Ruzsa and extends also for $t>2$ a recent result of Grynkiewicz, conjectured by Dicks--Ivanov (for non necessarily abelian groups) in connection to the famous Hanna Neumann problem in Group Theory.

math.NT

The lonely runner with seven runners

Suppose $k+1$ runners having nonzero constant speeds run laps on a unit-length circular track starting at the same time and place. A runner is said to be lonely if she is at distance at least $1/(k+1)$ along the track to every other runner. The lonely runner conjecture states that every runner gets lonely. The conjecture has been proved up to six runners ($k\le 5$). A formulation of the problem is related to the regular chromatic number of distance graphs. We use a new tool developed in this context to solve the first open case of the conjecture with seven runners.

math.CO

On complete subsets of the cyclic group

A subset $X$ of an abelian $G$ is said to be {\em complete} if every element of the subgroup generated by $X$ can be expressed as a nonempty sum of distinct elements from $X$. Let $A\subset \Z_n$ be such that all the elements of $A$ are coprime with $n$. Solving a conjecture of Erdős and Heilbronn, Olson proved that $A$ is complete if $n$ is a prime and if $|A|>2\sqrt{n}.$ Recently Vu proved that there is an absolute constant $c$, such that for an arbitrary large $n$, $A$ is complete if $|A|\ge c\sqrt{n},$ and conjectured that 2 is essentially the right value of $c$. We show that $A$ is complete if $|A|> 1+2\sqrt{n-4}$, thus proving the last conjecture.

math.NT