Source author record

Vsevolod F. Lev

Explore this author’s publications, collaborators and research topics. This catalog record has not been claimed by its author.

ResearcherUnclaimed source record

Catalog footprint

Research overview

16works
5topics
4close collaborators

Published work

16 publications

preprint2022arXiv

The structure of higher sumsets

Merging together a result of Nathanson from the early 70s and a recent result of Granville and Walker, we show that for any finite set AA of integers with min⁡(A)=0\min(A)=0 and gcd⁡(A)=1\gcd(A)=1 there exist two sets, the "head" and the "tail", such that if m≥max⁡(A)−∣A∣+2m\ge\max(A)-|A|+2, then the mm-fold sumset mAmA consists of the union of these sets and a long block of consecutive integers separating them. We give sharp estimates for the length of the block, and investigate the corresponding stability problem classifying those sets AA for which the bound max⁡(A)−∣A∣+2\max(A)-|A|+2 cannot be substantially improved.

preprint2016arXiv

Minimizing the sum of projections of a finite set

Consider the projections of a finite set A⊂RnA\subset R^n onto the coordinate hyperplanes. How small can the sum of the sizes of these projections be, given the size of AA? In a different form, this problem has been studied earlier in the context of edge-isoperimetric inequalities on graphs, and it is can be derived from the known results that there is a linear order on the set of nn-tuples with non-negative integer coordinates, such that the sum in question is minimised for the initial segments with respect to this order. We present a new, self-contained and constructive proof, enabling us to obtain a stability result and establish algebraic properties of the smallest possible projection sum. We also solve the problem of minimising the sum of the sizes of the one-dimensional projections.

preprint2016arXiv

Stability result for sets with 3A≠Z5n3A\ne{\mathbb Z}_5^n

As an easy corollary of Kneser's Theorem, if AA is a subset of the elementary abelian group Z5n{\mathbb Z}_5^n of density 5−n∣A∣>0.45^{-n}|A|>0.4, then 3A=Z5n3A={\mathbb Z}_5^n. We establish the complementary stability result: if 5−n∣A∣>0.35^{-n}|A|>0.3 and 3A≠Z5n3A\ne{\mathbb Z}_5^n, then AA is contained in a union of two cosets of an index-55 subgroup of Z5n{\mathbb Z}_5^n. Here the density bound 0.30.3 is sharp. Our argument combines combinatorial reasoning with a somewhat non-standard application of the character sum technique.

preprint2016arXiv

Symmetric Kneser's Theorem with Trios and 33-Transform

We give a new equivalent restatement and a new proof in terms of trios to the classical Kneser's theorem. In the finite case, our restatement takes the following, particularly symmetric shape: if AA, BB, and CC are subsets of a finite abelian group GG such that A+B+C≠GA+B+C\ne G, then, denoting by HH the period of the sumset A+B+CA+B+C, we have ∣A∣+∣B∣+∣C∣≤∣G∣+∣H∣. |A|+|B|+|C| \le |G|+|H|. The proof is based on an extension of the familiar Dyson transform onto set systems containing three (or more) sets.

preprint2015arXiv

Quadratic residues and difference sets

It has been conjectured by Sarkozy that with finitely many exceptions, the set of quadratic residues modulo a prime pp cannot be represented as a sumset {a+b ⁣:a∈A,b∈B}\{a+b\colon a\in A, b\in B\} with non-singleton sets A,B⊂FpA,B\subset F_p. The case A=BA=B of this conjecture has been recently established by Shkredov. The analogous problem for differences remains open: is it true that for all sufficiently large primes pp, the set of quadratic residues modulo pp is not of the form {a′−a" ⁣:a′,a"∈A, a′≠a"}\{a'-a"\colon a',a"\in A,\,a'\ne a"\} with A⊂FpA\subset F_p? We attack here a presumably more tractable variant of this problem, which is to show that there is no A⊂FpA\subset F_p such that every quadratic residue has a \emph{unique}representation as a′−a"a'-a" with a′,a"∈Aa',a"\in A, and no non-residue is represented in this form. We give a number of necessary conditions for the existence of such AA, involving for the most part the behavior of primes dividing p−1p-1. These conditions enable us to rule out all primes pp in the range 13<p<101813<p<10^{18} (the primes p=5p=5 and p=13p=13 being conjecturally the only exceptions).

preprint2014arXiv

Discrete norms of a matrix and the converse to the Expander Mixing Lemma

We define the discrete norm of a complex m×nm\times n matrix AA by ∥A∥Δ:=max⁡0≠ξ∈{0,1}n∥Aξ∥∥ξ∥, \|A\|_Δ:= \max_{0\neξ\in\{0,1\}^n} \frac{\|Aξ\|}{\|ξ\|}, and show that clog⁡h(A)+1 ∥A∥≤∥A∥Δ≤∥A∥, \frac c{\sqrt{\log h(A)+1}}\,\|A\| \le \|A\|_Δ\le \|A\|, where c>0c>0 is an explicitly indicated absolute constant, h(A)=∥A∥1∥A∥∞/∥A∥h(A)=\sqrt{\|A\|_1\|A\|_\infty}/\|A\|, and ∥A∥1,∥A∥∞\|A\|_1,\|A\|_\infty, and ∥A∥=∥A∥2\|A\|=\|A\|_2 are the induced operator norms of AA. Similarly, for the \emph{discrete Rayleigh norm} ∥A∥P:=max⁡0≠ξ∈{0,1}m0≠η∈{0,1}n∣ξtAη∣∥ξ∥∥η∥ \|A\|_P := \max_{\substack{0\neξ\in\{0,1\}^m \\ 0\neη\in\{0,1\}^n}} \frac{|ξ^tAη|}{\|ξ\|\|η\|} we prove the estimate clog⁡h(A)+1 ∥A∥≤∥A∥P≤∥A∥. \frac c{\log h(A)+1}\,\|A\| \le \|A\|_P \le \|A\|. These estimates are shown to be essentially best possible. As a consequence, we obtain another proof of the (slightly sharpened and generalized version of the) converse to the expander mixing lemma by Bollobas-Nikiforov and Bilu-Linial.

preprint2013arXiv

Approximate convexity and an edge-isoperimetric estimate

We study extremal properties of the function F(x):=min⁡{k∥x∥1−1/k ⁣:k≥1}, x∈[0,1], F(x) := \min\{k\|x\|^{1-1/k}\colon k\ge 1\},\ x\in[0,1], where ∥x∥=min⁡{x,1−x}\|x\|=\min\{x,1-x\}. In particular, we show that FF is the pointwise largest function of the class of all real-valued functions ff defined on the interval [0,1][0,1], and satisfying the relaxed convexity condition f(tx1+(1−t)x2)≤tf(x1)+(1−t)f(x2)+∣x2−x1∣, x1,x2,t∈[0,1] f(tx_1+(1-t)x_2) \le tf(x_1)+(1-t)f(x_2)+|x_2-x_1|, \ x_1,x_2,t\in[0,1] and the boundary condition max⁡{f(0),f(1)}≤0\max\{f(0),f(1)\}\le 0. As an application, we prove that if AA and SS are subsets of a finite abelian group GG, such that SS is generating and all of its elements have order at most mm, then the number of edges from AA to its complement G∖AG\setminus A in the directed Cayley graph induced by SS on GG is ∂S(A)≥1m∣G∣F(∣A∣/∣G∣). \partial_S(A) \ge \frac{1}{m} |G| F(|A|/|G|).

preprint2013arXiv

Flat-containing and shift-blocking sets in F2rF_2^r

For non-negative integers r≥dr\ge d, how small can a subset C⊂F2rC\subset F_2^r be, given that for any v∈F2rv\in F_2^r there is a dd-flat passing through vv and contained in C∪{v}C\cup\{v\}? Equivalently, how large can a subset B⊂F2rB\subset F_2^r be, given that for any v∈F2rv\in F_2^r there is a linear dd-subspace not blocked non-trivially by the translate B+vB+v? A number of lower and upper bounds are obtained.

preprint2012arXiv

Edge-isoperimetric problem for Cayley graphs and generalized Takagi function

Let GG be a finite abelian abelian group of exponent m≥2m\ge 2. For subsets A,S⊂GA,S\subset G, denote by ∂S(A)\partial_S(A) the number of edges from AA to its complement G∖AG\setminus A in the directed Cayley graph, induced by SS on GG. We show that if SS generates GG, and AA is non-empty, then ∂S(A)≥em ∣A∣ln⁡∣G∣∣A∣. \partial_S(A) \ge \frac{e}m\,|A|\ln\frac{|G|}{|A|}. Here the coefficient e=2.718...e=2.718... is best possible and cannot be replaced with a number larger than ee. For homocyclic groups GG of exponent mm, we find an explicit closed-form expression for ∂S(A)\partial_S(A) in the case where SS is a "standard" generating subset of GG, and AA is an initial segment of GG with respect to the lexicographic order, induced by SS on GG. Namely, we show that in this situation ∂S(A)=∣G∣ ωm(∣A∣/∣G∣), \partial_S(A) = |G|\,ω_m(|A|/|G|), where ω2ω_2 is the Takagi function, and ωmω_m for m≥3m\ge 3 is an appropriate generalization thereof. This particular case is of special interest, since for m∈{2,3,4}m\in\{2,3,4\} it is known to yield the smallest possible value of ∂S(A)\partial_S(A), over all sets A⊂GA\subset G of given size. We give this classical result a new proof, somewhat different from the standard one. We also give a new, short proof of the Boros-Pales inequality ω2(x+y2)≤ω2(x)+ω2(y)2+12 ∣y−x∣,ω_2(\frac{x+y}2) \le \frac{ω_2(x) + ω_2(y)}2 + \frac12\,|y-x|, establish an extremal characterization of the Takagi function as the (pointwise) maximal function, satisfying this inequality and the boundary condition max⁡{ω2(0),ω2(1)}≤0\max\{ω_2(0),ω_2(1)\}\le 0, and obtain similar results for the 3-adic analog ω3ω_3 of the Takagi function.

preprint2012arXiv

Solving a±b=2ca\pm b=2c in the elements of finite sets

We show that if AA and BB are finite sets of real numbers, then the number of triples (a,b,c)∈A×B×(A∪B)(a,b,c)\in A\times B\times (A\cup B) with a+b=2ca+b=2c is at most (0.15+o(1))(∣A∣+∣B∣)2(0.15+o(1))(|A|+|B|)^2 as ∣A∣+∣B∣→∞|A|+|B|\to\infty. As a corollary, if AA is antisymmetric (that is, $A\cap(-A)=\est$), then there are at most (0.3+o(1))∣A∣2(0.3+o(1))|A|^2 triples (a,b,c)(a,b,c) with a,b,c∈Aa,b,c\in A and a−b=2ca-b=2c. In the general case where AA is not necessarily antisymmetric, we show that the number of triples (a,b,c)(a,b,c) with a,b,c∈Aa,b,c\in A and a−b=2ca-b=2c is at most (0.5+o(1))∣A∣2(0.5+o(1))|A|^2. These estimates are sharp.

preprint2011arXiv

The Largest Eigenvalue and Bi-Average Degree of a Graph

We show that for a graph GG with the vertex set VV and the largest eigenvalue λmax⁡(G)λ_{\max}(G), letting M(G):=max⁡X,Y⊂Ve(X,Y)∣X∣∣Y∣ M(G) := \max_{X,Y \subset V} \frac{e(X,Y)}{\sqrt{|X||Y|}} (where e(X,Y)e(X,Y) denotes the number of edges between XX and YY), we have $$ M(G) \le λ_{\max}(G) \le \big(\frac14 \log|V| + 1 \big) \M(G). $$ Here the lower bound is attained if GG is regular or bi-regular, whereas the logarithmic factor in the upper bound, conjecturally, can be improved --- although we present an example showing that it cannot be replaced with a factor growing slower than (log⁡∣V∣/log⁡log⁡∣V∣)1/8(\log |V|/\log\log|V|)^{1/8}. Further refinements are established, particularly in the case where GG is bipartite.

preprint2010arXiv

Kakeya-type sets in finite vector spaces

For a finite vector space VV and a non-negative integer r≤dim⁡Vr\le\dim V we estimate the smallest possible size of a subset of VV, containing a translate of every rr-dimensional subspace. In particular, we show that if K⊂VK\subset V is the smallest subset with this property, nn denotes the dimension of VV, and qq is the size of the underlying field, then for rr bounded and r<n≤rqr−1r<n\le rq^{r-1} we have ∣V∖K∣=Θ(nqn−r+1)|V\setminus K|=Θ(nq^{n-r+1}). This improves previously known bounds ∣V∖K∣=Ω(qn−r+1)|V\setminus K|=Ω(q^{n-r+1}) and ∣V∖K∣=O(n2qn−r+1)|V\setminus K|=O(n^2q^{n-r+1}).

Research graph

See the researcher in context

Open full explorer

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.