arXiv subjects

Gábor Tardos

Publications and source records attributed to Gábor Tardos.

At least 19 recordsLinked to original sources

arXiv2026

Cutting a convex body into fat parts and approximating Euclidean distance by graph distances

Can one construct a graph $G$ on the set of integer points ${\mathbb Z}^2$ in the plane such that the length of the shortest path between any two vertices of $G$ differs from their Euclidean distance by at most an absolute constant? This question of Benjamini, Erd\H os, Kleiner, Kozma, Schramm, and the first-named author has been open for a long time. We give an affirmative answer to a weaker form of this question, based on the following geometric statement, which is of independent interest. There exists a constant $c>0$ such that for every $i=1,2,\ldots,$ every $ρ$-fat plane convex set $S$ can be cut into $2^i$ convex pieces of equal area, each of which is at least $cρ$-fat. (A convex set is $ρ$-fat if the ratio of its inradius to its circumradius is at least $ρ$.) We prove that there exists an (unweighted) spanning subgraph $G$ of an enlarged copy of ${\mathbb Z}^2$ such that, for every pair of vertices at Euclidean distance $d$, their shortest-path distance in $G$ lies between $d-O(1)$ and $d+o(d^{5/6})$. The same bound can be achieved by a planar graph with vertex set ${\mathbb Z}^2$, in which every edge joins two vertices at Euclidean distance at most 2.

math.CO↗

arXiv2026

Uniform Turán density -- palette classification

In the 1980s, Erdős and Sós initiated the study of Turán hypergraph problems with a uniformity condition on the distribution of edges, i.e., determining density thresholds for the existence of a hypergraph H in a host hypergraph with edges uniformly distributed. In particular, Erdős and Sós asked to determine the uniform Turán densities of the hypergraphs $K_4^{(3)-}$ and $K_4^{(3)}$. After more than 30 years, the former was solved by Glebov, Král' and Volec [Israel J. Math. 211 (2016), 349-366] and Reiher, Rödl and Schacht [J. Eur. Math. Soc. 20 (2018), 1139-1159], while the latter still remains open. In these two cases and several additional cases, the tight lower bounds are provided by a so-called palette construction. Lamaison [arXiv:2408.09643] has recently showed that the uniform Turán density of a 3-uniform hypergraph H is equal to the supremum of the densities of palettes that H is not colorable with. We give a necessary and sufficient condition, which is easy to verify, on the existence of a 3-uniform hypergraph colorable by a set of palettes and not colorable by another given set of palettes. We also demonstrate how our result can be used to prove the existence of 3-uniform hypergraphs with specific values of the uniform Turán density.

math.CO↗

arXiv2026

Piercing all maximum cliques in hypergraphs

Graphs whose maximum clique size exceeds half of the total number of vertices satisfy a classical property: the family of their maximum sized cliques can be pierced by a single vertex. This result dates back to a 1965 theorem by Hajnal. Motivated by this theorem, Jung, Keszegh, Pálvölgyi, and Yuditsky recently conjectured that an analogous result should hold for hypergraphs of larger uniformity, with an appropriate constant replacing the threshold $1/2$. In this paper we refute this conjecture in a strong form. We show that for any constant $c<1$ and integers $k\ge 3$ and $t\ge 1$, there exist $k$-uniform hypergraphs $G$ whose maximum clique size exceeds $c|V(G)|$, yet the family of maximum size cliques of $G$ cannot be pierced by $t$ vertices. This demonstrates that no universal constant threshold guarantees bounded piercing number for maximum cliques in uniform hypergraphs. We discuss further questions concerning the relationship between clique size and piercing maximum cliques in hypergraphs, and introduce a geometric variant of the problem using Helly's Theorem.

math.CO↗

arXiv2026

Unavoidable patterns and plane paths in dense topological graphs

Let $C_{s,t}$ be the complete bipartite geometric graph, with $s$ and $t$ vertices on two distinct parallel lines respectively, and all $s t$ straight-line edges drawn between them. In this paper, we show that every complete bipartite simple topological graph, with parts of size $2(k-1)^4 + 1$ and $2^{k^{5k}}$, contains a topological subgraph weakly isomorphic to $C_{k,k}$. As a corollary, every $n$-vertex simple topological graph not containing a plane path of length $k$ has at most $O_k(n^{2 - 8/k^4})$ edges. When $k = 3$, we obtain a stronger bound by showing that every $n$-vertex simple topological graph not containing a plane path of length 3 has at most $O(n^{4/3})$ edges. We also prove that $x$-monotone simple topological graphs not containing a plane path of length 3 have at most a linear number of edges.

math.CO↗

arXiv2024

Tight bounds for intersection-reverse sequences, edge-ordered graphs and applications

In 2006, Marcus and Tardos proved that if $A^1,\dots,A^n$ are cyclic orders on some subsets of a set of $n$ symbols such that the common elements of any two distinct orders $A^i$ and $A^j$ appear in reversed cyclic order in $A^i$ and $A^j$, then $\sum_{i} |A^i|=O(n^{3/2}\log n)$. This result is tight up to the logarithmic factor and has since become an important tool in Discrete Geometry. We improve this to the optimal bound $O(n^{3/2})$. In fact, we show that if $A^1,\dots,A^n$ are linear orders on some subsets of a set of $n$ symbols such that no three symbols appear in the same order in any two distinct linear orders, then $\sum_{i} |A^i|=O(n^{3/2})$. Using this result, we resolve several open problems in Discrete Geometry and Extremal Graph Theory as follows. We prove that every $n$-vertex topological graph that does not contain a self-crossing four-cycle has $O(n^{3/2})$ edges. This resolves a problem of Marcus and Tardos from 2006. We also show that $n$ pseudo-circles in the plane can be cut into $O(n^{3/2})$ pseudo-segments, which, in turn, implies new bounds on point-circle incidences and on other geometric problems. Moreover, we prove that the edge-ordered Turán number of the four-cycle $C_4^{1243}$ is $Θ(n^{3/2})$. This answers a question of Gerbner, Methuku, Nagy, Pálvölgyi, Tardos and Vizer. Using different methods, we determine the largest possible extremal number that an edge-ordered forest of order chromatic number two can have. Kucheriya and Tardos showed that every such graph has extremal number at most $n2^{O(\sqrt{\log n})}$, and conjectured that this can be improved to $n(\log n)^{O(1)}$. We disprove their conjecture by showing that for every $C>0$, there exists an edge-ordered tree of order chromatic number two whose extremal number is $Ω(n 2^{C\sqrt{\log n}})$.

math.CO↗

arXiv2024

On edge-ordered graphs with linear extremal functions

The systematic study of Turán-type extremal problems for edge-ordered graphs was initiated by Gerbner et al. in 2020. Here we characterize connected edge-ordered graphs with linear extremal functions and show that the extremal function of other connected edge-ordered graphs is $Ω(n\log n)$. This characterization and dichotomy are similar in spirit to results of Füredi et al. (2020) about vertex-ordered and convex geometric graphs. We also extend the study of extremal function of short edge-ordered paths by Gerbner et al. to some longer paths.

math.CO↗

arXiv2024

A Refutation of the Pach-Tardos Conjecture for 0-1 Matrices

The theory of forbidden 0-1 matrices generalizes Turan-style (bipartite) subgraph avoidance, Davenport-Schinzel theory, and Zarankiewicz-type problems, and has been influential in many areas, such as discrete and computational geometry, the analysis of self-adjusting data structures, and the development of the graph parameter twin width. The foremost open problems in this area is to resolve the Pach-Tardos conjecture from 2005, which states that if a forbidden pattern $P\in\{0,1\}^{k\times l}$ is the bipartite incidence matrix of an acyclic graph (forest), then $\mathrm{Ex}(P,n) = O(n\log^{C_P} n)$, where $C_P$ is a constant depending only on $P$. This conjecture has been confirmed on many small patterns, specifically all $P$ with weight at most 5, and all but two with weight 6. The main result of this paper is a clean refutation of the Pach-Tardos conjecture. Specifically, we prove that $\mathrm{Ex}(S_0,n),\mathrm{Ex}(S_1,n) \geq n2^{Ω(\sqrt{\log n})}$, where $S_0,S_1$ are the outstanding weight-6 patterns. We also prove sharp bounds on the entire class of alternating patterns $(P_t)$, specifically that for every $t\geq 2$, $\mathrm{Ex}(P_t,n)=Θ(n(\log n/\log\log n)^t)$. This is the first proof of an asymptotically sharp bound that is $ω(n\log n)$.

math.CO↗

arXiv2023

On the Extremal Functions of Acyclic Forbidden 0-1 Matrices

The extremal theory of forbidden 0-1 matrices studies the asymptotic growth of the function $\mathrm{Ex}(P,n)$, which is the maximum weight of a matrix $A\in\{0,1\}^{n\times n}$ whose submatrices avoid a fixed pattern $P\in\{0,1\}^{k\times l}$. This theory has been wildly successful at resolving problems in combinatorics, discrete and computational geometry, structural graph theory, and the analysis of data structures, particularly corollaries of the dynamic optimality conjecture. All these applications use acyclic patterns, meaning that when $P$ is regarded as the adjacency matrix of a bipartite graph, the graph is acyclic. The biggest open problem in this area is to bound $\mathrm{Ex}(P,n)$ for acyclic $P$. Prior results have only ruled out the strict $O(n\log n)$ bound conjectured by Furedi and Hajnal. It is consistent with prior results that $\forall P. \mathrm{Ex}(P,n)\leq n\log^{1+o(1)} n$, and also consistent that $\forall ε>0.\exists P. \mathrm{Ex}(P,n) \geq n^{2-ε}$. In this paper we establish a stronger lower bound on the extremal functions of acyclic $P$. Specifically, we give a new construction of relatively dense 0-1 matrices with $Θ(n(\log n/\log\log n)^t)$ 1s that avoid an acyclic $X_t$. Pach and Tardos have conjectured that this type of result is the best possible, i.e., no acyclic $P$ exists for which $\mathrm{Ex}(P,n)\geq n(\log n)^{ω(1)}$.

math.CO↗

arXiv2023

A characterization of edge-ordered graphs with almost linear extremal functions

The systematic study of Turán-type extremal problems for edge-ordered graphs was initiated by Gerbner et al. arXiv:2001.00849. They conjectured that the extremal functions of edge-ordered forests of order chromatic number 2 are $n^{1+o(1)}$. Here we resolve this conjecture proving the stronger upper bound of $n2^{O(\sqrt{\log n})}$. This represents a gap in the family of possible extremal functions as other forbidden edge-ordered graphs have extremal functions $Ω(n^c)$ for some $c>1$. However, our result is probably not the last word: here we conjecture that the even stronger upper bound of $n\log^{O(1)}n$ also holds for the same set of extremal functions.

math.CO↗

arXiv2023

Where have all the grasshoppers gone?

Let $P$ be an $N$-element point set in the plane. Consider $N$ (pointlike) grasshoppers sitting at different points of $P$. In a "legal" move, any one of them can jump over another, and land on its other side at exactly the same distance. After a finite number of legal moves, can the grasshoppers end up at a point set, similar to, but larger than $P$? We present a linear algebraic approach to answer this question. In particular, we solve a problem of Brunck by showing that the answer is yes if $P$ is the vertex set of a regular $N$-gon and $N\neq 3, 4, 6$. Some generalizations are also considered.

math.CO↗

arXiv2023

Planar Point Sets Determine Many Pairwise Crossing Segments

We show that any set of $n$ points in general position in the plane determines $n^{1-o(1)}$ pairwise crossing segments. The best previously known lower bound, $Ω\left(\sqrt n\right)$, was proved more than 25 years ago by Aronov, Erd\H os, Goddard, Kleitman, Klugerman, Pach, and Schulman. Our proof is fully constructive, and extends to dense geometric graphs.

math.CO↗

arXiv2021

Random necklaces require fewer cuts

It is known that any open necklace with beads of $t$ types in which the number of beads of each type is divisible by $k$, can be partitioned by at most $(k-1)t$ cuts into intervals that can be distributed into $k$ collections, each containing the same number of beads of each type. This is tight for all values of $k$ and $t$. Here, we consider the case of random necklaces, where the number of beads of each type is $km$. Then the minimum number of cuts required for a ``fair'' partition with the above property is a random variable $X(k,t,m)$. We prove that for fixed $k,t,$ and large $m$, this random variable is at least $(k-1)(t+1)/2$ with high probability. For $k=2$, fixed $t$, and large $m$, we determine the asymptotic behavior of the probability that $X(2,t,m)=s$ for all values of $s\le t $. We show that this probability is polynomially small when $s<(t+1)/2$, it is bounded away from zero when $s>(t+1)/2$, and decays like $Θ( 1/\log m)$ when $s=(t+1)/2$. We also show that for large $t$, $X(2,t,1)$ is at most $(0.4+o(1))t$ with high probability and that for large $t$ and large ratio $k/\log t$, $X(k,t,1)$ is $o(kt)$ with high probability.

math.CO↗

arXiv2021

Disjointness graphs of short polygonal chains

The {\em disjointness graph} of a set system is a graph whose vertices are the sets, two being connected by an edge if and only if they are disjoint. It is known that the disjointness graph $G$ of any system of segments in the plane is {\em $χ$-bounded}, that is, its chromatic number $χ(G)$ is upper bounded by a function of its clique number $ω(G)$. Here we show that this statement does not remain true for systems of polygonal chains of length $2$. We also construct systems of polygonal chains of length $3$ such that their disjointness graphs have arbitrarily large girth and chromatic number. In the opposite direction, we show that the class of disjointness graphs of (possibly self-intersecting) \emph{$2$-way infinite} polygonal chains of length $3$ is $χ$-bounded: for every such graph $G$, we have $χ(G)\le(ω(G))^3+ω(G).$

math.CO↗

arXiv2021

Turán problems for Edge-ordered graphs

In this paper we initiate a systematic study of the Turán problem for edge-ordered graphs. A simple graph is called $\textit{edge-ordered}$, if its edges are linearly ordered. An isomorphism between edge-ordered graphs must respect the edge-order. A subgraph of an edge-ordered graph is itself an edge-ordered graph with the induced edge-order. We say that an edge-ordered graph $G$ $\textit{avoids}$ another edge-ordered graph $H$, if no subgraph of $G$ is isomorphic to $H$. The $\textit{Turán number}$ of an edge-ordered graph $H$ is the maximum number of edges in an edge-ordered graph on $n$ vertices that avoids $H$. We study this problem in general, and establish an Erdős-Stone-Simonovits-type theorem for edge-ordered graphs -- we discover that the relevant parameter for the Turán number of an edge-ordered graph is its $\textit{order chromatic number}$. We establish several important properties of this parameter. We also study Turán numbers of edge-ordered paths, star forests and the cycle of length four. We make strong connections to Davenport-Schinzel theory, the theory of forbidden submatrices, and show an application in Discrete Geometry.

math.CO↗

arXiv2021

Convergence and limits of finite trees

Motivated by the work of Lovász and Szegedy on the convergence and limits of dense graph sequences, we investigate the convergence and limits of finite trees with respect to sampling in normalized distance. Based on separable real trees, we introduce the notion of a dendron and show that the limits of finite trees are exactly the dendrons. We also prove that the limit dendron is unique.

math.CO↗

arXiv2021

Unlabeled Compression Schemes Exceeding the VC-dimension

In this note we disprove a conjecture of Kuzmin and Warmuth claiming that every family whose VC-dimension is at most d admits an unlabeled compression scheme to a sample of size at most d. We also study the unlabeled compression schemes of the joins of some families and conjecture that these give a larger gap between the VC-dimension and the size of the smallest unlabeled compression scheme for them.

math.CO↗

arXiv2020

Crossings between non-homotopic edges

We call a multigraph {\em non-homotopic} if it can be drawn in the plane in such a way that no two edges connecting the same pair of vertices can be continuously transformed into each other without passing through a vertex, and no loop can be shrunk to its end-vertex in the same way. It is easy to see that a non-homotopic multigraph on $n>1$ vertices can have arbitrarily many edges. We prove that the number of crossings between the edges of a non-homotopic multigraph with $n$ vertices and $m>4n$ edges is larger than $c\frac{m^2}{n}$ for some constant $c>0$, and that this bound is tight up to a polylogarithmic factor. We also show that the lower bound is not asymptotically sharp as $n$ is fixed and $m$ tends to infinity.

math.CO↗

arXiv2020

Two extensions of the Erdős-Szekeres problem

According to Suk's breakthrough result on the Erdos-Szekeres problem, any point set in general position in the plane, which has no $n$ elements that form the vertex set of a convex $n$-gon, has at most $2^{n+O\left({n^{2/3}\log n}\right)}$ points. We strengthen this theorem in two ways. First, we show that the result generalizes to convexity structures induced by pseudoline arrangements. Second, we improve the error term. A family of $n$ convex bodies in the plane is said to be in convex position if the convex hull of the union of no $n-1$ of its members contains the remaining one. If any three members are in convex position, we say that the family is in general position. Combining our results with a theorem of Dobbins, Holmsen, and Hubard, we significantly improve the best known upper bounds on the following two functions, introduced by Bisztriczky and Fejes Toth and by Pach and Toth, respectively. Let $c(n)$ (and $c'(n)$) denote the smallest positive integer $N$ with the property that any family of $N$ pairwise disjoint convex bodies in general position (resp., $N$ convex bodies in general position, any pair of which share at most two boundary points) has an $n$-membered subfamily in convex position. We show that $c(n)\le c'(n)\leq 2^{n+O\left(\sqrt{n\log n}\right)}$.

math.CO↗