arXiv++ Combinatorics

A set in $\mathbb{R}^d$ is $s$-convex if it is the union of at most $s$ convex sets. A family $F$ satisfies the $(p,q)$ property if among any $p$ sets in $F$, some $q$ intersect. Let $\mathrm{HD}_d^{(s)}(p,q)$ be the minimum number of points needed to pierce a finite family of $s$-convex sets that satisfies the $(p,q)$-property. Alon and Kalai (1995) proved that $\mathrm{HD}_d^{(s)}(p,q)$ exists for any $p \geq q \geq d+1$ and any $s \geq 1$, but the quantitative bounds they obtained are very loose. We present several improved upper and lower bounds, for a general $d$ and for $s$-intervals of the line (i.e., $\mathrm{HD}_1^{(s)}(p,q)$). In particular, we prove the following: (i) For every $d\ge2$, $s \geq 1$ and $δ>0$, if $p>q$ and $q\ge C_d\log(e sp)$, then $\mathrm{HD}_d^{(s)}(p,q) \le p-q+1 + O_{d,δ}((s \cdot \tfrac{p}{q} \cdot \log \tfrac{esp}{q})^{ρ_d+δ}),$ where $ρ_d<d$ is the exponent in the weak epsilon-net theorem of Rubin (2022). (ii) For $s \geq 1$, $p \geq q \geq 2$ and $q\ge C_0s\log(2s)\log(ep)$, $p-q+s \leq \mathrm{HD}_1^{(s)}(p,q) \leq p-q+2s+1$. This result provides the first near-tight estimate for $\mathrm{HD}_1^{(s)}(p,q)$ for $q>2$. (iii) For any fixed $s$, there are an integer $κ_s\in\{s,\ldots,2s\}$ and constants $C_s,p_s>0$ such that, whenever $p\ge p_s$ and $q\ge C_s\log(ep)$, $ \mathrm{HD}_1^{(s)}(p,q)\in\{p-q+κ_s,\;p-q+κ_s+1\}. $ Interestingly, this two-value concentration result holds, although the exact value of the threshold remains unknown. (iv) For any $s \geq 1$, $\mathrm{HD}_3^{(s)}(p,4) \geq sp^{2-o(1)}$. Already for families of convex sets, this significantly improves the best known lower bound on $\mathrm{HD}_d^{(1)}(p,d+1)$, for all $d \geq 3$.

The VC-dimension is a fundamental measure of the complexity of a set system. In this paper, we introduce and study a colorful variant of VC-dimension that captures the behavior of set systems on colored ground sets. By studying this new notion, we obtain a variety of geometric results. First, we prove that separable abstract convexity spaces with Radon number $D$ admit a Tverberg theorem with Tverberg number $O(D^2 r \log r)$. This bound significantly improves the $O(Dr^2\log r)$ bound of Alon and Smorodinsky from SODA'26 and is the first quasi-linear bound in $r$, in which the dependence on $D$ is not super-exponential. Second, we prove the first colorful $k$-wise Tverberg theorem for separable abstract convexity spaces. Using this theorem, we obtain a colorful selection lemma with $O(D^3)$ colors, an uncolored selection lemma for subsets of size $O(D^3)$, a weak $\varepsilon$-net theorem with nets of size $O_D(\varepsilon^{-O(D^3)})$, and a $(p,q)$-theorem with exponent of $\mathrm{poly}(D)$. All these quantitative bounds are significantly better than the best previously known general bounds for abstract convexity spaces. Finally, we extend our method to obtain a colorful Tverberg theorem for unions of convex sets, generalizing the uncolored theorem of Alon and Smorodinsky (SODA'26).

The convexity number of a set $X \subset \mathbb{R}^2$ is the minimum number of convex subsets required to cover it. We study the following question: what is the largest possible convexity number $f(n)$ of $\mathbb{R}^2 \setminus S$, where $S$ is a set of $n$ points in general position in the plane? We prove that for all $n \geq 4$, $\lfloor\frac{n+5}{2}\rfloor \leq f(n) \leq \frac{7n+44}{11}$. We also show that for every $n \geq 4$, if the points of $S$ are in convex position then the convexity number of $\mathbb{R}^2 \setminus S$ is $\lfloor\frac{n+5}{2}\rfloor$. This solves a problem of Lawrence and Morris [Finite sets as complements of finite unions of convex sets, Disc. Comput. Geom. 42 (2009), 206-218].

Finite unions of convex sets are a central object of study in discrete and computational geometry. In this paper we initiate a systematic study of complements of such unions -- i.e., sets of the form $S=\mathbb{R}^d \setminus (\cup_{i=1}^n K_i)$, where $K_i$ are convex sets. In the first part of the paper we study isolated points in $S$, whose number is related to the Betti numbers of $\cup_{i=1}^n K_i$ and to its non-convexity properties. We obtain upper bounds on the number of such points, which are sharp for $n=3$ and significantly improve previous bounds of Lawrence and Morris (2009) for all $n \ll \frac{2^d}{d}$. In the second part of the paper we study coverings of $S$ by well-behaved sets. We show that $S$ can be covered by at most $g(d,n)$ flats of different dimensions, in such a way that each $x \in S$ is covered by a flat whose dimension equals the `local dimension' of $S$ in the neighborhood of $x$. Furthermore, we determine the structure of a minimum cover that satisfies this property. Then, we study quantitative aspects of this minimum cover and obtain sharp upper bounds on its size in various settings.

A convex geometric graph is a graph whose vertices are the corners of a convex polygon P in the plane and whose edges are boundary edges and diagonals of the polygon. It is called triangulation-free if its non-boundary edges do not contain the set of diagonals of some triangulation of P. Aichholzer et al. (2010) showed that the maximum number of edges in a triangulation-free convex geometric graph on n vertices is ${{n}\choose{2}}-(n-2)$, and subsequently, Keller and Stein (2020) and (independently) Ali et al. (2022) characterized the triangulation-free graphs with this maximum number of edges. We initiate the study of the saturation version of the problem, namely, characterizing the triangulation-free convex geometric graphs which are not of the maximum possible size, but yet the addition of any edge to them results in containing a triangulation. We show that, surprisingly, there exist saturated graphs with only g(n) = O(n log n) edges. Furthermore, we prove that for any $n > n_0$ and any $g(n)\leq t \leq {{n}\choose{2}}-(n-2)$, there exists a saturated graph with n vertices and t edges. In addition, we obtain a complete characterization of all saturated graphs whose number of edges is ${{n}\choose{2}}-(n-1)$, which is 1 less than the maximum.

An $ε$-net theorem for a hypergraph upper bounds the minimum size of a vertex set that pierces all $ε$-heavy hyperedges. A $(p,2)$-theorem bounds from above the minimum size of a vertex set that pierces all hyperedges, in terms of the maximum size of a set of pairwise disjoint hyperedges. Numerous works studied $ε$-net theorems and $(p,2)$-theorems that guarantee the existence of small-sized piercing sets. We focus on the question: In which settings the asymptotically smallest possible piercing sets -- i.e., $ε$-nets of size $O(\frac{1}ε)$ and piercing sets of size $O(p)$ in $(p,2)$-theorems, are guaranteed? We obtain several sufficient criteria for the existence of such linear $ε$-net theorems and $(p,2)$-theorems that unveil interesting connections to graph theory and improve and generalize several previous results. Most notably, we exhibit an unexpected relation of $ε$-nets to the classical Zarankiewicz's problem in graph theory. We show that a linear bound in the Zarankiewicz-type problem that asks for the maximum size of a bipartite graph with no copy of $K_{2,t}$, implies a linear $ε$-net theorem for the corresponding neighborhood hypergraph. We also show that hypergraphs with a hereditarily linear-sized Delaunay graph admit an almost linear $(p,2)$-theorem, and deduce that incidence hypergraphs of non-piercing regions in the plane admit a linear $(p,2)$-theorem, significantly improving previous results on such hypergraphs. Our work presents a landscape of sufficient conditions for the existence of linear $ε$-net theorems and $(p,2)$-theorems, with complex interrelations between them. Many of the interrelations are still unknown and call for future research.

The hypergraph Zarankiewicz's problem, introduced by Erdős in 1964, asks for the maximum number of hyperedges in an $r$-partite hypergraph with $n$ vertices in each part that does not contain a copy of $K_{t,t,\ldots,t}$. Erdős obtained a near optimal bound of $O(n^{r-1/t^{r-1}})$ for general hypergraphs. In recent years, several works obtained improved bounds under various algebraic assumptions -- e.g., if the hypergraph is semialgebraic. In this paper we study the problem in a geometric setting -- for $r$-partite intersection hypergraphs of families of geometric objects. Our main results are essentially sharp bounds for families of axis-parallel boxes in $\mathbb{R}^d$ and families of pseudo-discs. For axis-parallel boxes, we obtain the sharp bound $O_{d,r}(tn^{r-1}(\frac{\log n}{\log \log n})^{d-1})$. The best previous bound was larger by a factor of about $(\log n)^{d(2^{r-1}-2)}$. For pseudo-discs, we obtain the bound $O_r(tn^{r-1}(\log n)^{r-2})$, which is sharp up to logarithmic factors. As this hypergraph has no algebraic structure, no improvement of Erdős' 60-year-old $O(n^{r-1/t^{r-1}})$ bound was known for this setting. Futhermore, even in the special case of discs for which the semialgebraic structure can be used, our result improves the best known result by a factor of $\tildeΩ(n^{\frac{2r-2}{3r-2}})$. To obtain our results, we use the recently improved results for the graph Zarankiewicz's problem in the corresponding settings, along with a variety of combinatorial and geometric techniques, including shallow cuttings, biclique covers, transversals, and planarity.

The classical Zarankiewicz's problem asks for the maximum number of edges in a bipartite graph on $n$ vertices which does not contain the complete bipartite graph $K_{t,t}$. In one of the cornerstones of extremal graph theory, Kővári Sós and Turán proved an upper bound of $O(n^{2-\frac{1}{t}})$. In a celebrated result, Fox et al. obtained an improved bound of $O(n^{2-\frac{1}{d}})$ for graphs of VC-dimension $d$ (where $d<t$). Basit, Chernikov, Starchenko, Tao and Tran improved the bound for the case of semilinear graphs. At SODA'23, Chan and Har-Peled further improved Basit et al.'s bounds and presented (quasi-)linear upper bounds for several classes of geometrically-defined incidence graphs, including a bound of $O(n \log \log n)$ for the incidence graph of points and pseudo-discs in the plane. In this paper we present a new approach to Zarankiewicz's problem, via $ε$-t-nets - a recently introduced generalization of the classical notion of $ε$-nets. We show that the existence of `small'-sized $ε$-t-nets implies upper bounds for Zarankiewicz's problem. Using the new approach, we obtain a sharp bound of $O(n)$ for the intersection graph of two families of pseudo-discs, thus both improving and generalizing the result of Chan and Har-Peled from incidence graphs to intersection graphs. We also obtain a short proof of the $O(n^{2-\frac{1}{d}})$ bound of Fox et al., and show improved bounds for several other classes of geometric intersection graphs, including a sharp $O(n\frac{\log n}{\log \log n})$ bound for the intersection graph of two families of axis-parallel rectangles.

A family $\mathcal{F}$ of sets satisfies the $(p,q)$-property if among every $p$ members of $\mathcal{F}$, some $q$ can be pierced by a single point. The celebrated $(p,q)$-theorem of Alon and Kleitman asserts that for any $p \geq q \geq d+1$, any family $\mathcal{F}$ of compact convex sets in $\mathbb{R}^d$ that satisfies the $(p,q)$-property can be pierced by a finite number $c(p,q,d)$ of points. A similar theorem with respect to piercing by $(d-1)$-dimensional flats, called $(d-1)$-transversals, was obtained by Alon and Kalai. In this paper we prove the following result, which can be viewed as an $(\aleph_0,k+2)$-theorem with respect to $k$-transversals: Let $\mathcal{F}$ be an infinite family of closed balls in $\mathbb{R}^d$, and let $0 \leq k < d$. If among every $\aleph_0$ elements of $\mathcal{F}$, some $k+2$ can be pierced by a $k$-dimensional flat, then $\mathcal{F}$ can be pierced by a finite number of $k$-dimensional flats. We derive this result as a corollary of a more general result which proves the same assertion for families of not necessarily convex objects called \emph{near-balls}, to be defined below. This is the first $(p,q)$-theorem in which the assumption is weakened to an $(\infty,\cdot)$ assumption. Our proofs combine geometric and topological tools.

We introduce and study conflict-free colourings of $t$-subsets in hypergraphs. In such colourings, one assigns colours to all subsets of vertices of cardinality $t$ such that in any hyperedge of cardinality at least $t$ there is a uniquely coloured $t$-subset. The case $t=1$, i.e., vertex conflict-free colouring, is a well-studied notion. Already the case $t=2$ (i.e., colouring pairs) seems to present a new challenge. Many of the tools used for conflict-free colouring of geometric hypergraphs rely on hereditary properties of the underlying hypergraphs. When dealing with subsets of vertices, the properties do not pass to subfamilies of subsets. Therefore, we develop new tools, which might be of independent interest. (i) For any fixed $t$, we show that the $\binom n t$ $t$-subsets in any set $P$ of $n$ points in the plane can be coloured with $O(t^2 \log^2 n)$ colours so that any axis-parallel rectangle that contains at least $t$ points of $P$ also contains a uniquely coloured $t$-subset. (ii) For a wide class of "well behaved" geometrically defined hypergraphs, we provide near tight upper bounds on their $t$-subset conflict-free chromatic number. For $t=2$ we show that for each of those "well -behaved" hypergraphs $H$, the hypergraph $H'$ obtained by taking union of two hyperedges from $H$, admits a $2$-subset conflict-free colouring with roughly the same number of colours as $H$. For example, we show that the $\binom n 2$ pairs of points in any set $P$ of $n$ points in the plane can be coloured with $O(\log n)$ colours such that for any two discs $d_1,d_2$ in the plane with $|(d_1\cup d_2)\cap P|\geq 2$ there is a uniquely (in $d_1 \cup d_2$) coloured pair. (iii) We also show that there is no general bound on the $t$-subset conflict-free chromatic number as a function of the standard conflict-free chromatic number already for $t=2$.

We construct families of circles in the plane such that their tangency graphs have arbitrarily large girth and chromatic number. This provides a strong negative answer to Ringel's circle problem (1959). The proof relies on a (multidimensional) version of Gallai's theorem with polynomial constraints, which we derive from the Hales-Jewett theorem and which may be of independent interest.

For $n\geq s> r\geq 1$ and $k\geq 2$, write $n \rightarrow (s)_{k}^r$ if every hyperedge colouring with $k$ colours of the complete $r$-uniform hypergraph on $n$ vertices has a monochromatic subset of size $s$. Improving upon previous results by \textcite{AGLM14} and \textcite{EHMR84} we show that \[ \text{if } r \geq 3 \text{ and } n \nrightarrow (s)_k^r \text{ then } 2^n \nrightarrow (s+1)_{k+3}^{r+1}. \] This yields an improvement for some of the known lower bounds on multicolour hypergraph Ramsey numbers. Given a hypergraph $H=(V,E)$, we consider the Ramsey-like problem of colouring all $r$-subsets of $V$ such that no hyperedge of size $\geq r+1$ is monochromatic. We provide upper and lower bounds on the number of colours necessary in terms of the chromatic number $χ(H)$. In particular we show that this number is $O(\log^{(r-1)} (r χ(H)) + r)$.

Consider the hypergraph whose vertex set is a family of $n$ lines in general position in the plane, and whose hyperedges are induced by intersections with a family of pseudo-discs. We prove that the number of $t$-hyperedges is bounded by $O_t(n^2)$ and that the total number of hyperedges is bounded by $O(n^3)$. Both bounds are tight.

For a family $\mathcal{F}$ of sets in $\mathbb{R}^d$, the Krasnoselskii number of $\mathcal{F}$ is the smallest $m$ such that for any $S \in \mathcal{F}$, if every $m$ points of $S$ are visible from a common point in $S$, then any finite subset of $S$ is visible from a single point. More than 35 years ago, Peterson asked whether there exists a Krasnoselskii number for general sets in $\mathbb{R}^d$. Excluding results for special cases of sets with strong topological restrictions, the best known result is due to Breen, who showed that if such a Krasnoselskii number in $\mathbb{R}^2$ exists, then it is larger than $8$. In this paper we answer Peterson's question in the negative by showing that there is no Krasnoselskii number for the family of all sets in $\mathbb{R}^2$. The proof is non-constructive, and uses transfinite induction and the well ordering theorem. In addition, we consider Krasnoselskii numbers with respect to visibility through polygonal paths of length $ \leq n$, for which an analogue of Krasnoselskii's theorem was proved by Magazanik and Perles. We show, by an explicit construction, that for any $n \geq 2$, there is no Krasnoselskii number for the family of general sets in $\mathbb{R}^2$ with respect to visibility through paths of length $\leq n$. (Here the counterexamples are finite unions of line segments.)

We study a natural generalization of the classical $ε$-net problem (Haussler--Welzl 1987), which we call the "$ε$-$t$-net problem": Given a hypergraph on $n$ vertices and parameters $t$ and $ε\geq \frac t n$, find a minimum-sized family $S$ of $t$-element subsets of vertices such that each hyperedge of size at least $εn$ contains a set in $S$. When $t=1$, this corresponds to the $ε$-net problem. We prove that any sufficiently large hypergraph with VC-dimension $d$ admits an $ε$-$t$-net of size $O(\frac{ (1+\log t)d}ε \log \frac{1}ε)$. For some families of geometrically-defined hypergraphs (such as the dual hypergraph of regions with linear union complexity), we prove the existence of $O(\frac{1}ε)$-sized $ε$-$t$-nets. We also present an explicit construction of $ε$-$t$-nets (including $ε$-nets) for hypergraphs with bounded VC-dimension. In comparison to previous constructions for the special case of $ε$-nets (i.e., for $t=1$), it does not rely on advanced derandomization techniques. To this end we introduce a variant of the notion of VC-dimension which is of independent interest.

Let $P$ be a set of $n$ points in general position in the plane. Let $R$ be a set of $n$ points disjoint from $P$ such that for every $x,y \in P$ the line through $x$ and $y$ contains a point in $R$ outside of the segment delimited by $x$ and $y$. We show that $P \cup R$ must be contained in a cubic curve.

A family of sets $F$ is said to satisfy the $(p,q)$ property if among any $p$ sets in $F$, some $q$ have a non-empty intersection. Hadwiger and Debrunner (1957) conjectured that for any $p \geq q \geq d+1$ there exists $c=c_d(p,q)$, such that any family of compact convex sets in $\mathbb{R}^d$ that satisfies the $(p,q)$ property, can be pierced by at most $c$ points. In a celebrated result from 1992, Alon and Kleitman proved the conjecture. However, obtaining sharp bounds on $c_d(p,q)$, called `the Hadwiger-Debrunner numbers', is still a major open problem in discrete and computational geometry. The best currently known lower bound on the Hadwiger-Debrunner numbers in the plane is $c_2(p,q) = Ω( \frac{p}{q}\log(\frac{p}{q}))$ while the best known upper bound is $O(p^{(1.5+δ)(1+\frac{1}{q-2})})$. In this paper we improve the lower bound significantly by showing that $c_2(p,q) \geq p^{1+Ω(1/q)}$. Furthermore, the bound is obtained by a family of lines, and is tight for all families that have a bounded VC-dimension. Unlike previous bounds on the Hadwiger-Debrunner numbers which mainly used the weak epsilon-net theorem, our bound stems from a surprising connection of the $(p,q)$ problem to an old problem of Erdős on points in general position in the plane. We use a novel construction for the Erdős' problem, obtained recently by Balogh and Solymosi using the hypergraph container method, to get the lower bound on $c_2(p,3)$. We then generalize the bound to $c_2(p,q)$ for any $q \geq 3$.

Let G be a complete convex geometric graph, and let F be a family of subgraphs of G. A blocker for F is a set of edges, of smallest possible size, that has an edge in common with every element of F. In [C. Keller and M. A. Perles, Blockers for simple Hamiltonian paths in convex geometric graphs of even order, Disc. Comput. Geom., 60(1) (2018), pp. 1-8] we gave an explicit description of all blockers for the family of simple (i.e., non-crossing) Hamiltonian paths (SHPs) in G in the `even' case |V(G)|=2m. It turned out that all the blockers are simple caterpillar trees of a certain class. In this paper we give an explicit description of all blockers for the family of SHPs in the `odd' case |V(G)|=2m-1. In this case, the structure of the blockers is more complex, and in particular, they are not necessarily simple. Correspondingly, the proof is more complicated.

Let $G$ be a complete convex geometric graph whose vertex set $P$ forms a convex polygon $C$, and let $F$ be a family of subgraphs of $G$. A blocker for $F$ is a set of edges, of smallest possible size, that contains a common edge with every element of $F$. Previous works determined the blockers for various families $F$ of non-crossing subgraphs, including the families of all perfect matchings, all spanning trees, all Hamiltonian paths, etc. In this paper we present a complete characterization of the family $B$ of blockers for the family $T$ of triangulations of $C$. In particular, we show that $|B|=F_{2n-8}$, where $F_k$ is the $k$'th element in the Fibonacci sequence and $n=|P|$. We use our characterization to obtain a tight result on a geometric Maker-Breaker game in which the board is the set of diagonals of a convex $n$-gon $C$ and Maker seeks to occupy a triangulation of $C$. Namely, we show that in the $(1:1)$ triangulation game, Maker can ensure a win within $n-3$ moves, and that in the $(1:2)$ triangulation game, Breaker can ensure a win within $n-3$ moves. In particular, the threshold bias for the game is $2$.

Let $P$ be a set of $n \geq 5$ points in convex position in the plane. The path graph $G(P)$ of $P$ is an abstract graph whose vertices are non-crossing spanning paths of $P$, such that two paths are adjacent if one can be obtained from the other by deleting an edge and adding another edge. We prove that the automorphism group of $G(P)$ is isomorphic to $D_{n}$, the dihedral group of order $2n$. The heart of the proof is an algorithm that first identifies the vertices of $G(P)$ that correspond to boundary paths of $P$, where the identification is unique up to an automorphism of $K(P)$ as a geometric graph, and then identifies (uniquely) all edges of each path represented by a vertex of $G(P)$. The complexity of the algorithm is $O(N \log N)$ where $N$ is the number of vertices of $G(P)$.