arXiv++ Combinatorics

The length polyhedron of an interval order $P$ is the convex hull of integer vectors representing the interval lengths in possible interval representations of $P$ in which all intervals have integer endpoints. This polyhedron is an integral translation of a polyhedral cone, with its apex corresponding to the canonical interval representation of $P$ (also known as the minimal endpoint representation). In earlier work, we introduced an arc-weighted directed graph model, termed the key graph, inspired by this canonical representation. We showed that cycles in the key graph correspond, via Fourier-Motzkin elimination, to inequalities that describe supporting hyperplanes of the length polyhedron. These cycle inequalities derived from the key graph form a complete system of linear inequalities defining the length polyhedron. By applying a theorem due to Cook, we establish here that this system of inequalities is totally dual integral (TDI). Leveraging circulations, total dual integrality, and the special structure of the key graph, our main theorem demonstrates that a cycle inequality is a positive linear combination of other cycle inequalities if and only if it is a positive integral linear combination of smaller cycle inequalities (where `smaller' here refers a natural weak ordering among these cycle inequalities). This yields an efficient method to remove redundant cycle inequalities and ultimately construct the unique minimal TDI-system, also known as the Schrijver system, for the length polyhedron. Notably, if the key graph contains a polynomial number of cycles, this gives a polynomial-time algorithm to compute the Schrijver system for the length polyhedron. Lastly, we provide examples of interval orders where the Schrijver system has an exponential size.

The interval count problem, a classical question in the study of interval orders, was introduced by Ronald Graham in the 1980s. This problem asks: given an interval order $P$, what is the minimum number of distinct interval lengths required to construct an interval representation of $P$? Interval orders that can be represented with just one interval length are known as semiorders, and their characterizations are well known. However, the characterization of interval orders that require at most $k$ interval lengths -- termed $k$-count interval orders -- remains an open and challenging problem for $k\geq 2$. Our investigation into $2$-count interval orders led us naturally to consider a related problem, interval representations of permutations, which we introduce in this paper. Specifically, we characterize permutations that have a $2$-count interval representation. We prove that a permutation admits a $2$-count interval representation if and only if its longest decreasing subsequences have length at most $2$. For larger values of $k$, however, a similar characterization does not hold. There are permutations that do not permit a $3$-count interval representation despite having decreasing subsequences of length at most $3$. Characterizing $k$-count permutations remains open for $k \geq 3$. The $k$-count permutation representation problem appears to capture essential aspects of the broader problem of characterizing $k$-count interval orders. To support this connection, we apply our findings on interval representations of permutations to demonstrate that a height-$3$ interval order is $2$-count if and only if it has depth at most $2$, where the depth of an interval order refers to the length of the longest nested chain of intervals required in any interval representation of the order.

The length polyhedron $Q_P$ of an interval order $P$ is the convex hull of integral vectors representing the interval lengths in interval representations of $P$. This polyhedron has been studied by various authors, including Fishburn and Isaak. Notably, $Q_P$ forms a pointed affine cone, a property inherited from being a projection of the representation polyhedron, a structure explored also by Doignon and Pauwels. The apex of the length polyhedron corresponds to Greenough's minimal endpoint representation, which is, in fact, the length vector of the canonical interval representation -- an interval representation that minimizes the sum of the interval lengths. Building on a combinatorial perspective of canonical representations, we refine Isaak's graph-theoretical model by introducing a new and simpler directed graph. From directed cycles of this key digraph, we extract a linear system of inequalities that precisely characterizes the length polyhedron $Q_P$. This combinatorial approach also reveals the unique Hilbert basis of the polyhedron. We prove that the intersection graph of the sets corresponding to these binary rays are Berge graphs; therefore they are perfect graphs. As a result, for interval orders with bounded width, the length polyhedron has a polynomial-sized Hilbert basis, which can be computed in polynomial time. We also provide an example of interval orders with a Hilbert basis of exponential size. In a companion paper we determine the Schrijver system for the length polyhedron. We conclude with open problems.

Let $H$ be a complete $r$-uniform hypergraph such that two vertices are marked in each edge as its `boundary' vertices. A linear ordering of the vertex set of $H$ is called an {\em agreeing linear order}, provided all vertices of each edge of $H$ lie between its two boundary vertices. We prove the following Helly-type theorem: if there is an {agreeing linear order} on the vertex set of every subhypergraph of $H$ with at most $2r-2$ vertices, then there is an agreeing linear order on the vertex set of $H$. We also show that the constant $2r-2$ cannot be reduced in the theorem. The case $r=3$ of the theorem has particular interest in the axiomatic theory of betweenness. Similar results are obtained for further $r$-uniform hypergraphs ($r\geq 3$), where one or two vertices are marked in each edge, and the linear orders need to satisfy various rules of agreement. In one of the cases we prove that no such Helly-type statement holds.

Consider a $3$-uniform hypergraph of order $n$ with clique number $k$ such that the intersection of all its $k$-cliques is empty. Szemerédi and Petruska proved $n\leq 8m^2+3m$, for fixed $m=n-k$, and they conjectured the sharp bound $n \leq {m+2 \choose 2}$. This problem is known to be equivalent to determining the maximum order of a $τ$-critical $3$-uniform hypergraph with transversal number $m$ (details may also be found in a companion paper: arXiv:2204.02859). The best known bound, $n\leq \frac{3}{4}m^2+m+1$, was obtained by Tuza using the machinery of $τ$-critical hypergraphs. Here we propose an alternative approach, a combination of the iterative decomposition process introduced by Szemerédi and Petruska with the skew version of Bollobás's theorem on set pair systems. The new approach improves the bound to $n\leq {m+2 \choose 2} + O(m^{{5}/{3}})$, resolving the conjecture asymptotically.

Recently we asymptotically resolved the long-standing Szemerédi and Petruska conjecture. Several decades ago Gyárfás et al. observed, via a straightforward but unpublished argument, that this conjecture is equivalent to the problem of determining the maximum order of a $3$-uniform $τ$-critical hypergraph. Consequently, an asymptotically tight upper bound for the maximum order of a $3$-uniform $τ$-critical hypergraph follows from our recent work, reawakening interest in this equivalence. In this companion paper we supply a simple proof of this equivalence. We also present related background with open problems, and mention combinatorial geometry applications of the Szemerédi and Petruska conjecture.

For $n\leq d$, a family ${\cal F}=\{C_0,C_1,\ldots, C_n\}$ of compact convex sets in $R^d$ is called an $n$-critical family provided any $n$ members of ${\cal F}$ have a non-empty intersection, but $\bigcap_{i=0}^n C_i=\varnothing$. If $n=d$ then a lemma on the intersection of convex sets due to Klee implies that the $d+1$ members of the $d$-critical family enclose a `hollow' in $R^d$, a bounded connected component of $R^d\setminus\bigcup_{i=0}^n C_i.$ Here we prove that the closure of the convex hull of a hollow in $R^d$ is a $d$-simplex.

Let H be a 3-uniform hypergraph of order n with clique number k such that the intersection of all maximum cliques of H is empty. For fixed m=n-k, Szemerédi and Petruska conjectured the sharp bound $n\leq {m+2\choose 2}$. In this note the conjecture is verified for m=2,3 and 4.

Given $2k-1$ convex sets in $R^2$ such that no point of the plane is covered by more than $k$ of the sets, is it true that there are two among the convex sets whose union contains all $k$-covered points of the plane? This question due to Gy. Petruska has an obvious affirmative answer for $k=1,2,3$; we show here that the claim is also true for $k=4$, and we present a counterexample for $k=5$. We explain how Petruska's geometry question fits into the classical hypergraph extremal problems, called arrow problems, proposed by P. Erdős.

Let $G=P_6\Box P_6$ be the $6\times 6$ grid, the Cartesian product of two paths of six vertices. Let $T$ be the set of eight distinct vertices of $G$, called terminals, and assume that $T$ is partitioned into four terminal pairs $\{s_i,t_i\}$, $1\leq i\leq 4$. We prove that $G$ is $4$-path-pairable, that is, for every $T$ there exist in $G$ pairwise edge disjoint $s_i,t_i$-paths, $1\leq i\leq 4$.

Let $G$ be the Cartesian product of two finite paths, called a grid, and let $T$ be the set of eight distinct vertices of $G$, called terminals. Assume that $T$ is partitioned into four terminal pairs $\{s_i,t_i\}$, $1\leq i\leq 4$, to be linked in $G$ by using edge disjoint paths. To prove that such a linkage always exists we need a sequence of technical lemmas making possible for some terminals to `escape' from a $3\times 3$ corner of $Q\subset G$, called a `quadrant'. Here we state those lemmas, and give a proof for the cases when $Q$ contains at most $4$ terminals.

Let $Q$ be a finite subgraph of the integer grid $G$ in the plane, and let $T$ be a set of pairs of distinct vertices in $G$, called `terminal pairs'. Escaping a subset $X\subset T\cap Q$ from $Q$ means finding edge disjoint paths from the terminals in $X$ into distinct vertices of a set $L$ in the boundary of $Q$. Here we prove several lemmas for the cases where $Q$ is a $3\times 3$ grid, $L$ is the union of a vertical and horizontal boundary line of $Q$, furthermore, $T$ is a set of four terminal pairs in $G$ such that $|T\cap Q|\geq 5$. These lemmas are applied in [4] and complete the proof that the Cartesian product of two (one way) infinite paths has path-pairability number four.

Motivated by the problem of Gallai on $(1-1)$-transversals of $2$-intervals, it was proved by the authors in 1969 that if the edges of a complete graph $K$ are colored with red and blue (both colors can appear on an edge) so that there is no monochromatic induced $C_4$ and $C_5$ then the vertices of $K$ can be partitioned into a red and a blue clique. Aharoni, Berger, Chudnovsky and Ziani recently strengthened this by showing that it is enough to assume that there is no induced monochromatic $C_4$ and there is no induced $C_5$ in {\em one of the colors}. Here this is strengthened further, it is enough to assume that there is no monochromatic induced $C_4$ and there is no $K_5$ on which both color classes induce a $C_5$. We also answer a question of Kaiser and Rabinovich, giving an example of six $2$-convex sets in the plane such that any three intersect but there is no $(1-1)$-transversal for them.