arXiv++ Combinatorics

Let $H$ be an inclusion-free hypergraph on $n$ vertices. A weight assignment $w:[n]\to[d]$ is isolating if there is a unique edge $e$ whose weight $w(e) = \sum_{i \in e} w(i)$ is minimum. We show that the number of isolating weight assignments is at least $$ n\sum_{j=0}^{d-1} j^{n-1}, $$ a bound which is attained with equality by the hypergraph consisting of the $n$ singleton edges. This proves the conjecture stated in Faber & Harris (2018). We also prove the bound for a more general class of edge-weight objectives, including arbitrary edge offsets.

We study packet routing in the Kautz digraph K(d,D), where every ordered pair of distinct vertices is connected by a unique shortest directed path. The regular routing introduced in earlier work schedules all ordered pairs in tau(d,D) = (D-1)d^(D-2) + D d^(D-1) steps. We show that, for every fixed outdegree d at least 2 and all sufficiently large diameters D, no shortest-path routing scheme can match this makespan. More precisely, we prove that K(d,D) contains an edge whose shortest-path congestion strictly exceeds tau(d,D) when D is sufficiently large. Our construction uses edge-words drawn from a subset of ternary unbordered square-free words, together with a trimming inequality that propagates large congestion at distance D down to shorter distances. Computations for d=2 and small D show that for all D at least 4 there is an edge in K(2,D) with congestion greater than tau(2,D).

For the Kautz digraph $K(d,D)$, let $ρ_k(d,D)$ be the number of oriented edges whose shortest directed cycle has length $k+1$, and define $Δ_k(d,D) = ρ_k(d,D) - ρ_k(d,D-1)$. We give an exact, finite-dimensional matrix product that computes $Δ_k(d,D)$ directly, without first computing $ρ$. In particular, $Δ_k(d,D)=0$ for $k < D/2+2$. and $Δ_k(d,D)$ is positive for every larger $k$ up to $D-1$.

We investigate the existence of spanning 1-factorizations in vertex-transitive digraphs of out-degree d. The open question is whether every such digraph admits a spanning 1-factorization that includes, for each vertex v, all d out-edges (v,F_i(v)) from v. This paper focuses on the case d=2. Using the structure of alternating cycles and block systems, we develop a block/phase framework that yields sufficient conditions for including both F_1,F_2. We show that certain block obstructions can prevent their simultaneous inclusion, while sharply transitive sets (and hence spanning 1-factorizations) always exist. Our results provide general constraints on feasible block sizes, describe the role of phase distributions, and illustrate the theory with concrete families, including coset digraphs on A_5. The necessity of the block criterion remains open, even in degree 2.

Efficiency of routing on a regular digraph often involves finding opitmal properties of the graph. For example, the diameter of a digraph is the maximum distance between any two vertices. We show how we can study these problems algebraically in terms of quasigroups, 1-factors, and permutation groups. Our investigation originated from the study of graphs as the Cayley graphs of groupoids with $d$ generators, a left identity, and right cancellation; that is, a right quasigroup. This enables us to provide compact algebraic definitions for some important graphs that are either given as explicit edge lists or as the Cayley coset graphs of groups larger than the graph. One such example is a single expression for the Hoffman-Singleton graph. From there, we notice that the groupoids can be represented uniquely by a set of disjoint permutations and we explore the consequences of that observation.

This paper concerns all-to-all network routing on regular digraphs. In previous work we focused on efficient routing in highly symmetric digraphs with low diameter for fixed degree. Here, we show that every connected regular digraph has an all-to-all routing scheme and associated schedule with no waiting. In fact, this routing scheme becomes more efficient as the diameter goes down with respect to the degree and number of vertices. Lastly, we examine the simple scheduling algorithm called ``farthest-distance-first'' and prove that it yields optimal schedules for all-to-all communication in networks of interest, including Kautz graphs.

In 2020, a paper [arXiv:2010.13443] appeared in the arXiv claiming to prove that a Moore graph of diameter 2 and degree 57 does not exist. (The paper is in Russian; we include a link to a translation of this paper kindly provided to us by Konstantin Selivanov.) The proof technique is reasonable. It employs the fact that such a graph must be distance regular and that there exists a large set of relations which such a graph must satisfy. The argument proceeds by a case analysis that shows that this set of relations cannot be satisfied. We show that this seems not to be correct. The system of equations factors into small diagonal blocks all of which have solutions. As an alternative, we show that there is a family of systems of permutations with the property that the Moore graph exists if and only if there is a member of the family with no solutions.

We discuss an open problem and its converse first posed by Dougherty and Faber in [3], "Network routing on regular directed graphs from spanning factorizations." Does every vertex transitive digraph have a spanning 1=factorization? We show relationships between various properties a regular digraph might have: vertex transitivity, left or right cancellation, tree-like or neighborhood preserving spanning factorizations.

We consider spherical quadrangulations -- spherical embeddings of multigraphs, possibly with loops, so that every face has boundary walk of length 4 -- in which all vertices have degree 3 or 4. Interpreting each degree 4 vertex as a crossing, these embeddings can also be thought of as transversal immersions of cubic graphs which we refer to as the {\it extracted graphs}. We also consider quadrangulations of the disk in which interior vertices have degree 3 or 4 and boundary vertices have degree 2 or 3. First, we classify all such quadrangulations of the disk. Then, we provide four methods for constructing spherical quadrangulations, two of which use quadrangulations of the disk as input. Two of these methods provide one-parameter families of quadrangulations, for which we prove that the sequence of isomorphism types of extracted graphs is periodic. We close with a description of computer computations which yielded spherical quadrangulations for all but three cubic multigraphs on eight vertices.

The $k$-truss, introduced by Cohen (2005), is a graph where every edge is incident to at least $k$ triangles. This is a relaxation of the clique. It has proved to be a useful tool in identifying cohesive subnetworks in a variety of real-world graphs. Despite its simplicity and its utility, the combinatorial and algorithmic aspects of trusses have not been thoroughly explored. We provide nearly-tight bounds on the edge counts of $k$-trusses. We also give two improved algorithms for finding trusses in large-scale graphs. First, we present a simplified and faster algorithm, based on approach discussed in Wang & Cheng (2012). Second, we present a theoretical algorithm based on fast matrix multiplication; this converts a triangle-generation algorithm of Bjorklund et al. (2014) into a dynamic data structure.

Motivated by the Erdős-Faber-Lovász (EFL) conjecture for hypergraphs, we consider the list edge coloring of linear hypergraphs. We show that if the hyper-edge sizes are bounded between $i$ and $C_{i,ε} \sqrt{n}$ inclusive, then there is a list edge coloring using $(1 + ε) \frac{n}{i - 1}$ colors. The dependence on $n$ in the upper bound is optimal (up to the value of $C_{i,ε}$).

Motivated by the Erdős-Faber-Lovász (EFL) conjecture for hypergraphs, we consider the list edge coloring of linear hypergraphs. We discuss several conjectures for list edge coloring linear hypergraphs that generalize both EFL and Vizing's theorem for graphs. For example, we conjecture that in a linear hypergraph of rank 3, the list edge chromatic number is at most 2 times the maximum degree plus 1. We show that for sufficiently large fixed rank and sufficiently large degree, the conjectures are true.

Given a hypergraph $H$ and a weight function $w: V \rightarrow \{1, \dots, M\}$ on its vertices, we say that $w$ is isolating if there is exactly one edge of minimum weight $w(e) = \sum_{i \in e} w(i)$. The Isolation Lemma is a combinatorial principle introduced in Mulmuley et. al (1987) which gives a lower bound on the number of isolating weight functions. Mulmuley used this as the basis of a parallel algorithm for finding perfect graph matchings. It has a number of other applications to parallel algorithms and to reductions of general search problems to unique search problems (in which there are one or zero solutions). The original bound given by Mulmuley et al. was recently improved by Ta-Shma (2015). In this paper, we show improved lower bounds on the number of isolating weight functions, and we conjecture that the extremal case is when $H$ consists of $n$ singleton edges. When $M \gg n$ our improved bound matches this extremal case asymptotically. We are able to show that this conjecture holds in a number of special cases: when $H$ is a linear hypergraph or is 1-degenerate, or when $M = 2$. We also show that it holds asymptotically when $M \gg n \gg 1$.

Motivated by the Erdos-Faber Lovasz conjecture (EFL) for hypergraphs, we explore relationships between several conjectures on the edge coloring of linear hypergraphs. In particular, we are able to increase the class of hypergraphs for which EFL is true.

Let H be a hypergraph on n vertices with the property that no edge contains another. We prove some results for a special case of the Isolation Lemma when the label set for the edges of H can only take two values. Given any set of vertices S and an edge e, the weight of S in e is the size of e plus the size of the intersection of S and e. A versal S for an edge e is a set of vertices with weight in e smaller than the weight in any other edge. We show that H always has at least n + 1 versals except if H is either the set of all singletons T_n or the complement of T_n or the 4-cycle graph. In those exceptional cases there are only n versals.

Networks with a high degree of symmetry are useful models for parallel processor networks. In earlier papers, we defined several global communication tasks (universal exchange, universal broadcast, universal summation) that can be critical tasks when complex algorithms are mapped to parallel machines. We showed that utilizing the symmetry can make network optimization a tractable problem. In particular, we showed that Cayley graphs have the desirable property that certain routing schemes starting from a single node can be transferred to all nodes in a way that does not introduce conflicts. In this paper, we define the concept of spanning factorizations and show that this property can also be used to transfer routing schemes from a single node to all other nodes. We show that all Cayley graphs and many (perhaps all) vertex transitive graphs have spanning factorizations.

We discuss transpose (sometimes called universal exchange or all-to-all) on vertex symmetric networks. We provide a method to compare the efficiency of transpose schemes on two different networks with a cost function based on the number processors and wires needed to complete a given algorithm in a given time.

Cycle prefix digraphs have been proposed as an efficient model of symmetric interconnection networks for parallel architecture. It has been discovered that the cycle prefix networks have many attractive communication properties. In this paper, we determine the automorphism group of the cycle prefix digraphs. We show that the automorphism group of a cycle prefix digraph is isomorphic to the symmetric group on its underlying alphabet. Our method can be applied to other classes of graphs built on alphabets including the hypercube, the Kautz graph,and the de Bruijn graph.

Kelner, Orecchia, Sidford, and Zhu have given a randomized iterative method for approximating the solution to the discrete Laplace equation that has expected running time nearly linear in the size of the problem. The goal of this note is to explain this algorithm in the language of difference operators on graphs.

We discuss several combinatorial problems that arise when one looks at computational algorithms for highly symmetric networks of processors. More specifically, we are interested in minimal times associated with four communication tasks (defined more precisely below): universal broadcast, every processor has a vector that it wishes to broadcast to all the others; universal accumulation, every processor wishes to receive the sum of all the vectors being sent to it by all the other processors; universal exchange, every processor wishes to exchange a vector with each other processor; and global summation, every processor wants the sum of the vectors in all the processors