arXiv++ Combinatorics

Let $t(n)$ denote the minimum number of transversals in a Latin square of odd order $n$. Improving upon a recent bound of Dai, Divoux and Kelly, we prove that for every $n$ such that $n \equiv 3 \pmod 6$, \[ t(n) \leq \left( \left(1+o(1)\right) \frac{2n}{3e^2}\right)^n . \] Our proof is based on a family of $3 \times 3$ block Latin squares whose transversals are constrained to either lie entirely in the diagonal blocks or avoid them altogether.

The $n\times n$ doubly stochastic matrices constitute a polytope in $\mathbb{R}^{n^2}$, and by Birkhoff's theorem, its vertex set coincides with the set of order-$n$ permutation matrices.\\ A tristochastic array is an $n \times n\times n$ array of nonnegative reals, where each row, column, and shaft sums to one. These arrays constitute a polytope $Δ_n$ in $\mathbb{R}^{n^3}$. In analogy, it is easy to see that each of the $L_n$ order-$n$ Latin squares is a vertex of $Δ_n$, but in contrast to Birkhoff's theorem, Latin squares form a vanishingly small subset of $Δ_n$'s vertex set. We show here that $Δ_n$ has at least $L_n^{2-o(1)}$ vertices.

The $n$-queens puzzle is to place $n$ mutually non-attacking queens on an $n \times n$ chessboard. We present a simple two stage randomized algorithm to construct such configurations. In the first stage, a random greedy algorithm constructs an approximate \textit{toroidal} $n$-queens configuration. In this well-known variant the diagonals wrap around the board from left to right and from top to bottom. We show that with high probability this algorithm succeeds in placing $(1-o(1))n$ queens on the board. In the second stage, the method of absorbers is used to obtain a complete solution to the non-toroidal problem. By counting the number of choices available at each step of the random greedy algorithm we conclude that there are more than $\left( \left( 1 - o(1) \right) n e^{-3} \right)^n$ solutions to the $n$-queens problem. This proves a conjecture of Rivin, Vardi, and Zimmerman in a strong form.

The fundamental group of the $2$-dimensional Linial-Meshulam random simplicial complex $Y_2(n,p)$ was first studied by Babson, Hoffman and Kahle. They proved that the threshold probability for simple connectivity of $Y_2(n,p)$ is about $p\approx n^{-1/2}$. In this paper, we show that this threshold probability is at most $p\le (γn)^{-1/2}$, where $γ= 4^4/3^3$, and conjecture that this threshold is sharp. In fact, we show that $p=(γn)^{-1/2}$ is a sharp threshold probability for the stronger property that every cycle of length $3$ is the boundary of a subcomplex of $Y_2(n,p)$ that is homeomorphic to a disk. Our proof uses the Poisson paradigm, and relies on a classical result of Tutte on the enumeration of planar triangulations.

Consider the random process in which the edges of a graph $G$ are added one by one in a random order. A classical result states that if $G$ is the complete graph $K_{2n}$ or the complete bipartite graph $K_{n,n}$, then typically a perfect matching appears at the moment at which the last isolated vertex disappears. We extend this result to arbitrary $k$-regular bipartite graphs $G$ on $2n$ vertices for all $k = ω\left( \frac{n}{\log^{1/3} n} \right)$. Surprisingly, this is not the case for smaller values of $k$. Using a construction due to Goel, Kapralov and Khanna, we show that there exist bipartite $k$-regular graphs in which the last isolated vertex disappears long before a perfect matching appears.

Let $m \leq n \leq k$. An $m \times n \times k$ 0-1 array is a Latin box if it contains exactly $mn$ ones, and has at most one $1$ in each line. As a special case, Latin boxes in which $m = n = k$ are equivalent to Latin squares. Let $\mathcal{M}(m,n,k;p)$ be the distribution on $m \times n \times k$ 0-1 arrays where each entry is $1$ with probability $p$, independently of the other entries. The threshold question for Latin squares asks when $\mathcal{M}(n,n,n;p)$ contains a Latin square with high probability. More generally, when does $\mathcal{M}(m,n,k;p)$ support a Latin box with high probability? Let $\varepsilon>0$. We give an asymptotically tight answer to this question in the special cases where $n=k$ and $m \leq \left(1-\varepsilon \right) n$, and where $n=m$ and $k \geq \left(1+\varepsilon \right) n$. In both cases, the threshold probability is $Θ\left( \log \left( n \right) / n \right)$. This implies threshold results for Latin rectangles and proper edge-colorings of $K_{n,n}$.

In how many ways can $n$ queens be placed on an $n \times n$ chessboard so that no two queens attack each other? This is the famous $n$-queens problem. Let $Q(n)$ denote the number of such configurations, and let $T(n)$ be the number of configurations on a toroidal chessboard. We show that for every $n$ of the form $4^k+1$, $T(n)$ and $Q(n)$ are both at least $n^{Ω(n)}$. This result confirms a conjecture of Rivin, Vardi and Zimmerman for these values of $n$. We also present new upper bounds on $T(n)$ and $Q(n)$ using the entropy method, and conjecture that in the case of $T(n)$ the bound is asymptotically tight. Along the way, we prove an upper bound on the number of perfect matchings in regular hypergraphs, which may be of independent interest.

A Hamiltonian cycle in a graph is a spanning subgraph that is homeomorphic to a circle. With this in mind, it is natural to define a Hamiltonian d-sphere in a d-dimensional simplicial complex as a spanning subcomplex that is homeomorphic to a d-dimensional sphere. We consider the Linial-Meshulam model for random simplicial complexes, and prove that there is a sharp threshold at $p=\sqrt{\frac{e}{γn}}$ for the appearance of a Hamiltonian $2$-sphere in a random $2$-complex, where $γ= 4^4/3^3$.

The theory of $k$-regular graphs is closely related to group theory. Every $k$-regular, bipartite graph is a Schreier graph with respect to some group $G$, a set of generators $S$ (depending only on $k$) and a subgroup $H$. The goal of this paper is to begin to develop such a framework for $k$-regular simplicial complexes of general dimension $d$. Our approach does not directly generalize the concept of a Schreier graph, but still presents an extensive family of $k$-regular simplicial complexes as quotients of one universal object: the $k$-regular $d$-dimensional arboreal complex, which is itself a simplicial complex originating in one specific group depending only on $d$ and $k$. Along the way we answer a question from [PR12] on the spectral gap of higher dimensional Laplacians and prove a high dimensional analogue of Leighton's graph covering theorem. This approach also suggests a random model for $k$-regular $d$-dimensional multicomplexes.

We introduce a new model of random $d$-dimensional simplicial complexes, for $d\geq 2$, whose $(d-1)$-cells have bounded degrees. We show that with high probability, complexes sampled according to this model are coboundary expanders. The construction relies on Keevash's recent result on designs [Ke14], and the proof of the expansion uses techniques developed by Evra and Kaufman in [EK15]. This gives a full solution to a question raised in [DK12], which was solved in the two-dimensional case by Lubotzky and Meshulam [LM13].

A Hamilton cycle in a graph $Γ$ is a cycle passing through every vertex of $Γ$. A Hamiltonian decomposition of $Γ$ is a partition of its edge set into disjoint Hamilton cycles. One of the oldest results in graph theory is Walecki's theorem from the 19th century, showing that a complete graph $K_n$ on an odd number of vertices $n$ has a Hamiltonian decomposition. This result was recently greatly extended by Kühn and Osthus. They proved that every $r$-regular $n$-vertex graph $Γ$ with even degree $r=cn$ for some fixed $c>1/2$ has a Hamiltonian decomposition, provided $n=n(c)$ is sufficiently large. In this paper we address the natural question of estimating $H(Γ)$, the number of such decompositions of $Γ$. Our main result is that $H(Γ)=r^{(1+o(1))nr/2}$. In particular, the number of Hamiltonian decompositions of $K_n$ is $n^{(1-o(1))n^2/2}$.

Let $L$ be an order-$n$ Latin square. For $X, Y, Z \subseteq \{1, ... ,n\}$, let $L(X, Y. Z)$ be the number of triples $i\in X, j\in Y, k\in Z$ such that $L(i,j) = k$. We conjecture that asymptotically almost every Latin square satisfies $|L(X, Y, Z) - \frac 1n |X||Y||Z||\le O(\sqrt{|X||Y||Z|})$ for every $X, Y$ and $Z$. Let $\varepsilon(L):= \max |X||Y||Z|$ when $L(X, Y, Z)=0$. The above conjecture implies that $\varepsilon(L) \le O(n^2)$ holds asymptotically almost surely (this bound is obviously tight). We show that there exist Latin squares with $\varepsilon(L) \le O(n^2)$, and that $\varepsilon(L) \le O(n^2 \log^2 n)$ for almost every order-$n$ Latin square. On the other hand, we recall that $\varepsilon(L)\geq Ω(n^{33/14})$ if $L$ is the multiplication table of an order-$n$ group. Some of these results extend to higher dimensions. Many open problems remain.

Let $T(n)$ denote the maximal number of transversals in an order-$n$ Latin square. Improving on the bounds obtained by McKay et al., Taranenko recently proved that $T(n) \leq \left((1+o(1))\frac{n}{e^2}\right)^{n}$, and conjectured that this bound is tight. We prove via a probabilistic construction that indeed $T(n) = \left((1+o(1))\frac{n}{e^2}\right)^{n}$. Until the present paper, no superexponential lower bound for $T(n)$ was known. We also give a simpler proof of the upper bound.

Consider the Birkhoff polytope of n by n doubly-stochastic matrices. As the Birkhoff-von Neumann theorem famously states, its vertex set coincides with the set of all n by n permutation matrices. Here we seek a higher-dimensional analog of this basic fact. Namely, consider the polytope which consists of all tristochastic arrays of order n. These are n by n by n arrays with nonnegative entries in which every line sums to 1. What can be said about its vertex set? It is well-known that an order-n Latin square may be viewed as a tristochastic array where every line contains n-1 zeros and a single 1 entry. Indeed, every Latin square of order n is a vertex, but as we show, such vertices constitute only a vanishingly small part of the total number of vertices. More concretely, we show that the number of vertices is at least (L_n)^{3/2-o(1)}, where L_n is the number of order-n Latin squares. We also briefly consider similar problems concerning the polytope of n by n by n arrays where the entries in every coordinate hyperplane sum to 1. Several open questions are presented as well.

Let STS(n) denote the number of Steiner triple systems on n vertices, and let F(n) denote the number of 1-factorizations of the complete graph on n vertices. We prove the following upper bound. STS(n) <= ((1 + o(1)) (n/e^2))^(n^2/6) F(n) <= ((1 + o(1)) (n/e^2))^(n^2/2) We conjecture that the bound is sharp. Our main tool is the entropy method.

What is the higher-dimensional analog of a permutation? If we think of a permutation as given by a permutation matrix, then the following definition suggests itself: A d-dimensional permutation of order n is an [n]^(d+1) array of zeros and ones in which every "line" contains a unique 1 entry. A line here is a set of entries of the form {(x_1,...,x_{i-1},y,x_{i+1},...,x_{d+1})}, for y between 1 and n, some index i between 1 and d+1 and some choice of x_j in [n] for all j except i. It is easy to observe that a one-dimensional permutation is simply a permutation matrix and that a two-dimensional permutation is synonymous with an order-n Latin square. We seek an estimate for the number of d-dimensional permutations. Our main result is the following upper bound on their number: ((1+o(1))(n/e^d))^(n^d). We tend to believe that this is actually the correct number, but the problem of proving the complementary lower bound remains open. Our main tool is an adaptation of Bregman's proof of the Minc conjecture on permanents. More concretely, our approach is very close in spirit to Radhakrishnan's proof of Bregman's theorem.

We present a randomized approximation algorithm for counting contingency tables, mxn non-negative integer matrices with given row sums R=(r_1, ..., r_m) and column sums C=(c_1, ..., c_n). We define smooth margins (R,C) in terms of the typical table and prove that for such margins the algorithm has quasi-polynomial N^{O(ln N)} complexity, where N=r_1+...+r_m=c_1+...+c_n. Various classes of margins are smooth, e.g., when m=O(n), n=O(m) and the ratios between the largest and the smallest row sums as well as between the largest and the smallest column sums are strictly smaller than the golden ratio (1+sqrt{5})/2 = 1.618. The algorithm builds on Monte Carlo integration and sampling algorithms for log-concave densities, the matrix scaling algorithm, the permanent approximation algorithm, and an integral representation for the number of contingency tables.