Latin Transversal Research Articles

A computational approach to analyze the Hadamard quasigroup product

<abstract><p>Based on the binary product described by any Latin square, the Hadamard quasigroup product is introduced in this paper as a natural generalization of the classical Hadamard product of matrices. The successive iteration of this new product is endowed with a cyclic behaviour that enables one to define a pair of new isomorphism invariants of Latin squares. Of particular interest is the set of Latin squares for which this iteration preserves the Latin square property, which requires the existence of successive localized Latin transversals within the Latin square under consideration. In order to enumerate and classify, up to isomorphism, these Latin squares, we propose a computational algebraic geometry approach based on the computation of reduced Gröbner bases. To illustrate this point, we obtain the classification of the sought Latin squares, for order up to six, by using the open computer algebra system for polynomial computations Singular.</p></abstract>

New bounds for the Moser‐Tardos distribution

The Lovász local lemma (LLL) is a probabilistic tool to generate combinatorial structures with good “local” properties. The “LLL‐distribution” further shows that these structures have good global properties in expectation. The seminal algorithm of Moser and Tardos turned the simplest, variable‐based form of the LLL into an efficient algorithm; this has since been extended to other probability spaces including random permutations. One can similarly define an “MT‐distribution” for these algorithms, that is, the distribution of the configuration they produce. We show new bounds on the MT‐distribution in the variable and permutation settings which are significantly stronger than those known to hold for the LLL‐distribution. As some example illustrations, we show a nearly tight bound on the minimum implicate size of a CNF Boolean formula, and we obtain improved bounds on weighted Latin transversals and partial Latin transversals.

Rainbow Hamilton cycles and lopsidependency

The Lovász Local Lemma is a powerful probabilistic tool used to prove the existence of combinatorial structures which avoid a set of constraints. A standard way to apply the local lemma is to prove that the set of constraints satisfy a lopsidependency condition and obtain a lopsidependency graph. For instance, Erdős and Spencer used this framework to posit the existence of Latin transversals in matrices provided no symbol appears too often in the matrix.The local lemma has been used in various ways to infer the existence of rainbow Hamilton cycles in complete graphs when each colour is used at most O(n) times. However, the existence of a lopsidependency graph for Hamilton cycles has neither been proved nor refuted. All previous approaches have had to prove a variant of the local lemma or reduce the problem of finding Hamilton cycles to finding another combinatorial structure, such as Latin transversals. In this paper, we revisit the question of whether or not Hamilton cycles have a lopsidependency graph and give a positive answer for this question. We also use the resampling oracle framework of Harvey and Vondrák to give a polynomial time algorithm for finding rainbow Hamilton cycles in complete graphs.

Algorithmic and Enumerative Aspects of the Moser-Tardos Distribution

Moser and Tardos have developed a powerful algorithmic approach (henceforth MT) to the Lovász Local Lemma (LLL); the basic operation done in MT and its variants is a search for “bad” events in a current configuration. In the initial stage of MT, the variables are set independently. We examine the distributions on these variables that arise during intermediate stages of MT. We show that these configurations have a more or less “random” form, building further on the MT-distribution concept of Haeupler et al. in understanding the (intermediate and) output distribution of MT. This has a variety of algorithmic applications; the most important is that bad events can be found relatively quickly, improving on MT across the complexity spectrum. It makes some polynomial-time algorithms sublinear (e.g., for Latin transversals, which are of basic combinatorial interest), gives lower-degree polynomial runtimes in some settings, transforms certain superpolynomial-time algorithms into polynomial-time algorithms, and leads to Las Vegas algorithms for some coloring problems for which only Monte Carlo algorithms were known. We show that, in certain conditions when the LLL condition is violated, a variant of the MT algorithm can still produce a distribution that avoids most of the bad events. We show in some cases that this MT variant can run faster than the original MT algorithm itself and develop the first-known criterion for the case of the asymmetric LLL. This can be used to find partial Latin transversals—improving on earlier bounds of Stein (1975)—among other applications. We furthermore give applications in enumeration, showing that most applications (for which we aim for all or most of the bad events to be avoided) have large solution sets. We do this by showing that the MT distribution has large Rényi entropy.

Rainbow Perfect Matchings in Complete Bipartite Graphs: Existence and Counting

A perfect matchingMin an edge-coloured complete bipartite graphKn,nis rainbow if no pair of edges inMhave the same colour. We obtain asymptotic enumeration results for the number of rainbow perfect matchings in terms of the maximum number of occurrences of each colour. We also consider two natural models of random edge-colourings ofKn,nand show that if the number of colours is at leastn, then there is with high probability a rainbow perfect matching. This in particular shows that almost every square matrix of ordernin which every entry appearsntimes has a Latin transversal.

Rainbow Matchings: existence and counting

Abstract A perfect matching M in an edge–colored complete bipartite graph K n , n is rainbow if no pair of edges in M have the same color. We obtain asymptotic enumeration results for the number of rainbow matchings in terms of the maximum number of occurrences of a color. We also consider two natural models of random edge–colored K n , n and show that, if the number of colors is at least n, then there is whp a random matching. This in particular shows that almost every matrix in which every entry appears n times has a Latin transversal.

An Improvement of the Lovász Local Lemma via Cluster Expansion

An old result by Shearer relates the Lovász local lemma with the independent set polynomial on graphs, and consequently, as observed by Scott and Sokal, with the partition function of the hard-core lattice gas on graphs. We use this connection and a recent result on the analyticity of the logarithm of the partition function of the abstract polymer gas to get an improved version of the Lovász local lemma. As an application we obtain tighter bounds on conditions for the existence of Latin transversal matrices.

An additive theorem and restricted sumsets

Let $G$ be any additive abelian group with cyclic torsion subgroup, and let $A$, $B$ and $C$ be finite subsets of $G$ with cardinality $n>0$. We show that there is a numbering $\{a_i\}_{i=1}^n$ of the elements of $A$, a numbering $\{b_i\}_{i=1}^n$ of the elements of $B$ and a numbering $\{c_i\}_{i=1}^n$ of the elements of $C$, such that all the sums $a_i+b_i+c_i (1\ls i\ls n)$ are (pairwise) distinct. Consequently, each subcube of the Latin cube formed by the Cayley addition table of $\Z/N\Z$ contains a Latin transversal. This additive theorem is an essential result which can be further extended via restricted sumsets in a field.

Additive Latin transversals and group rings

LetA={a 1, …,a k} and {b 1, …,b k} be two subsets of an abelian groupG, k≤|G|. Snevily conjectured that, when |G| is odd, there is a numbering of the elements ofB such thata i+b i,1≤i≤k are pairwise distinct. By using a polynomial method, Alon affirmed this conjecture for |G| prime, even whenA is a sequence ofk<|G| elements. With a new application of the polynomial method, Dasgupta, Karolyi, Serra and Szegedy extended Alon’s result to the groupsZ p r andZ p rin the casek<p and verified Snevily’s conjecture for every cyclic group. In this paper, by employing group rings as a tool, we prove that Alon’s result is true for any finite abelianp-group withk<√2p, and verify Snevily’s conjecture for every abelian group of odd order in the casek<√p, wherep is the smallest prime divisor of |G|.

Covering with latin transversals

Given an n × n matrix A = [ a ij ], a transversal of A is a set of elements, one from each row and one from each column. A transversal is a latin transversal if no two elements are the same. Erdös and Spencer showed that there always exists a latin transversal in any n × n matrix in which no element appears more than s times, for s⩽ ( n — 1)/16. Here we show that, in fact, the elements of the matrix can be partitioned into n disjoint latin transversals, provided n is a power of 2 and no element appears more than εn times for some fixed ε>0. The assumption that n is a power of 2 can be weakened, but at the moment we are unable to prove the theorem for all values of n.

Lopsided Lovász Local Lemma and Latin transversals

A new version of the Lovász Local lemma is used to prove the existence of Latin transversals in matrices where no symbol appears too often.

Has Every Latin Square of Order n a Partial Latin Transversal of Size n - 1?

(1988). Has Every Latin Square of Order n a Partial Latin Transversal of Size n - 1? The American Mathematical Monthly: Vol. 95, No. 5, pp. 428-430.

Has Every Latin Square of Order n a Partial Latin Transversal of Size n - 1?

Has Every Latin Square of Order <i>n</i> a Partial Latin Transversal of Size <i>n</i> - 1?