Quasi‐carousel tournaments

A tournament is called locally transitive if the outneighborhood and the inneighborhood of every vertex are transitive. Equivalently, a tournament is locally transitive if it avoids the tournaments W₄ and L₄, which are the only tournaments up to isomorphism on four vertices containing a unique 3‐cycle. On the other hand, a sequence of tournaments  with  is called almost balanced if all but  vertices of  have outdegree . In the same spirit of quasi‐random properties, we present several characterizations of tournament sequences that are both almost balanced and asymptotically locally transitive in the sense that the density of W₄ and L₄ in  goes to zero as n goes to infinity.     

Ramsey Results for Cycle Spectra

Let denote the set of lengths of cycles of a graph G of order n and let denote the complement of G. We show that if , then contains all odd ? with and all even ? with , where and denote the maximum odd and the maximum even integer in , respectively. From this we deduce that the set contains at least integers, which is sharp.     

An Asymptotic Version of the Multigraph 1‐Factorization Conjecture

We give a self‐contained proof that for all positive integers r and all , there is an integer such that for all any regular multigraph of order 2n with multiplicity at most r and degree at least is 1‐factorizable. This generalizes results of Perkovi? and Reed (Discrete Math 165/166 (1997), 567–578) and Plantholt and Tipnis (J London Math Soc 44 (1991), 393–400).     

Two Questions of Erdős on Hypergraphs above the Turán Threshold

For ordinary graphs it is known that any graph G with more edges than the Turán number of must contain several copies of , and a copy of , the complete graph on vertices with one missing edge. Erd?s asked if the same result is true for , the complete 3‐uniform hypergraph on s vertices. In this note, we show that for small values of n, the number of vertices in G, the answer is negative for . For the second property, that of containing a , we show that for the answer is negative for all large n as well, by proving that the Turán density of is greater than that of .     

Edge‐transitive homogeneous factorizations of complete uniform hypergraphs

For a finite set V and a positive integer k with , letting be the set of all k‐subsets of V, the pair is called the complete k‐hypergraph on V, while each k‐subset of V is called an edge. A factorization of the complete k‐hypergraph of index , simply a ‐factorization of order n, is a partition of the edges into s disjoint subsets such that each k‐hypergraph , called a factor, is a spanning subhypergraph of . Such a factorization is homogeneous if there exist two transitive subgroups G and M of the symmetric group of degree n such that G induces a transitive action on the set and M lies in the kernel of this action. In this article, we give a classification of homogeneous factorizations of that admit a group acting transitively on the edges of . It is shown that, for and , there exists an edge‐transitive homogeneous ‐factorization of order n if and only if is one of (32, 3, 5), (32, 3, 31), (33, 4, 5), , and , where and q is a prime power with .     

Non‐Critical Vertices and Long Circuits in Strong Tournaments of Order n and Diameter d

Let T be a strong tournament of order with diameter . A vertex w in T is non‐critical if the subtournament is also strong. In the opposite case, it is a critical vertex. In the present article, we show that T has at most critical vertices. This fact and Moon's vertex‐pancyclic theorem imply that for , it contains at least circuits of length . We describe the class of all strong tournaments of order with diameter for which this lower bound is achieved and show that for , the minimum number of circuits of length in a tournament of this class is equal to . In turn, the minimum among all strong tournaments of order with diameter 2 grows exponentially with respect to n for any given . Finally, for , we select a strong tournament of order n with diameter d and conjecture that for any strong tournament T of order n whose diameter does not exceed d, the number of circuits of length ? in T is not less than that in for each possible ?. Moreover, if these two numbers are equal to each other for some given , where , then T is isomorphic to either or its converse . For , this conjecture was proved by Las Vergnas. In the present article, we confirm it for the case . In an Appendix, some problems concerning non‐critical vertices are considered for generalizations of tournaments.     

Packing Triangles in Regular Tournaments

We prove that a regular tournament with n vertices has more than pairwise arc‐disjoint directed triangles. On the other hand, we construct regular tournaments with a feedback arc set of size less than , so these tournaments do not have pairwise arc‐disjoint triangles.     

The decycling number and maximum genus of cubic graphs

Let and denote the minimum size of a decycling set and maximum genus of a graph G, respectively. For a connected cubic graph G of order n, it is shown that . Applying the formula, we obtain some new results on the decycling number and maximum genus of cubic graphs. Furthermore, it is shown that the number of vertices of a decycling set S in a k‐regular graph G is , where c and are the number of components of and the number of edges in , respectively. Therefore, S is minimum if and only if is minimum. As an application, this leads to a lower bound for of a k‐regular graph G. In many cases this bound may be sharp.     

Packing odd T‐joins with at most two terminals

Take a graph G, an edge subset , and a set of terminals where is even. The triple is called a signed graft. A T‐join is odd if it contains an odd number of edges from Σ. Let ν be the maximum number of edge‐disjoint odd T‐joins. A signature is a set of the form where and is even. Let τ be the minimum cardinality a T‐cut or a signature can achieve. Then and we say that packs if equality holds here. We prove that packs if the signed graft is Eulerian and it excludes two special nonpacking minors. Our result confirms the Cycling Conjecture for the class of clutters of odd T‐joins with at most two terminals. Corollaries of this result include, the characterizations of weakly and evenly bipartite graphs, packing two‐commodity paths, packing T‐joins with at most four terminals, and a new result on covering edges with cuts.     

Excluding pairs of tournaments

The Erdős–Hajnal conjecture states that for every given undirected graph H there exists a constant such that every graph G that does not contain H as an induced subgraph contains a clique or a stable set of size at least . The conjecture is still open. Its equivalent directed version states that for every given tournament H there exists a constant such that every H‐free tournament T contains a transitive subtournament of order at least . In this article, we prove that for several pairs of tournaments, H₁ and H₂, there exists a constant such that every ‐free tournament T contains a transitive subtournament of size at least . In particular, we prove that for several tournaments H, there exists a constant such that every ‐free tournament T, where stands for the complement of H, has a transitive subtournament of size at least . To the best of our knowledge these are first nontrivial results of this type.     

Full subgraphs

Let be a graph of density p on n vertices. Following Erdős, Łuczak, and Spencer, an m‐vertex subgraph H of G is called full if H has minimum degree at least . Let denote the order of a largest full subgraph of G. If is a nonnegative integer, define Erdős, Łuczak, and Spencer proved that for , In this article, we prove the following lower bound: for , Furthermore, we show that this is tight up to a multiplicative constant factor for infinitely many p near the elements of . In contrast, we show that for any n‐vertex graph G, either G or contains a full subgraph on vertices. Finally, we discuss full subgraphs of random and pseudo‐random graphs, and several open problems.     

On the Structure of Graphs with Given Odd Girth and Large Minimum Degree

We study minimum degree conditions for which a graph with given odd girth has a simple structure. For example, the classical work of Andrásfai, Erd?s, and Sós implies that every n‐vertex graph with odd girth and minimum degree bigger than must be bipartite. We consider graphs with a weaker condition on the minimum degree. Generalizing results of Häggkvist and of Häggkvist and Jin for the cases and 3, we show that every n‐vertex graph with odd girth and minimum degree bigger than is homomorphic to the cycle of length . This is best possible in the sense that there are graphs with minimum degree and odd girth that are not homomorphic to the cycle of length . Similar results were obtained by Brandt and Ribe‐Baumann.     

More on the Bipartite Decomposition of Random Graphs

For a graph , let denote the minimum number of pairwise edge disjoint complete bipartite subgraphs of G so that each edge of G belongs to exactly one of them. It is easy to see that for every graph G , , where is the maximum size of an independent set of G . Erd?s conjectured in the 80s that for almost every graph G equality holds, that is that for the random graph , with high probability, that is with probability that tends to 1 as n tends to infinity. The first author showed that this is slightly false, proving that for most values of n tending to infinity and for , with high probability. We prove a stronger bound: there exists an absolute constant so that with high probability.     

On an anti‐Ramsey threshold for sparse graphs with one triangle

For graphs G and H, let  denote the property that for every proper edge‐coloring of G (with an arbitrary number of colors) there is a rainbow copy of H in G, that is, a copy of H with no two edges of the same color. The authors (2014) proved that, for every graph H, the threshold function  of this property for the binomial random graph  is asymptotically at most , where denotes the so‐called maximum 2‐density of H. Nenadov et al. (2014) proved that if H is a cycle with at least  seven vertices or a complete graph with at least 19 vertices, then . We show that there exists a fairly rich, infinite family of graphs F containing a triangle such that if for suitable constants and , where , then almost surely. In particular, for any such graph F.     

On a Conjecture of Erdős,Gallai, and Tuza

Erd?s, Gallai, and Tuza posed the following problem: given an n‐vertex graph G, let denote the smallest size of a set of edges whose deletion makes G triangle‐free, and let denote the largest size of a set of edges containing at most one edge from each triangle of G. Is it always the case that ? We have two main results. We first obtain the upper bound , as a partial result toward the Erd?s–Gallai–Tuza conjecture. We also show that always , where m is the number of edges in G; this bound is sharp in several notable cases.     

Low Polynomial Exclusion of Planar Graph Patterns

The celebrated grid exclusion theorem states that for every h‐vertex planar graph H , there is a constant such that if a graph G does not contain H as a minor then G has treewidth at most . We are looking for patterns of H where this bound can become a low degree polynomial. We provide such bounds for the following parameterized graphs: the wheel , the double wheel , any graph of pathwidth at most 2 , and the yurt graph .     

Large Induced Forests in Graphs

In this article, we prove three theorems. The first is that every connected graph of order n and size m has an induced forest of order at least with equality if and only if such a graph is obtained from a tree by expanding every vertex to a clique of order either 4 or 5. This improves the previous lower bound of Alon–Kahn–Seymour for , and implies that such a graph has an induced forest of order at least for . This latter result relates to the conjecture of Albertson and Berman that every planar graph of order n has an induced forest of order at least . The second is that every connected triangle‐free graph of order n and size m has an induced forest of order at least . This bound is sharp by the cube and the Wagner graph. It also improves the previous lower bound of Alon–Mubayi–Thomas for , and implies that such a graph has an induced forest of order at least for . This latter result relates to the conjecture of Akiyama and Watanabe that every bipartite planar graph of order n has an induced forest of order at least . The third is that every connected planar graph of order n and size m with girth at least 5 has an induced forest of order at least with equality if and only if such a graph is obtained from a tree by expanding every vertex to one of five specific graphs. This implies that such a graph has an induced forest of order at least , where was conjectured to be the best lower bound by Kowalik, Lu?ar, and ?krekovski.     

Nonseparating Cycles Avoiding Specific Vertices

Thomassen proved that every ‐connected graph G contains an induced cycle C such that is k‐connected, establishing a conjecture of Lovász. In general, one could ask the following question: For any positive integers , does there exist a smallest positive integer such that for any ‐connected graph G, any with , and any , there is an induced cycle C in such that and is l‐connected? The case when is a well‐known conjecture of Lovász that is still open for . In this article, we prove and . We also consider a weaker version: For any positive integers , is there a smallest positive integer such that for every ‐connected graph G and any with , there is an induced cycle C in such that is l‐connected? The case when was studied by Thomassen. We prove and .     

Sufficient conditions for the existence of a path‐factor which are related to odd components

In this article, we are concerned with sufficient conditions for the existence of a ‐factor. We prove that for , there exists such that if a graph G satisfies for all , then G has a ‐factor, where is the number of components C of with . On the other hand, we construct infinitely many graphs G having no ‐factor such that for all .     

List Coloring with a Bounded Palette

Král' and Sgall (J Graph Theory 49(3) (2005), 177–186) introduced a refinement of list coloring where every color list must be subset to one predetermined palette of colors. We call this ‐choosability when the palette is of size at most ? and the lists must be of size at least k . They showed that, for any integer , there is an integer , satisfying as , such that, if a graph is ‐choosable, then it is C‐choosable, and asked if C is required to be exponential in k . We demonstrate it must satisfy . For an integer , if is the least integer such that a graph is ‐choosable if it is ‐choosable, then we more generally supply a lower bound on , one that is super‐polynomial in k if , by relation to an extremal set theoretic property. By the use of containers, we also give upper bounds on that improve on earlier bounds if .