The Erdős–Simonovits stability theorem states that for all ε > 0 there exists α > 0 such that if G is a Kr+1-free graph on n vertices with e(G) > ex(n, Kr+1)– α n2, then one can remove εn2 edges from G to obtain an r-partite graph. Füredi gave a short proof that one can choose α = ε. We give a bound for the relationship of α and ε which is asymptotically sharp as ε → 0.

Given a fixed graph H that embeds in a surface $\Sigma$ , what is the maximum number of copies of H in an n-vertex graph G that embeds in $\Sigma$ ? We show that the answer is $\Theta(n^{f(H)})$ , where f(H) is a graph invariant called the ‘flap-number’ of H, which is independent of $\Sigma$ . This simultaneously answers two open problems posed by Eppstein ((1993) J. Graph Theory17(3) 409–416.). The same proof also answers the question for minor-closed classes. That is, if H is a $K_{3,t}$ minor-free graph, then the maximum number of copies of H in an n-vertex $K_{3,t}$ minor-free graph G is $\Theta(n^{f'(H)})$ , where f′(H) is a graph invariant closely related to the flap-number of H. Finally, when H is a complete graph we give more precise answers.

For fixed integers p and q, let f(n,p,q) denote the minimum number of colours needed to colour all of the edges of the complete graph Kn such that no clique of p vertices spans fewer than q distinct colours. Any edge-colouring with this property is known as a (p,q)-colouring. We construct an explicit (5,5)-colouring that shows that f(n,5,5) ≤ n1/3 + o(1) as n → ∞. This improves upon the best known probabilistic upper bound of O(n1/2) given by Erdős and Gyárfás, and comes close to matching the best known lower bound Ω(n1/3).

Polyhedral embeddings of cubic graphs by means of certain operations are studied. It is proved that some known families of snarks have no (orientable) polyhedral embeddings. This result supports a conjecture of Grünbaum that no snark admits an orientable polyhedral embedding. This conjecture is verified by computer for all snarks having fewer than 30 vertices. On the other hand, for every non-orientable surface $S$, there exists a non-3-edge-colourable graph which polyhedrally embeds in $S$.

In 1978, Dhar suggested a model of a lattice gas whose states are partial orders. In this context he raised the question of determining the number of partial orders with a fixed number of comparable pairs. Dhar conjectured that in order to find a good approximation to this number, it should suffice to enumerate families of layer posets. In this paper we prove this conjecture and thereby prepare the ground for a complete answer to the question.

We prove the following 30 year-old conjecture of Győri and Tuza: the edges of every n-vertex graph G can be decomposed into complete graphs C1,. . .,Cℓ of orders two and three such that |C1|+···+|Cℓ| ≤ (1/2+o(1))n2. This result implies the asymptotic version of the old result of Erdős, Goodman and Pósa that asserts the existence of such a decomposition with ℓ ≤ n2/4.

Bollobás and Riordan introduce a Tutte polynomial for coloured graphs and matroids in [3]. We observe that this polynomial has an expansion as a sum indexed by the subsets of the ground-set of a coloured matroid, generalizing the subset expansion of the Tutte polynomial. We also discuss similar expansions of other contraction–deletion invariants of graphs and matroids.

Assemblies are decomposable combinatorial objects characterized by a sequence mi that counts the number of possible components of size i. Permutations on n elements, mappings from a set containing n elements into itself, 2-regular graphs on n vertices, and set partitions on a set of size n are all assemblies with natural decompositions. Logarithmic assemblies are characterized by constants θ > 0 and κ0 > 0 such that miκi0/ (i−1)! → θ. Random mappings, permutations and 2-regular graphs are all logarithmic assemblies, but set partitions are not.

Given a logarithmic assembly, all representatives having total size n are chosen uniformly and a component counting process C(n) = (C1(n), C2(n), …, Cn(n)) is defined, where Ci(n) is the number of components of size i. Our results also apply to C(n) distributed as the Ewens sampling formula with parameter θ. Denote the component counting process up to size at most b by Cb(n) = (C1(n), C2(n), …, Cb(n)). It is natural to approximate Cb by Zb = (Z1, Z2, …, Zb), the b-dimensional process of independent Poisson variables Zi for which the ith variable has expectation [ ]Zi = miκi0 exp((1−θ)i/n)/i!. We find asymptotics for the total variation distance between Cb(n) and Zb.

A partition of a positive integer n is a finite sequence of positive integers a1, a2, . . ., ak such that a1+a2+ċ ċ ċ+ak=n and ai+1 ≥ ai for all i. Let d be a fixed positive integer. We say that we have an ascent of size d or more if ai+1 ≥ ai+d.

We determine the mean, the variance and the limiting distribution of the number of ascents of size d or more (equivalently, the number of distinct part sizes of multiplicity d or more) in the partitions of n.

The hard-core model has attracted much attention across several disciplines, representing lattice gases in statistical physics and independent sets in discrete mathematics and computer science. On finite graphs, we are given a parameter λ, and an independent set I arises with probability proportional to λ|I|. On infinite graphs a Gibbs measure is defined as a suitable limit with the correct conditional probabilities, and we are interested in determining when this limit is unique and when there is phase coexistence, i.e., existence of multiple Gibbs measures.

It has long been conjectured that on ℤ2 this model has a critical value λc ≈ 3.796 with the property that if λ < λc then it exhibits uniqueness of phase, while if λ > λc then there is phase coexistence. Much of the work to date on this problem has focused on the regime of uniqueness, with the state of the art being recent work of Sinclair, Srivastava, Štefankovič and Yin showing that there is a unique Gibbs measure for all λ < 2.538. Here we explore the other direction and prove that there are multiple Gibbs measures for all λ > 5.3506. We also show that with the methods we are using we cannot hope to replace 5.3506 with anything below 4.8771.

Our proof begins along the lines of the standard Peierls argument, but we add two innovations. First, following ideas of Kotecký and Randall, we construct an event that distinguishes two boundary conditions and always has long contours associated with it, obviating the need to accurately enumerate short contours. Second, we obtain improved bounds on the number of contours by relating them to a new class of self-avoiding walks on an oriented version of ℤ2.

Let a, b and n be integers with 2[les ]a[les ]b and n[ges ]a+b. Suppose that [Ascr ]⊂([n]a) and [Bscr ]⊂([n]b) are nontrivial cross-intersecting families. Then [mid ][Ascr ][mid ]+[mid ][Bscr ][mid ][les ]2+(nb) −2(n−ab)+ (n−2ab). This result is best possible.

A sharper form of the Szarek–Talagrand ‘isomorphic’ version of the Sauer–Shelah lemma is proved. Also we prove an analogous ‘isomorphic’ version of the Karpovsky–Milman lemma, which is a generalization of that due to Sauer and Shelah.

Let $G$ be a finite group of order $n$ and let $k$ be a natural number. Let $\{x_i : i\in I\}$ be a family of elements of $G$ such that $|I|= n+k-1$. Let $v$ be the most repeated value of the family. Let $ \{ \sigma_i : 1\leq i \leq k \} $ be a family of permutations of $G$ such that $\sigma_i(1)=1$ for all $i$. We obtain the following result.

There are pairwise distinct elements $i_1, i_2, \dots ,i_k\in I$ such that \[ \prod_{1\leq j\leq k } \sigma_j \big(v^{-1}x_ {i_j }\big) =1.\]

We give an efficient algorithm that, given a graph G and a partition V1,…,Vm of its vertex set, finds either an independent transversal (an independent set {v1,…,vm} in G such that ${v_i} \in {V_i}$ for each i), or a subset ${\cal B}$ of vertex classes such that the subgraph of G induced by $\bigcup\nolimits_{\cal B}$ has a small dominating set. A non-algorithmic proof of this result has been known for a number of years and has been used to solve many other problems. Thus we are able to give algorithmic versions of many of these applications, a few of which we describe explicitly here.

We explore the tree limits recently defined by Elek and Tardos. In particular, we find tree limits for many classes of random trees. We give general theorems for three classes of conditional Galton–Watson trees and simply generated trees, for split trees and generalized split trees (as defined here), and for trees defined by a continuous-time branching process. These general results include, for example, random labelled trees, ordered trees, random recursive trees, preferential attachment trees, and binary search trees.

Let $P(n)$ and $C(n)$ denote, respectively, the maximum possible numbers of Hamiltonian paths and Hamiltonian cycles in a tournament on n vertices. The study of $P(n)$ was suggested by Szele [14], who showed in an early application of the probabilistic method that $P(n) \geq n!2^{-n+1}$, and conjectured that $\lim ( {P(n)}/ {n!} )^{1/n}= 1/2.$ This was proved by Alon [2], who observed that the conjecture follows from a suitable bound on $C(n)$, and showed $C(n) <O(n^{3/2}(n-1)!2^{-n}).$ Here we improve this to $C(n)<O\big(n^{3/2-\xi}(n-1)!2^{-n}\big),$ with $\xi = 0.2507$… Our approach is mainly based on entropy considerations.

Random constraint satisfaction problems play an important role in computer science and combinatorics. For example, they provide challenging benchmark examples for algorithms, and they have been harnessed in probabilistic constructions of combinatorial structures with peculiar features. In an important contribution (Krzakala et al. 2007, Proc. Nat. Acad. Sci.), physicists made several predictions on the precise location and nature of phase transitions in random constraint satisfaction problems. Specifically, they predicted that their satisfiability thresholds are quite generally preceded by several other thresholds that have a substantial impact both combinatorially and computationally. These include the condensation phase transition, where long-range correlations between variables emerge, and the reconstruction threshold. In this paper we prove these physics predictions for a broad class of random constraint satisfaction problems. Additionally, we obtain contiguity results that have implications for Bayesian inference tasks, a subject that has received a great deal of interest recently (e.g. Banks et al. 2016, Proc. 29th COLT).

The present paper deals with two graph parameters related to cover graphs and acyclic orientations of graphs.

The parameter $c(G)$ of a graph $G$, introduced by B. Bollobás, G. Brightwell and J. Nešetřil [Order3 245–255], is defined as the minimum number of edges one needs to delete from $G$ in order to obtain a cover graph. Extending their results, we prove that, for $\delta >0$, $(1-\delta) \frac{1}{l} \frac{n^2p}{2} \leq c({\mathcal G}_{n,p}) \leq (1+\delta) \frac{1}{l} \frac{n^2p}{2}$ asymptotically almost surely as long as $C n^{-1 + \frac{1}{l}} \leq p(n) \leq c n^{-1 + \frac{1}{ l-1} }$ for some positive constants $c$ and $C$. Here, as usual, ${\mathcal G}_{n,p}$ is the random graph.

Given an acyclic orientation of a graph $G$, an arc is called dependent if its reversal creates an oriented cycle. Let $d_{\min}(G)$ be the minimum number of dependent arcs in any acyclic orientation of $G$. We determine the supremum, denoted by $r_{\chi,g}$, of $d_{\min}(G)/e(G)$ in the class of graphs $G$ with chromatic number $\chi$ and girth $g$. Namely, we show that $r_{\chi,g} = {(\scriptsize\begin{array}{@{}c@{}}{\chi}-g+2\\ 2\end{array})} / {(\scriptsize\begin{array}{@{}c@{}}{\chi}\\ 2\end{array})}$. This extends results of D. C. Fisher, K. Fraughnaugh, L. Langley and D. B. West [J. Combin. Theory Ser. B71 73–78].

Let $T$ denote a real function defined on random subsets of a given family of finite sets. The random variable $T$ is decomposed into the sum of the linear, the quadratic, the cubic etc. parts which are mutually uncorrelated. Applications of this decomposition to the asymptotics of the probability distribution of $T$ (as the sizes of random subsets and of finite sets increase) are discussed.

Let G be an additive abelian group, let n ⩾ 1 be an integer, let S be a sequence over G of length |S| ⩾ n + 1, and let ${\mathsf h}$(S) denote the maximum multiplicity of a term in S. Let Σn(S) denote the set consisting of all elements in G which can be expressed as the sum of terms from a subsequence of S having length n. In this paper, we prove that either ng ∈ Σn(S) for every term g in S whose multiplicity is at least ${\mathsf h}$(S) − 1 or |Σn(S)| ⩾ min{n + 1, |S| − n + | supp (S)| − 1}, where |supp(S)| denotes the number of distinct terms that occur in S. When G is finite cyclic and n = |G|, this confirms a conjecture of Y. O. Hamidoune from 2003.