Let n ≥ 1 be an integer. Given a vector a=(a1,. . ,an)∈, write(the ‘projection of a onto the positive orthant’). For a set A⊆ put A+:={a+: a ∈ A} and A−A:={a−b: a, b ∈ A}. Improving previously known bounds, we show that |(A−A)+| ≥ |A|3/5/6 for any finite set A⊆, and that |(A−A)+| ≥ c|A|6/11/(log |A|)2/11 with an absolute constant c>0 for any finite set A⊆ such that |A| ≥ 2.

The λ-dilate of a set A is λċA={λa : a∈A}. We give an asymptotically sharp lower bound on the size of sumsets of the form λ1ċA+ċċċ+λkċA for arbitrary integers λ1,. . .,λk and integer sets A. We also establish an upper bound for such sums, which is similar to, but often stronger than Plünnecke's inequality.

The number of spanning trees in the giant component of the random graph (n, c/n) (c > 1) grows like exp{m(f(c)+o(1))} as n → ∞, where m is the number of vertices in the giant component. The function f is not known explicitly, but we show that it is strictly increasing and infinitely differentiable. Moreover, we give an explicit lower bound on f′(c). A key lemma is the following. Let PGW(λ) denote a Galton–Watson tree having Poisson offspring distribution with parameter λ. Suppose that λ*>λ>1. We show that PGW(λ*) conditioned to survive forever stochastically dominates PGW(λ) conditioned to survive forever.

We consider the problem of minimizing the size of a family of sets such that every subset of {1,. . ., n} can be written as a disjoint union of at most k members of , where k and n are given numbers. This problem originates in a real-world application aiming at the diversity of industrial production. At the same time, the question of finding the minimum of || so that every subset of {1,. . ., n} is the union of two sets in was asked by Erdős and studied recently by Füredi and Katona without requiring the disjointness of the sets. A simple construction providing a feasible solution is conjectured to be optimal for this problem for all values of n and k and regardless of the disjointness requirement; we prove this conjecture in special cases including all (n, k) for which n≤3k holds, and some individual values of n and k.

In 1998 Łuczak Rödl and Szemerédi [7] proved, by means of the Regularity Lemma, that there exists n0 such that, for any n ≥ n0 and two-edge-colouring of Kn, there exists a pair of vertex-disjoint monochromatic cycles of opposite colours covering the vertices of Kn. In this paper we make use of an alternative method of finding useful structure in a graph, leading to a proof of the same result with a much smaller value of n0. The proof gives a polynomial-time algorithm for finding the two cycles.

We use a greedy probabilistic method to prove that, for every ε > 0, every m × n Latin rectangle on n symbols has an orthogonal mate, where m = (1 − ε)n. That is, we show the existence of a second Latin rectangle such that no pair of the mn cells receives the same pair of symbols in the two rectangles.

Both the hopcount HN (the number of links) and the weight WN (the sum of the weights on links) of the shortest path between two arbitrary nodes in the complete graph KN with i.i.d. exponential link weights is computed. We consider the joint distribution of the pair (HN, WN) and derive, after proper scaling, the joint limiting distribution. One of the results is that HN and WN, properly scaled, are asymptotically independent.

Semple and Welsh [5] introduced the concept of correlated matroids, which relate to conjectures by Grimmett and Winkler [2], and Pemantle [4], respectively, that the uniformly random forest and the uniformly random connected subgraph of a finite graph have the edge-negative-association property. In this paper, we extend results of Semple and Welsh, and show that the Grimmett and Winkler, and Pemantle conjectures are equivalent to statements about correlated graphic matroids. We also answer some open questions raised in [5] regarding correlated matroids, and in particular show that the 2-sum of correlated matroids is correlated.

For a graph G and an integer t we let mcct(G) be the smallest m such that there exists a colouring of the vertices of G by t colours with no monochromatic connected subgraph having more than m vertices. Let be any non-trivial minor-closed family of graphs. We show that mcc2(G) = O(n2/3) for any n-vertex graph G ∈ . This bound is asymptotically optimal and it is attained for planar graphs. More generally, for every such , and every fixed t we show that mcct(G)=O(n2/(t+1)). On the other hand, we have examples of graphs G with no Kt+3 minor and with mcct(G)=Ω(n2/(2t−1)).

It is also interesting to consider graphs of bounded degrees. Haxell, Szabó and Tardos proved mcc2(G) ≤ 20000 for every graph G of maximum degree 5. We show that there are n-vertex 7-regular graphs G with mcc2(G)=Ω(n), and more sharply, for every ϵ > 0 there exists cϵ > 0 and n-vertex graphs of maximum degree 7, average degree at most 6 + ϵ for all subgraphs, and with mcc2(G) ≥ cϵn. For 6-regular graphs it is known only that the maximum order of magnitude of mcc2 is between and n.

We also offer a Ramsey-theoretic perspective of the quantity mcct(G).

Given a set of s points and a set of n2 lines in three-dimensional Euclidean space such that each line is incident to n points but no n lines are coplanar, we show that s = Ω(n11/4). This is the first non-trivial answer to a question recently posed by Jean Bourgain.

Let the random graph Rn be drawn uniformly at random from the set of all simple planar graphs on n labelled vertices. We see that with high probability the maximum degree of Rn is Θ(ln n). We consider also the maximum size of a face and the maximum increase in the number of components on deleting a vertex. These results extend to graphs embeddable on any fixed surface.

For a fixed ρ ∈ [0, 1], what is (asymptotically) the minimal possible density g3(ρ) of triangles in a graph with edge density ρ? We completely solve this problem by proving thatwhere is the integer such that .

If G is a graph with vertex set [n] then is G-intersecting if, for all , either A ∩ B ≠ ∅ or there exist a ∈ A and b ∈ B such that a ~Gb.

The question of how large a k-uniform G-intersecting family can be was first considered by Bohman, Frieze, Ruszinkó and Thoma [2], who identified two natural candidates for the extrema depending on the relative sizes of k and n and asked whether there is a sharp phase transition between the two. Our first result shows that there is a sharp transition and characterizes the extremal families when G is a matching. We also give an example demonstrating that other extremal families can occur.

Our second result gives a sufficient condition for the largest G-intersecting family to contain almost all k-sets. In particular we show that if Cn is the n-cycle and k > αn + o(n), where α = 0.266. . . is the smallest positive root of (1 − x)3(1 + x) = 1/2, then the largest Cn-intersecting family has size .

Finally we consider the non-uniform problem, and show that in this case the size of the largest G-intersecting family depends on the matching number of G.

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.

We provide an estimate, sharp up to poly-logarithmic factors, of the asymptotic almost sure mixing time of the graph created by long-range percolation on the cycle of length N (). While it is known that the asymptotic almost sure diameter drops from linear to poly-logarithmic as the exponent s decreases below 2 [4, 9], the asymptotic almost sure mixing time drops from N2 only to Ns-1 (up to poly-logarithmic factors).

In a random passive intersection graph model the edges of the graph are decided by taking the union of a fixed number of cliques of random size. We give conditions for a random passive intersection graph model to have a limiting vertex degree distribution, in particular to have a Poisson limiting vertex degree distribution. We give related conditions which, in addition to implying a limiting vertex degree distribution, imply convergence of expectation.

In the job shop scheduling problem k-units-J m , there are m machines and each machine has an integer processing time of at most k time units. Each job consists of a permutation of m tasks corresponding to all machines and thus all jobs have an identicaldilation D. The contribution of this paper are the following results;(i) for $d=o(\sqrt{D})$ jobs and every fixed k, the makespan of an optimal schedule is at most D+ o(D), which extends the result of [3]for k=1; (ii) a randomized on-line approximation algorithm for k-units-J m  ispresented. This is the on-line algorithm with the best known competitiveratio against an oblivious adversary for $d = o(\sqrt{D})$ and k > 1; (iii) different processing times yield harder instances than identicalprocessing times. There is no 5/3 competitive deterministic on-linealgorithm for k-units-J m , whereas the competitive ratio of the randomizedon-line algorithm of (ii) still tends to 1 for $d = o(\sqrt{D})$ .

We consider relational periods, where the relation is acompatibility relation on words induced by a relation on letters. Weprove a variant of the theorem of Fine and Wilf for a (pure) periodand a relational period.

Codd defined the relational algebra [E.F. Codd, Communications of the ACM13 (1970) 377–387;E.F. Codd, Relational completeness of data base sublanguages, in Data Base Systems, R. Rustin, Ed., Prentice-Hall (1972) 65–98] as the algebra with operations projection, join, restriction, union and difference. His projection operator can drop, permute and repeat columns of a relation. This permuting and repeating of columns does not really add expressive power to the relational algebra. Indeed, using the join operation, one can rewrite any relational algebra expression into an equivalent expression where no projection operator permutes or repeats columns. The fragment of the relational algebra known as the semijoin algebra, however, lacks a full join operation. Nevertheless, we show that any semijoin algebra expression can still be simulated in a natural way by a set of expressions where no projection operator permutes or repeats columns.

We show that the termination of Mohri's algorithm is decidable for polynomially ambiguous weighted finite automata over the tropical semiring which gives a partial answer to a question by Mohri [29]. The proof relies on an improvement of the notion of the twins property and a Burnside type characterization for the finiteness of the set of states produced by Mohri's algorithm.