The cyclic shift of a language L, defined as SHIFT(L) = {vu | uv ∈ L},is an operation known to preserve both regularity and context-freeness.Its descriptional complexity has been addressed in Maslov'spioneering paper on the state complexity of regular language operations[Soviet Math. Dokl.11 (1970) 1373–1375],where a high lower bound for partial DFAs using a growing alphabet was given.We improve this result by using a fixed 4-letter alphabet, obtaining a lower bound (n-1)! . 2(n-1)(n-2),which shows that the state complexity of cyclic shift is2n2+ nlogn - O(n) for alphabets with at least 4 letters.For 2- and 3-letter alphabets, we prove $2^{\Theta(n^2)}$ state complexity.We also establish a tight 2n2+1 lower bound forthe nondeterministic state complexity of this operation using a binary alphabet.

We propose a variation of Wythoff's game on three piles of tokens, in the sense that the losing positions can be derived from the Tribonacci word instead of the Fibonacci word for the two piles game. Thanks to the corresponding exotic numeration system built on the Tribonacci sequence, deciding whether a game position islosing or not can be computed in polynomial time.

We investigate the Laplacian eigenvalues of sparse random graphs Gnp. We show that in the case that the expected degree d = (n-1)p is bounded, the spectral gap of the normalized Laplacian is o(1). Nonetheless, w.h.p. G = Gnp has a large subgraph core(G) such that the spectral gap of is as large as 1-O (d−1/2). We derive similar results regarding the spectrum of the combinatorial Laplacian L(Gnp). The present paper complements the work of Chung, Lu and Vu [8] on the Laplacian spectra of random graphs with given expected degree sequences. Applied to Gnp, their results imply that in the ‘dense’ case d ≥ ln2n the spectral gap of is 1-O (d−1/2) w.h.p.

A simple explicit construction is provided of a partition-valued fragmentation process whose distribution on partitions of [n] = 1,. . .,n at time θ ≥ 0 is governed by the Ewens sampling formula with parameter θ. These partition-valued processes are exchangeable and consistent, as n varies. They can be derived by uniform sampling from a corresponding mass fragmentation process defined by cutting a unit interval at the points of a Poisson process with intensity θx−1dx on/mathbbR+, arranged to beintensifying as θ increases.

Szemerédi's regularity lemma for graphs has proved to be a powerful tool with many subsequent applications. The objective of this paper is to extend the techniques developed by Nagle, Skokan, and the authors and obtain a stronger and more ‘user-friendly’ regularity lemma for hypergraphs.

We continue the study of regular partitions of hypergraphs. In particular, we obtain corresponding counting lemmas for the regularity lemmas for hypergraphs from our paper ‘Regular Partitions of Hypergraphs: Regularity Lemmas’ (in this issue).

A widely studied model for generating binary sequences is to ‘evolve’ them on a tree according to a symmetric Markov process. We show that under this model distinguishing the true (model) tree from a false one is substantially ‘easier’ (in terms of the sequence length needed) than determining the true tree. The key tool is a new and near-tight Ramsey-type result for binary trees.

In 1972, Rosenfeld asked if every triangle-free graph could be embedded in the unit sphere Sd in such a way that two vertices joined by an edge have distance more than (ie, distance more than 2π/3 on the sphere). In 1978, Larman [LAR] disproved this conjecture, constructing a triangle-free graph for which the minimum length of an edge could not exceed . In addition, he conjectured that the right answer would be , which is not better than the class of all graphs. Larman'sconjecture was independently proved by Rosenfeld [MR] and Rödl [VR[. In this last paper it was shown that no bound better than can be found for graphs with arbitrarily large odd girth. We prove in this paper that this is stilltrue for arbitrarily large girth. We discuss then the case of triangle-free graphs with linear minimum degree.

The vertex-nullity interlace polynomial of a graph, described by Arratia, Bollobás and Sorkin in [3] as evolving from questions of DNA sequencing, and extended to a two-variable interlace polynomial by the same authors in [5], evokes many open questions. These include relations between the interlace polynomial and the Tutte polynomial and the computational complexity of the vertex-nullity interlace polynomial. Here, using the medial graph of a planar graph, we relate the one-variable vertex-nullity interlace polynomial to the classical Tutte polynomial when x=y, and conclude that, like the Tutte polynomial, it is in general #P-hard to compute. We also show a relation between the two-variable interlace polynomial and the topological Tutte polynomial of Bollobás and Riordan in [13].

We define the γ invariant as the coefficient of x1 in the vertex-nullity interlace polynomial, analogously to the β invariant, which is the coefficientof x1 in the Tutte polynomial. We then turn to distance hereditary graphs, characterized by Bandelt and Mulder in [9] as being constructed by a sequence ofadding pendant and twin vertices, and show that graphs in this class have γ invariant of 2n+1 when n true twins are added intheir construction. We furthermore show that bipartite distance hereditary graphs are exactly the class of graphs with γ invariant 2, just as the series-parallel graphs are exactly the class of graphs with β invariant 1. In addition, we show that a bipartite distance hereditary graph arises precisely as the circle graph of an Euler circuitin the oriented medial graph of a series-parallel graph. From this we conclude that the vertex-nullity interlace polynomial is polynomial time to compute for bipartite distancehereditary graphs, just as the Tutte polynomial is polynomial time to compute for series-parallel graphs.

We consider the parallel approximability of two problems arisingfrom high multiplicity scheduling, namely the unweightedmodel with variable processing requirements and the weighted model with identical processing requirements. These twoproblems are known to be modelled by a class of quadratic programsthat are efficiently solvable in polynomial time. On the parallelsetting, both problems are P-complete and hence cannot beefficiently solved in parallel unless P = NC. To deal with theparallel approximablity of these problems, we show first aparallel additive approximation procedure to a subclass ofmulti-valued quadratic programming, called smooth multi-valuedQP, which is defined by imposing certain restrictions onthe coefficients of the instance. We use this procedure to obtainparallel approximation to dense instances of the two problems by observing that denseinstances of these problems are instances of smooth multi-valuedQP. The dense instances of the problemsconsidered here are defined similarly as for other combinatorialproblems in the literature. For such instances we can find inparallel a near optimal schedule. The definition of smoothmulti-valued QP as well as the procedure forapproximating it in parallel are of interest independently of theapplication to the scheduling problems considered in this paper.

We study the problem of learning regular tree languages from text. We show that the framework of function distinguishability, as introduced by the author in [Theoret. Comput. Sci.290 (2003) 1679–1711], can be generalized from the case of string languages towards tree languages. This provides a large source of identifiable classes of regular tree languages. Each of these classes can be characterized in various ways. Moreover, we present a generic inference algorithm with polynomial update time and prove its correctness. In this way, we generalize previous works of Angluin, Sakakibara and ourselves. Moreover, we show that this way all regular tree languages can be approximately identified.

Computing the image of a regular language by the transitive closure of a relation is a central question in regular model checking. In a recent paper Bouajjani et al. [IEEE Comput. Soc. (2001) 399–408] proved that the class of regular languages L – called APC – of the form Uj L0,j L1,j L2,j ...Lkj,j , where the union is finite and each Li,j is either a single symbol or a language of the form B* with B a subset of the alphabet, is closed under all semi-commutation relations R. Moreover a recursive algorithm on the regular expressions was given to compute R*(L). This paper provides a new approach, based on automata, for the same problem. Our approach produces a simpler and more efficient algorithm which furthermore works for a larger class of regular languages closed under union, intersection, semi-commutation relations and conjugacy. The existence of this new class, PolC, answers the open question proposed in the paper of Bouajjani et al.

Nous établissons quelques propriétés des mots sturmiens et classifions, ensuite, les mots infinis qui possèdent, pour tout entier naturel non nul n, exactement n+2 facteurs de longueur n. Nous définissons également la notion d'insertion k à k sur les mots infinis puis nous calculons la complexité des mots obtenus en appliquant cette notion aux mots sturmiens. Enfin nous étudions l'équilibre et la palindromie d'une classe particulière de mots de complexité n+2 que nous appelons mots quasi-sturmiens par insertion et que nous caractérisons à l'aide des vecteurs de Parikh.