Etant donné trois entiers k, s et a, nous prouvons dans le premier chapitre qu'il existe un graphe k-régulier fini (resp. un graphe k-régulier connexe fini) dont le groupe d'automorphismes a exactement s orbites sur l'ensemble des sommets et a orbites sur l'ensemble des arêtes si et seulement si (s,a)=(1,0) si k=0, (s,a)=(1,1) si k=1, s=a>0 si k=2, 0< s <= 2a <= 2ks si k>2. (resp. (s,a)=(1,0) si k=0, (s,a)=(1,1) si k=1 ou 2, s-1<=a<=(k-1)s+1 et s,a>0 si k>2.) Nous étudions les polyèdres convexes de R³ dans le second chapitre. Pour tout polyèdre convexe P, nous notons Isom(P) l'ensemble des isométries de R³ laissant P invariant. Si G est un sous-groupe de Isom(P), le f_G-vecteur de P est le triple d'entiers (s,a,f) tel que G ait exactement s orbites sur l'ensemble sommets de P, a orbites sur l'ensemble des arêtes de P et f orbites sur l'ensemble des faces de P. Remarquons que (s,a,f) est le f_{id}-vecteur (appelé f-vecteur dans la littérature) d'un polyèdre si ce dernier possède exactement s sommets, a arêtes et f faces. Nous généralisons un théorème de Steinitz décrivant tous les f-vecteurs possibles. Pour tout groupe fini G d'isométries de R³, nous déterminons l'ensemble des triples (s,a,f) pour lesquels il existe un polyèdre convexe ayant (s,a,f) comme f_G-vecteur. Ces résultats nous permettent de caractériser les triples (s,a,f) pour lesquels il existe un polyèdre convexe tel que Isom(P) a s orbites sur l'ensemble des sommets, a orbites sur l'ensemble des arêtes et f orbites sur l'ensemble des faces. La structure d'incidence I(P) associée à un polyèdre P consiste en la donnée de l'ensemble des sommets de P, l'ensemble des arêtes de P, l'ensemble des faces de P et de l'inclusion entre ces différents éléments (la notion de distance ne se trouve pas dans I(P)). Nous déterminons également l'ensemble des triples d'entiers (s,a,f) pour lesquels il existe une structure d'incidence I(P) associée à un polyèdre P dont le groupe d'automorphismes a exactement s orbites de sommets, a orbites d'arêtes et f orbites de sommets.

If every vertex in a graph G has the same degree, then the graph is called a regular graph. That is, if deg(v) = r for all vertices in the graph, then it is denoted as an r-regular graph. A graph G is said to be vertex-weighted if all of the vertices are assigned weights. A generalized definition for degree regularity for vertex-weighted graphs can be stated as follows: A vertex-weighted graph is said to be rw-regular if the sum of the weights in the neighborhood of every vertex is rw. If all vertices are assigned the unit weight of 1, then this is equivalent to the definition for r-regular graphs. In this thesis, we determine if a graph has a weighting scheme that makes it a weighted regular graph or prove no such scheme exists for a number of special classes of graphs such as paths, stars, caterpillars, spiders and wheels.

In 1975, Sheehan conjectured that every simple 4-regular hamiltonian graph has a second Hamilton cycle. If Sheehan's Conjecture holds, then the result can be extended to all simple d-regular hamiltonian graphs with d at least 3. First, we survey some previous results which verify the existence of a second Hamilton cycle if d is large enough. We will then demonstrate some techniques for finding a second Hamilton cycle that will be used throughout this paper. Finally, we use these techniques and show that for certain 4-regular Hamiltonian graphs whose automorphism group is large enough, a second Hamilton cycle exists.

This thesis is concerned with the quantum Dirac magnetic monopole and two classes of its generalisations. The first of these are certain analogues of the Dirac magnetic monopole on coadjoint orbits of compact Lie groups, equipped with the normal metric. The original Dirac magnetic monopole on the unit sphere S^2 corresponds to the particular case of the coadjoint orbits of SU(2). The main idea is that the Hilbert space of the problem, which is the space of L^2-sections of a line bundle over the orbit, can be interpreted algebraically as an induced representation. The spectrum of the corresponding Schodinger operator is described explicitly using tools of representation theory, including the Frobenius reciprocity and Kostant's branching formula. In the second part some discrete versions of Dirac magnetic monopoles on S^2 are introduced and studied. The corresponding quantum Hamiltonian is a magnetic Schodinger operator on a regular polyhedral graph. The construction is based on interpreting the vertices of the graph as points of a discrete homogeneous space G/H, where G is a binary polyhedral subgroup of SU(2). The edges are constructed using a specially selected central element from the group algebra, which is used also in the definition of the magnetic Schrodinger operator together with a character of H. The spectrum is computed explicitly using representation theory by interpreting the Hilbert space as an induced representation.

In 1975, Sheehan conjectured that every simple 4-regular hamiltonian graph has a second Hamilton cycle. If Sheehan's Conjecture holds, then the result can be extended to all simple d-regular hamiltonian graphs with d at least 3. First, we survey some previous results which verify the existence of a second Hamilton cycle if d is large enough. We will then demonstrate some techniques for finding a second Hamilton cycle that will be used throughout this paper. Finally, we use these techniques and show that for certain 4-regular Hamiltonian graphs whose automorphism group is large enough, a second Hamilton cycle exists.

Let n be a positive integer. A Latin square of order n is an n×n array L such that each element of some n-set occurs in each row and in each column of L exactly once. It is well-known that one may construct a 4-class association scheme on the positions of a Latin square, where the relations are the identity, being in the same row, being in the same column, having the same entry, and everything else. We describe the subconstituent (Terwilliger) algebras of such an association scheme. One also may construct several strongly regular graphs on the positions of a Latin square, where adjacency corresponds to any subset of the nonidentity relations described above. We describe the local spectrum and subconstituent algebras of such strongly regular graphs. Finally, we study various notions of isomorphism for subconstituent algebras using Latin squares as examples.

It is a well established fact, that – in the case of classical random graphs like variants of Gn,p or random regular graphs – spectral methods yield efficient algorithms for clustering (e. g. colouring or bisec- tion) problems. The theory of large networks emerging recently provides convincing evidence that such networks, albeit looking random in some sense, cannot sensibly be described by classical random graphs. A vari- ety of new types of random graphs have been introduced. One of these types is characterized by the fact that we have a fixed expected degree sequence, that is for each vertex its expected degree is given. Recent theoretical work confirms that spectral methods can be success- fully applied to clustering problems for such random graphs, too – pro- vided that the expected degrees are not too small, in fact &ge; log⁶ n. In this case however the degree of each vertex is concentrated about its expectation. We show how to remove this restriction and apply spectral methods when the expected degrees are bounded below just by a suitable constant. Our results rely on the observation that techniques developed for the classical sparse Gn,p random graph (that is p = c/n) can be transferred to the present situation, provided we consider a suitably normalized ad- jacency matrix: We divide each entry of the adjacency matrix by the product of the expected degrees of the incident vertices. Given the host of spectral techniques developed for Gn,p this observation should be of independent interest.

It is a well established fact, that – in the case of classical random graphs like variants of Gn,p or random regular graphs – spectral methods yield efficient algorithms for clustering (e. g. colouring or bisec- tion) problems. The theory of large networks emerging recently provides convincing evidence that such networks, albeit looking random in some sense, cannot sensibly be described by classical random graphs. A vari- ety of new types of random graphs have been introduced. One of these types is characterized by the fact that we have a fixed expected degree sequence, that is for each vertex its expected degree is given. Recent theoretical work confirms that spectral methods can be success- fully applied to clustering problems for such random graphs, too – pro- vided that the expected degrees are not too small, in fact &ge; log⁶ n. In this case however the degree of each vertex is concentrated about its expectation. We show how to remove this restriction and apply spectral methods when the expected degrees are bounded below just by a suitable constant. Our results rely on the observation that techniques developed for the classical sparse Gn,p random graph (that is p = c/n) can be transferred to the present situation, provided we consider a suitably normalized ad- jacency matrix: We divide each entry of the adjacency matrix by the product of the expected degrees of the incident vertices. Given the host of spectral techniques developed for Gn,p this observation should be of independent interest.

Pairwise comparison matrices (PCMs) play a key role in multi-criteria decision making, especially in the analytic hierarchy process. It could be necessary for an expert to compare alternatives based on various criteria. However, for a variety of reasons, such as lack of time or insufficient knowledge, it may happen that the expert cannot provide judgments on all pairs of alternatives. In this case, an incomplete pairwise comparison matrix is formed. In the first research part, an optimization algorithm is proposed for the optimal completion of an incomplete PCM. It is intended to numerically minimize a constrained eigenvalue problem, in which the objective function is difficult to write explicitly in terms of variables. Numerical simulations are carried out to examine the performance of the algorithm. The simulation results show that the proposed algorithm is capable of solving the minimization of the constrained eigenvalue problem. In the second part, a comparative analysis of eleven completion methods is studied. The similarity of the eleven completion methods is analyzed on the basis of numerical simulations and hierarchical clustering. Numerical simulations are performed for PCMs of different orders considering various numbers of missing comparisons. The results suggest the existence of a cluster of five extremely similar methods, and a method significantly dissimilar from all the others. In the third part, the filling in patterns (arrangements of known comparisons) of incomplete PCMs based on their graph representation are investigated under given conditions: regularity, diameter and number of vertices, but without prior information. Regular and quasi-regular graphs with minimal diameter are proposed. Finally, the simulation results indicate that the proposed graphs indeed provide better weight vectors than alternative graphs with the same number of comparisons. This research problem’s contributions include a list of (quasi-)regular graphs with diameters of 2 and 3, and vertices from 5 up to 24.

Uma decomposição de um grafo G é um conjunto D = {H_1,... , H_k } de subgrafos de G dois-a-dois aresta-disjuntos que cobre o conjunto das arestas de G. Se H_i é isomorfo a um grafo fixo H, para 1<=i<=k, então dizemos que D é uma H-decomposição de G. Neste trabalho, estudamos o caso em que H é um caminho de comprimento fixo. Para isso, primeiramente decompomos o grafo dado em trilhas, e depois fazemos uso de um lema de desemaranhamento, que nos permite transformar essa decomposição em trilhas numa decomposição somente em caminhos. Com isso, obtemos resultados para três conjecturas sobre H-decomposição de grafos no caso em que H=P_\\ell é o caminho de comprimento \\ell. Dois desses resultados resolvem versões fracas das Conjecturas de Kouider e Lonc (1999) e de Favaron, Genest e Kouider (2010), ambas para grafos regulares. Provamos que, para todo inteiro positivo \\ell, (i) existe um inteiro positivo m_0 tal que se G é um grafo 2m\\ell-regular com m>=m_0, então G admite uma P_\\ell-decomposição; (ii) se \\ell é ímpar, existe um inteiro positivo m_0 tal que se G é um grafo m\\ell-regular com m>=m_0, e G contém um m-fator, então G admite uma P_\\ell-decomposição. O terceiro resultado diz respeito a grafos altamente aresta- conexos: existe um inteiro positivo k_\\ell tal que se G é um grafo k_\\ell-aresta-conexo cujo número de arestas é divisível por \\ell, então G admite uma P_\\ell-decomposição. Esse resultado prova que a Decomposition Conjecture de Barát e Thomassen (2006), formulada para árvores, é verdadeira para caminhos. / A decomposition of a graph G is a set D = {H_1,...,H_k} of pairwise edge-disjoint subgraphs of G that cover the set of edges of G. If H_i is isomorphic to a fixed graph H, for 1<=i<=k, then we say that D is an H-decomposition of G. In this work, we study the case where H is a path of fixed length. For that, we first decompose the given graph into trails, and then we use a disentangling lemma, that allows us to transform this decomposition into one consisting only of paths. With this approach, we tackle three conjectures on H-decomposition of graphs and obtain results for the case H=P_\\ell is the path of length \\ell. Two of these results solve weakenings of a conjecture of Kouider and Lonc (1999) and a conjecture of Favaron, Genest and Kouider (2010), both for regular graphs. We prove that, for every positive integer \\ell, (i) there is a positive integer m_0 such that, if G is a 2m\\ell-regular graph with m>=m_0, then G admits a P_\\ell-decomposition; (ii) if \\ell is odd, there is a positive integer m_0 such that, if G is an m\\ell-regular graph with m>=m_0 containing an m-factor, then G admits a P_\\ell-decomposition. The third result concerns highly edge-connected graphs: there is a positive integer k_\\ell such that if G is a k_\\ell-edge-connected graph whose number of edges is divisible by \\ell, then G admits a P_\\ell-decomposition. This result verifies for paths the Decomposition Conjecture of Barát and Thomassen (2006), on trees.