In this thesis we study the homological behaviour of configuration spaces as the number of objects in the configuration goes to infinity. For unordered configurations of distinct points (possibly equipped with some internal parameters) in a connected, open manifold it is a well-known result, going back to G. Segal and D. McDuff in the 1970s, that these spaces enjoy the property of homological stability. In Chapter 2 we prove that this property also holds for so-called oriented configuration spaces, in which the points of a configuration are equipped with an ordering up to even permutations. There are two important differences from the unordered setting: the rate (or slope) of stabilisation is strictly slower, and the stabilisation maps are not in general split-injective on homology. This can be seen by some explicit calculations of Guest-Kozlowski-Yamaguchi in the case of surfaces. In Chapter 3 we refine their calculations to show that, for an odd prime p, the difference between the mod-p homology of the oriented and the unordered configuration spaces on a surface is zero in a stable range whose slope converges to 1 as p goes to infinity. In Chapter 4 we prove that unordered configuration spaces satisfy homological stability with respect to finite-degree twisted coefficient systems, generalising the corresponding result of S. Betley for the symmetric groups. We deduce this from a general “twisted stability from untwisted stability” principle, which also applies to the configuration spaces studied in the next chapter. In Chapter 5 we study configuration spaces of submanifolds of a background manifold M. Roughly, these are spaces of pairwise unlinked, mutually isotopic copies of a fixed closed, connected manifold P in M. We prove that if the dimension of P is at most (dim(M)−3)/2 then these configuration spaces satisfy homological stability w.r.t. the number of copies of P in the configuration. If P is a sphere this upper bound on its dimension can be increased to dim(M)−3.

La théorie des groupes de tresses s'inscrit au croisement de plusieurs domaines des mathématiques, en particulier, l'algèbre et la géométrie. La recherche actuelle s'étend dans chacune de ces directions, et de riches développements naissent du mariage de ces deux aspects. D'un point de vue géométrique, le groupe des tresses à n brins est vu comme le groupe modulaire d'un disque à n trous, avec composante de bord. On peut représenter une tresse par un diagramme de courbes, c'est-à-dire l'image d'une famille fixée d'arcs sur le disque, par l'élément correspondant du groupe modulaire. Dans cette thèse est présenté l'algorithme de relaxations par la droite, qui permet de retrouver, étant donné un diagramme de courbes, la tresse à partir de laquelle il a été obtenu. Cet algorithme aide à faire le lien entre des propriétés géométriques du diagramme de courbes, et des propriétés algébriques du mot de tresse, en permettant de repérer de grandes puissances d'un générateur sous forme de spirales dans le diagramme de courbes. D'un point de vue algébrique, le groupe de tresses est l'exemple classique de groupe de Garside. L'un des objectifs actuels des recherches en théorie de Garside est d'obtenir un algorithme de résolution en temps polynomial du problème de conjugaison dans les groupes de tresses. À cette fin, on cherche à exploiter les propriétés de certains ensembles finis de conjugués d'une tresse, qui sont des invariants de conjugaison. L'un des résultats de cette thèse concerne la taille d'un de ces invariants, l'ensemble super-sommital : on exhibe une famille de tresses pseudo-anosoviennes dont l'ensemble super-sommital est de taille exponentielle. González-Meneses avait déjà établi le résultat similaire pour une famille de tresses réductibles. La conséquence de ces résultats est qu'on ne peut pas espérer résoudre le problème de conjugaison en temps polynomial au moyen de cet ensemble, et qu'il vaut mieux chercher à exploiter des invariants plus petits. Dans le cas des tresses pseudo-anosoviennes, des espoirs résident actuellement en l'ensemble des circuits glissants. Dans cette thèse, un algorithme en temps polynomial s'appuyant sur ce dernier ensemble résout génériquement le problème de conjugaison, c'est-à-dire qu'il le résout pour une proportion de tresses tendant exponentiellement vite vers 1 lorsque la longueur de la tresse tend vers l'infini. On montre également que, dans une boule du graphe de Cayley avec pour générateurs les tresses simples, une tresse générique est pseudo-anosovienne, ce qui était une conjecture bien connue des spécialistes de la théorie de Garside. / The theory of braid groups is at the intersection of several areas of mathematics, especially algebra and geometry. The current research extends in each of these directions, leading to rich developments. From a geometrical point of view, the braid group on n strands is seen as the mapping class group of a disc with n punctures, with boundary component. A braid can be represented by a curve diagram, that is to say, the image of a family of arcs attached to the disc, by the corresponding mapping class. In this thesis we present the algorithm of relaxations from the right, which, given a curve diagram, determines the braid from which it was obtained. This algorithm helps us to make the link between geometric properties of the curve diagram and algebraic properties of the braid word, allowing us to identify great powers of a generator as spirals in the curve diagram. From an algebraic point of view, the braid group is the classical example of a Garside group. One of the objectives of current research in Garside theory is to obtain a polynomial time algorithm to solve the conjugacy problem in braid groups. For this, a possibility is to exploit the properties of some finite sets of conjugates of a braid, which are invariants of the conjugacy classes. One of the results of this thesis concerns the size of one of these invariants, the super summit set: we construct a family of pseudo-Anosov braids whose super summit set has exponential size. González- Meneses had already established the similar result for a family of reducible braids. These results implies that we cannot hope to solve the conjugacy problem in polynomial time through this set, and it is better to try to use smaller invariants. In the case of pseudo-Anosov braids, one may hope that the so-called sliding circuit set is more useful. In this thesis, we present a polynomial time algorithm based on this last set which generically solves the conjugacy problem, that is to say, it solves it for a proportion of braids that tends exponentially fast to 1 as the length of the braid tends to infinity. We also show that, in a ball of the Cayley graph with generators the simple braids, a braid is generically pseudo-Anosov, which was a well-known conjecture for the specialists in Garside theory.

Neste trabalho, estudamos os grupos de tranças em superfícies visando encontrar apresentações para estes grupos em superfícies fechadas orientáveis de gênero g >= 1 ou superfícies fechadas não orientáveis de gênero g >= 2. Uma consequência destas apresentações é resolvermos o problema da palavra, que consiste em encontrar um algoritmo para decidir quando uma dada palavra num grupo definido por seus geradores e suas relações é a palavra trivial / In this work, we find presentations for surface braid groups either in closed orientable surfaces of genus g >= 1 or in closed non-orientable surfaces of genus g >= 2. A consequence of this presentations is to solve the word problem, which consists in finding an algorithm to decide when a given word in a group defined by its generators and its relations is the trivial word

Dada uma superfície M, definiremos os grupos de tranças de M, denotado por \'B IND. n\' (M), geometricamente e usando a noção de espaços de confiuração. Mostraremos a equivalência das definições. Na mesma linha de raciocínio, definiremos os grupos de tranças puras de superfícies \'P IND. n\' (M). Apresentaremos as propriedades mais importantes dos grupos de tranças do plano e mostraremos que \'B IND. n\' (\'R POT. 2\') injeta em \'B IND. n\' (M), para muitas superfícies M. Mais detalhadamente, obteremos a apresentação de \'B IND. n\' (\'RP POT. 2\' ) e \'P IND. n\'(\'RP POT. 2\') / For a surface M, we define the braid groups of M, \'B IND. n\'(M), geometricaly and using the notion of configuration spaces. We show the equivalence of these definitions. In the sequence, we define the pure braid group of M, \'P IND. n\' (M). We present the most important properties of braid groups of the plane and we show that \'B IND. n\'\'(\'R POT. 2\') embedds in \'B IND. n\' (M), for almost all M. In a more detailed fashion, we present \'B IND. n\' (\'RP POT. 2\') and \'P IND. n\' (\'RP POT. 2)

En 2004, motivés par des constructions de quasimorphismes sur des groupes d'homéomorphismes et de difféomorphismes, Gambaudo et Ghys démontrèrent une formule liant les ω-signatures d'un entrelacs et les propriétés symplectiques d'une représentation du groupe de tresses.Le but de la thèse est d’étendre le résultat de Gambaudo et Ghys en termes d’un invariant algébrique associé à une tresse : la classe de Witt de sa forme de Blanchfield. Il est en effet possible de définir des invariants d'entrelacs en étudiant l'homologie des revêtements cycliques. Les groupes d’homologie et de cohomologie mis en jeu sont munis de structures de modules sur l’anneau du groupe Λ = Z[π].La forme de Blanchfield d’un entrelacs est ainsi la généralisation de la forme d’enlacement définie sur la partie de torsion du premier groupe d’homologie d’une variété fermée de dimension 3. Elle définit alors pour chaque tresse β une classe L(β) dans un groupe de Witt WT(Λ) .Théorème. Soit α et β deux tresses. On a l’égalité suivante, dans WT(Λ) : L(αβ) － L(α) － L(β) = －∂ Meyer(Burau(α), Burau(β)), où le cocycle de Meyer est maintenant défini sur le sous-groupe des éléments de GLn(Λ) préservant la forme de Squier, à valeurs dans le groupe de Witt W(Q(t)). On retrouve essentiellement le résultat original en spécifiant t = ω dans la formule précédente. / In 2004, Gambaudo and Ghys proved a formula establishing a connection between the ω-signatures of a link and the symplectic features of a representation of the braid group. Their main motivation was the construction on quasimorphisms on homeomorphism and diffeomorphism groups.The main goal of this thesis is to extend this result in terms of an algebraic invariant of a braids: the Witt class of the Blanchfield form. Some link invariants are defined through the cyclic covering spaces of their exterior. (Co)homology groups are then equipped with module structures over the ring Λ = Z[π]. For example, the Blanchfield form of a link is a generalisation of the linking form of a 3-manifold, which is a bilinear form on the torsionpart of its first homology group. In particular, every braid β defines a class L(β) in a Witt group WT(Λ) .Theorem. Let α and β be two braids. Then, in WT(Λ):L(αβ) － L(α) － L(β) = －∂ Meyer(Burau(α), Burau(β)),where the Meyer cocycle is defined on the sub group of GLn(Λ) whose elements preserve the Squier form.The result by Gambaudo and Ghys can essentially be recovered from this equality.

Dada uma superfície M, definiremos os grupos de tranças de M, denotado por \'B IND. n\' (M), geometricamente e usando a noção de espaços de confiuração. Mostraremos a equivalência das definições. Na mesma linha de raciocínio, definiremos os grupos de tranças puras de superfícies \'P IND. n\' (M). Apresentaremos as propriedades mais importantes dos grupos de tranças do plano e mostraremos que \'B IND. n\' (\'R POT. 2\') injeta em \'B IND. n\' (M), para muitas superfícies M. Mais detalhadamente, obteremos a apresentação de \'B IND. n\' (\'RP POT. 2\' ) e \'P IND. n\'(\'RP POT. 2\') / For a surface M, we define the braid groups of M, \'B IND. n\'(M), geometricaly and using the notion of configuration spaces. We show the equivalence of these definitions. In the sequence, we define the pure braid group of M, \'P IND. n\' (M). We present the most important properties of braid groups of the plane and we show that \'B IND. n\'\'(\'R POT. 2\') embedds in \'B IND. n\' (M), for almost all M. In a more detailed fashion, we present \'B IND. n\' (\'RP POT. 2\') and \'P IND. n\' (\'RP POT. 2)

Neste trabalho, estudamos os grupos de tranças em superfícies visando encontrar apresentações para estes grupos em superfícies fechadas orientáveis de gênero g >= 1 ou superfícies fechadas não orientáveis de gênero g >= 2. Uma consequência destas apresentações é resolvermos o problema da palavra, que consiste em encontrar um algoritmo para decidir quando uma dada palavra num grupo definido por seus geradores e suas relações é a palavra trivial / In this work, we find presentations for surface braid groups either in closed orientable surfaces of genus g >= 1 or in closed non-orientable surfaces of genus g >= 2. A consequence of this presentations is to solve the word problem, which consists in finding an algorithm to decide when a given word in a group defined by its generators and its relations is the trivial word

Cette thèse concerne principalement deux objets classiques étroitement liés: d'une part la variété des polynômes complexes unitaires de degré $d>1$ à une variable, et à racines simples (donc de discriminant différent de zéro), et d'autre part, les groupes de tresses d'Artin avec d brins. Le travail présenté dans cette thèse propose une nouvelle approche permettant des calculs cohomologiques explicites à coefficients dans n'importe quel faisceau. En vue de calculs cohomologiques explicites, il est souhaitable d'avoir à sa disposition un bon recouvrement au sens de Čech. L'un des principaux objectifs de cette thèse est de construire un tel recouvrement basé sur des graphes (appelés signatures) qui rappellent les `dessins d'enfant' et qui sont associées aux polynômes complexes classifiés par l'espace de polynômes. Cette décomposition de l'espace de polynômes fournit une stratification semi-algébrique. Le nombre de composantes connexes de chaque strate est calculé dans le dernier chapitre ce cette thèse. Néanmoins, cette partition ne fournit pas immédiatement un recouvrement adapté au calcul de la cohomologie de Čech (avec n'importe quels coefficients) pour deux raisons liées et évidentes: d'une part les sous-ensembles du recouvrement ne sont pas ouverts, et de plus ils sont disjoints puisqu'ils correspondent à différentes signatures. Ainsi, l'objectif principal du chapitre 6 est de ``corriger'' le recouvrement de départ afin de le transformer en un bon recouvrement ouvert, adapté au calcul de la cohomologie Čech. Cette construction permet ensuite un calcul explicite des groupes de cohomologie de Čech à valeurs dans un faisceau localement constant. / This thesis mainly concerns two closely related classical objects: on the one hand, the variety of unitary complex polynomials of degree $ d> 1 $ with a variable, and with simple roots (hence with a non-zero discriminant), and on the other hand, the $d$ strand Artin braid groups. The work presented in this thesis proposes a new approach allowing explicit cohomological calculations with coefficients in any sheaf. In order to obtain explicit cohomological calculations, it is necessary to have a good cover in the sense of Čech. One of the main objectives of this thesis is to construct such a good covering, based on graphs that are reminiscent of the ''dessins d'enfants'' and which are associated to the complex polynomials. This decomposition of the space of polynomials provides a semi-algebraic stratification. The number of connected components in each stratum is counted in the last chapter of this thesis. Nevertheless, this partition does not immediately provide a ''good'' cover adapted to the computation of the cohomology of Čech (with any coefficients) for two related and obvious reasons: on the one hand the subsets of the cover are not open, and moreover they are disjoint since they correspond to different signatures. Therefore, the main purpose of Chapter 6 is to ''correct'' the cover in order to transform it into a good open cover, suitable for the calculation of the Čech cohomology. It is explicitly verified that there is an open cover such that all the multiple intersections are contractible. This allows an explicit calculation of cohomology groups of Čech with values in a locally constant sheaf.

A configuration space is a space whose points represent the possible states of a given physical system. As such they appear naturally both in theoretical physics and technical applications. For an example of the former, in analytical mechanics, the Lagrangian and Hamiltonian formulations of classical mechanics depend heavily on the use of a physical system’s configuration space for the description of its kinematical and dynamical behavior, and importantly, its evolution in time. As an example of a technical application, consider robotics, where the space of possible configurations of the mechanical linkages that make up a robot is an important tool in motion planning. In this case it is of particular interest to study the singularities of these mechanical linkages, to see if a given configuration is singular or not. This can be done with the help of configuration spaces and their topological properties. Arguably, the simplest configuration space possible arises when the system is just a collection of point-like particles in a plane. Despite its simplicity, the corresponding configuration space has substantial complexity and is of great interest in mathematics, physics and technology: For instance, it arises naturally in the mathematical modelling of robots performing tasks in a warehouse. In this thesis we go through the mathematics necessary to study the behaviour of paths in this space, which corresponds to motions of the particles. We use the theory of groups, algebraic topology, and manifolds to examine the properties of the configuration space of point-like particles in a plane. An important role in the discussion will be played by braids, which are certain collections of curves, interlaced in three-space. They are connected to many different topics in algebra, geometry, and mathematical physics, such as representation theory, the Yang-Baxter equation and knot theory. They are also important in their own right. Here we focus on their relation to configurations of points.

A relação entre os grupos de tranças de superfícies e os grupos de homotopia das esferas é atualmente um tópico de bastante interesse. Nos últimos anos tem sido feitos avanços consideráveis no estudo desta relação no caso dos grupos de tranças de Artin com n cordas, denotado por Bn, da esfera e do plano projetivo. Nessa tese analisamos com detalhes as interações entre a teoria de tranças e a teoria de homotopia, e mostramos novos resultados que estabelecem conexões entre os grupos de homotopia da 2-esfera S2 e os grupos de tranças sobre qualquer superfície. No andamento deste trabalho, descobrimos uma conexão surpreendente dos grupos de tranças com os grupos cristalográficos e de Bieberbach: para n maior ou igual que 3, o grupo quociente Bn/[Pn, Pn] é um grupo cristalográfico que contém grupos de Bieberbach como subgrupos, onde Pn é o subgrupo de tranças puras de Bn. Com isto obtivemos uma formulação de um Teorema de Auslander e Kuranishi para 2-grupos finitos e exibimos variedades Riemannianas compactas planas que admitem difeomorfismo de Anosov e cujo grupo de holonomia é Z2k . Além disso, durante esta tese, detectamos e, quando possível, corrigimos algumas imprecisões em dois importantes artigos nessa área de estudo, escritos por J. Berrick, F. R. Cohen, Y. L. Wong e J. Wu (Jour. Amer. Math. Soc. - 2006) assim como por J. Y. Li e J.Wu (Proc. London Math. Soc. - 2009). / The relation between surface braid groups and homotopy groups of spheres is currently a subject of great interest. Considerable progress has been made in recent years in the study of these relations in the case of the n-string Artin braid groups, denoted by Bn, the sphere and the projective plane. In this thesis we analyse in detail the interactions between braid theory and homotopy theory, and we present new results that establish connections between the homotopy groups of the 2-sphere S2 and the braid groups of any surface. During the course of this work, we discovered an unexpected connection of braid groups with crystallographic and Bieberbach groups: for n greater or equal than 3, the quotient group Bn/[Pn, Pn] is a crystallographic group that contains Bieberbach groups as subgroups, where Pn is the pure braid subgroup of Bn. This enables us to obtain a formulation of a theorem of Auslander and Kuranishi for finite 2-groups, and to exhibit Riemannian compact flat manifolds that admit Anosov diffeomorphisms and whose holonomy group is Z2k. In addition, during the thesis, we have detected, and where possible, corrected some inaccuracies in two important papers in the area of study, by J. Berrick, F. R. Cohen, Y. L. Wong and J. Wu (Jour. Amer. Math. Soc. - 2006), and by J. Y. Li and J. Wu (Proc. London Math. Soc. - 2009).