La thèse porte sur le raccordement de surfaces canal en modélisation géométriques en utilisant des morceaux de cyclides de Dupin. Elle tente de répondre à un problème de reconstruction de pièces controlées et usinées par le CEA de Valduc. En se plaçant dans l'espace adéquat, l'espace des sphères, dans lequel nous pouvons manipuler à la fois les points, les sphères et les surfaces canal, nous simplifions considérablement certains problèmes. Cet espace est représenté par une quadrique de dimension 4 dans un espace de dimension 5, muni de la forme de Lorentz : l'espace de Lorentz. Dans l'espace des sphères, les problèmes de recollements de surfaces canal par des morceaux de cyclides de Dupin se simplifient en problèmes linéaires. Nous donnons les algorithmes permettant de réaliser ce type de jointures en utilisant l'espace des sphères puis nous revenons dans l'espace à 3 dimensions usuel. Ces jointures se font toujours le long de cercles caractéristiques des surfaces considérées. En résolvant le problème dit des trois conditions de contact, nous mettons en évidence une autre courbe particulière, sur une famille à un paramètre de cyclides, que nous appellons courbe de contact qui permettrait d'effectuer des jointures le long d'autres courbes / The thesis deals with the blending of canal surfaces in geometric modeling using pieces of Dupin Cyclides. We try to solve a problem of reconstructing real parts manufactured and controlled by the CEA of Valduc. Using the space of spheres in which we can manipulate both points, spheres and canal surfaces, we simplify some problems. This space is represented by a 4-dimensional quadric in a 5-dimensional space, equipped with the Lorentz form, it is the Lorentz space. In the space of spheres, problems of blending canal surfaces by pieces of Dupin cyclides are simplified in linear problems. We give algorithms to make such blends using the space of spheres and after we come back to 3 dimensions to draw the result. These blends are always made along characteristics circles of the considered surfaces. By solving the problem of three contact conditions, we highlight another particular curve, on a one parameter familly of cyclides, that we call contact curve along which we could also make this kind of blends

Ce travail traite de la modélisation de structures composites intégrant des éléments amortissants passifs. Un modèle de plaque "équivalent simple couche" générique utilisant des fonctions de description du cisaillement transverse est présenté. Plusieurs méthodes d'obtention de ces fonctions sont décrites, permettant de retrouver des modèles classiques ou issus de la littérature. Deux nouvelles méthodes d'obtention de ces fonctions sont aussi présentées.Plusieurs méthodes de discrétisation adaptées au modèle générique sont étudiées. La méthode de Navier permet de tester la qualité de chaque modèle associé à un jeu de fonctions de description du cisaillement transverse. La méthode de Rayleigh-Ritz permet l'étude du comportement vibratoire d'une plaque rectangulaire munie d'un ou plusieurs patchs viscocontraints. Plusieurs éléments finis issus de la littérature, adaptés au modèle, sont aussi présentés.À l'aide de la méthode de Navier, une étude numérique du comportement statique et dynamique de plusieurs configurations de plaques permet la comparaison des différents modèles présentés. La méthode de Rayleigh-Ritz est utilisée pour étudier le comportement vibratoire d'une plaque munie d'un patch viscocontraint. Une comparaison des résultats obtenus avec le modèle présenté et ceux issus de calculs éléments finis tridimensionnels permet de valider notre modèle. Une étude énergétique de la plaque patchée permet d'illustrer le comportement du patch. Enfin une méthode inverse d'identification des matériaux viscoélastiques, basées sur une combinaison du modèle décrit et d'un algorithme génétique, montre une application du modèle. / This work is on the subject of modelization of structures treated with passive damping elements. A generic "equivalent single layer" plate model using transverse shear warping functions is presented. Several methods to obtain these functions are described, allowing the implementation of classical models and others issued from the litterature. Two new methods for obtaining these functions are also presented.Several discretization methods adapted to the generic plate model are studied. Navier's procedure allows the testing of the quality of each model associated with a set of transverse shear warping functions. Rayleigh-Ritz method allows the study of the vibrational behavior of a rectangular plate treated with one or several constrained damping patches. Several finite elements issued from the literature are also presented.Using Navier's procedure, a numerical study of the static and dynamic behavior of several plate configurations allows the comparison of the different plate models. Rayleigh-Ritz method is used to study the vibrational response of a plate treated with a constrained damping patch. A comparison of the results with those obtained with three dimensional finite element calculations permits the model validation. An energetic study of the patched plate allow us to understand the constrainted damping patch behavior. Finally, an inverse method, allowing the identification of the properties of viscoelastic materials, based on a combination of the presented model and a genetic algorithm, shows a possible application of the model.

Nous nous intéressons à la modélisation de formes complexes de type structures arborescences, formes lacunaires ou surfaces rugueuses. Ces formes sont intéressantes de par leurs propriétés physiques particulières :objets légers, économie de matière, résistance mécanique, absorption acoustique importante. Les modèles basés sur le concept de la géométrie fractale permettent de générer de telles formes et notamment les formes auto-similaires. A partir des travaux de Barnsley sur les systèmes itérés de fonctions, Tosan et al, ont proposé une extension, Boundary Controled Iterated Funcions Systems (BCIFS) pour contrôler plus facilement les formes et faciliter leur description. Nous nous intéressons aux propriétés différentielles des formes décrites par BCIFS. Nous proposons une définition plus générale d'espace tangent qui permet de caractériser le comportement de cas non-classiquement différentiables.Nous montrons que l'étude du comportement différentiel peut alors se faire simplement par analyse des valeurs propres et vecteurs propres généralisés des opérateurs de subdivision. Il devient alors possible de contrôler ces propriétés différentielles. Nous présentons une application de nos résultats, en proposant une méthode pour construire des raccords entre deux structures définies par des processus de subdivision différents. Cette méthode est appliquée pour la construction d'un raccord entre une surface de subdivision de Doo-Sabin(schéma dual) et une surface de subdivision de Catmull-Clark (schéma primal) / The fractal geometry is a relatively new branch of mathematics that studies complex objects of non-integer dimensions. It finds applications in many branches of science as objects of such complex structure often poses interesting properties. In 1988 Barnsley presented the Iterative Func-tion System (IFS) model that allows modelling complex fractal shapes with only a limited set of contractive transformations. Later many other models were based on the IFS model such as Language-Restricted IFS,Projective IFS, Controlled IFS and Boundary Controlled IFS. The lastto allow modelling complex shapes with control points and specific topol-ogy. These models cover classical geometric models such as B-splines and subdivision surfaces as well as fractal shapes.This thesis focuses on the analysis of the differential behaviour of the shapes described with Controlled IFS and Boundary Controlled IFS. Wederive the necessary and sufficient conditions for differentiability for ev-erywhere dense set of points. Our study is based on the study of the eigenvalues and eigenvectors of the transformations composing the IFS. We apply the obtained conditions to modelling curves in surfaces. We describe different examples of differential behaviour presented in shapes modelled with Controlled IFS and Boundary Controlled IFS. We also use the Boundary Controlled IFS to solve the problem of connecting different subdivision schemes. We construct a junction between Doo-Sabin and Catmull-Clark subdivision surfaces and analyse the differential behaviour of the intermediate surface

Le but de cette thèse est de fournir une caractérisation topologique de tresses virtuelles. Les tresses virtuelles sont des classes d’équivalence de diagrammes de type tresses tracés sur le plan. La relation d’équivalence est générée par l’isotopie, les mouvements de Reidemeister et les mouvements de Reidemeister virtuels. L’ensemble des tresses virtuelles est munie d’une opération de groupe. On parlera alors du groupe de tresses virtuelles. Dans le Chapitre 1, nous introduisons les notions de base de la théorie de noeuds virtuels, nous évoquons certains propriétés du groupe tresses virtuelles, et des liens qu’il a avec le groupe de tresses classiques. Dans le Chapitre 2, nous introduisons la notion de diagramme de Gauss tressé (ou diagramme de Gauss horizontal), et on démontre qu’il s’agit là d’une bonne réinterprétation combinatoire pour les tresses virtuelles. On généralise en particulier certains résultats connus en théorie de noeuds virtuels. Un application est de retrouver la présentation classique du groupe de tresses virtuelles pures à l’aide des diagrammes de Gauss tressés. Dans le Chapitre 3, on introduit les tresses abstraites et on montre qu’elles sont en correspondance bijective avec les tresses virtuelles. Les tresses abstraites sont des classes d’équivalence des diagrammes de type tresses tracés sur une surface orientable avec deux composantes de bord. La relation d’équivalence est générée par l’isotopie, la compatibilité, la stabilité et les mouvements de Reidemeister. La compatibilité est la relation d’équivalence générée par les difféomorphismes préservant l’orientation. La stabilité est la relation d’équivalence générée par l’addition ou la suppression d’anses à la surface, dans le complémentaire du diagramme. Dans le Chapitre 4, on démontre que tout tresse abstraite admets une unique représentant de genre minimal, à compatibilité et mouvements de Reidemeister prés. En particulier, les tresses classiques se plongent dans les tresses abstraites. / The purpose of this thesis is to give a topological characterization of virtual braids. Virtual braids are equivalence classes of planar braid-like diagrams identified up to isotopy, Reidemeister and virtual Reidemeister moves. The set of virtual braids admits a group structure and is called the virtual braid group. In Chapter 1 we present a general introduction to the theory of virtual knots, and we discuss some properties of virtual braids and their relations with classical braids. In Chapter 2 we introduce braid-Gauss dia- grams, and we prove that they are a good combinatorial reinterpretation of virtual braids. In particular this generalizes some results known in virtual knot theory. As an application, we use braid-Gauss diagrams to recover a well known presentation of the pure virtual braid group. In Chapter 3 we introduce abstract braids and we prove that they are in a bijective cor- respondence with virtual braids. Abstract braids are equivalence classes of braid-like diagrams on an orientable surface with two boundary components. The equivalence relation is generated by isotopy, compatibility, stability and Reidemeister moves. Compatibility is the equivalence relation generated by orientation preserving diffeomorphisms. Stability is the equivalence relation generated by adding handles to or deleting handles from the surface in the complement of the braid-like diagram. In Chapter 4 we prove that for any abstract braid, there is a unique representative of minimal genus, up to compatibility and Reidemeister equivalence. In particular this implies that classical braids embed in abstract braids.

In this thesis, we investigate methods of finding a local minimum for unconstrained problems of non-convex functions with n variables, by following the solution curve of a system of ordinary differential equations. The motivation for this was the fact that existing methods (e.g. those based on Newton methods with line search) sometimes terminate at a non-stationary point when applied to functions f(x) that do not a have positive-definite Hessian (i.e. ∇²f → 0) for all x. Even when methods terminate at a stationary point it could be a saddle or maximum rather than a minimum. The only method which makes intuitive sense in non-convex region is the trust region approach where we seek a step which minimises a quadratic model subject to a restriction on the two-norm of the step size. This gives a well-defined search direction but at the expense of a costly evaluation. The algorithms derived in this thesis are gradient based methods which require systems of equations to be solved at each step but which do not use a line search in the usual sense. Progress along the Continuous Steepest Descent Path (CSDP) is governed both by the decrease in the function value and measures of accuracy of a local quadratic model. Numerical results on specially constructed test problems and a number of standard test problems from CUTEr [38] show that the approaches we have considered are more promising when compared with routines in the optimization tool box of MATLAB [46], namely the trust region method and the quasi-Newton method. In particular, they perform well in comparison with the, superficially similar, gradient-flow method proposed by Behrman [7].

We develop expansion methods as a new computational approach towards high-dimensional partial differential equations (PDEs), particularly of such type as arising in the valuation of financial derivatives. The proposed methods are extended from [41] and use principal component analysis (PCA) of the underlying process in combination with a Taylor expansion of the value function into solutions to low-dimensional PDEs. They enable calculation of highly accurate approximate solutions with computational complexity polynomial in the number of dimensions for PDEs with a low number of dominant principal components. For the case of PDEs with constant coefficients, we show existence of expansion solutions and prove theoretical error bounds. We give a precise characterisation of when our methods can be applied and construct specific examples of a first and second order version. We provide numerical results showing that the empirically observed convergence speeds are in agreement with the theoretical predictions. For the case of PDEs with varying coefficients, we give a heuristic motivation using the Parametrix approach and empirically test the methods' accuracy for a range of variable parameter stock models. We demonstrate the applicability of our expansion methods to real-world securities pricing problems by considering path-dependent and early-exercise options in the LIBOR market model. Using the example of Bermudan swaptions and Ratchet floors, which are considered difficult benchmark problems, we give a careful analysis of the numerical accuracy and computational complexity. We are able to demonstrate that for problems with medium to high dimensionality, around 60-100, and moderate time horizons, the presented PDE methods deliver results comparable in accuracy to benchmark state-of-the-art Monte Carlo methods in similar or (significantly) faster run time.

Computer simulation of electromagnetic behaviour of a device is a common practice in modern engineering. Maxwell's equations are solved on a computer with help of numerical methods. Contemporary devices constantly grow in size and complexity. Therefore, new numerical methods should be highly efficient. Many industrial and research applications of numerical methods need to account for the frequency dependent materials. The Finite-Difference Time-Domain (FDTD) method is one of the most widely adopted algorithms for the numerical solution of Maxwell's equations. A major drawback of the FDTD method is the interdependence of the spatial and temporal discretisation steps, known as the Courant-Friedrichs-Lewy (CFL) stability criterion. Due to the CFL condition the simulation of a large object with delicate geometry will require a high spatio-temporal resolution everywhere in the FDTD grid. Application of subgridding increases the efficiency of the FDTD method. Subgridding decomposes the simulation domain into several subdomains with different spatio-temporal resolutions. The research project described in this dissertation uses the Huygens Subgridding (HSG) method. The frequency dependence is included with the Auxiliary Differential Equation (ADE) approach based on the one-pole Debye relaxation model. The main contributions of this work are (i) extension of the one-dimensional (1D) frequency-dependent HSG method to three dimensions (3D), (ii) implementation of the frequency-dependent HSG method, termed the dispersive HSG, in Fortran 90, (iii) implementation of the radio environment setting from the PGM-files, (iv) simulation of the electromagnetic wave propagating from the defibrillator through the human torso and (v) analysis of the computational requirements of the dispersive HSG program.

We determine the asymptotic limit of a sequence of weak solutions for the interaction problem between several rigid bodies and an incompressible fluid in an arbitrary domain \(\Omega\subseteq\mathbb{R}^d\) containing them. The limit is performed by letting the maximal diameter of the rigid bodies tend to zero and their number is allowed to grow to infinity, bounded by the logarithm of the maximal diameter. Both the fluid and the rigid bodies are allowed to be inhomogeneous and the viscosity of the fluid is allowed to depend on its density. The initial densities shall be bounded in \(L^{p_0}(\Omega)\) for some \(d<p_0\leq\infty\) and converge in \(L^1(\Omega)\), the initial momenta are assumed to converge weakly in \(L^2(\Omega)^d\) and the initial energies shall converge. Under these assumptions, the rigid bodies are negligible in the limit, meaning that up to a subsequence, density and velocity for the fluid-structure interaction problem converge weakly to a solution of the inhomogeneous incompressible Navier-Stokes equations. The result holds in any dimension \(d\geq2\). We apply the asymptotic limit to the special case of initially periodically distributed rigid bodies of equal shape and obtain that the rigid bodies are negligible in the asymptotic limit, provided their size tends to zero fast enough compared to the period of the initial distribution. / Wir bestimmen den asymptotischen Grenzwert für eine Folge schwacher Lösungen des Problems der Wechselwirkung mehrerer starrer Körper mit einer inkompressiblen Flüssigkeit in einem beliebigen Gebiet \(\Omega\subseteq\mathbb{R}^d\), welche jene enthält. Der Grenzwert wird derart gebildet, dass der maximale Durchmesser der starren Körper gegen Null konvergiert, während die Anzahl starrer Körper gegen unendlich streben darf, beschränkt durch den Logarithmus des maximalen Durchmessers. Sowohl die Flüssigkeit als auch die starren Körper dürfen inhomogen sein, und die Viskosität der Flüssigkeit kann von ihrer Dichte abhängen. Die Anfangsbedingungen für die Dichte sollen beschränkt sein in \(L^{p_0}(\Omega)\) für ein \(d<p_0\leq\infty\) und in \(L^1(\Omega)\) konvergieren, die Anfangsbedingungen des Impulses konvergieren schwach in \(L^2(\Omega)^d\) und die Anfangsenergien konvergieren. Unter diesen Annahmen sind die starren Körper im asymptotischen Grenzwert vernachlässigbar, was heißt, dass, zumindest auf einer Teilfolge, Dichte und Geschwindigkeit für die Fluid-Struktur Wechselwirkung schwach gegen eine Lösung der inhomogenen inkompressiblen Navier-Stokes Gleichungen konvergieren. Wir zeigen dieses Resultat in jeder Dimension \(d\geq2\). Weiter wenden wir es auf den Spezialfall von zu Beginn periodisch verteilten starren Körpern gleicher Form an. Wir erhalten, dass die starren Körper im asymptotischen Grenzwert vernachlässigbar sind, falls ihre Größe im Vergleich zur Periode der Anfangsverteilung schnell genug klein wird.

The first contribution of this thesis is a new regularity theorem for time harmonic Maxwell's equations with less than Lipschitz complex anisotropic coefficients. By using the L^p theory for elliptic equations, it is possible to prove H¹ and Hölder regularity results, provided that the coefficients are W^1,p for some p = 3. This improves previous regularity results, where the assumption W^1,∞ for the coefficients was believed to be optimal. The method can be easily extended to the case of bi-anisotropic materials, for which a separate approach turns out to be unnecessary. The second focus of this work is the boundary control of the Helmholtz and Maxwell equations to enforce local constraints inside the domain. More precisely, we look for suitable boundary conditions such that the corresponding solutions and their derivatives satisfy certain local non-zero constraints. Complex geometric optics solutions can be used to construct such illuminations, but are impractical for several reasons. We propose a constructive approach to this problem based on the use of multiple frequencies. The suitable boundary conditions are explicitly constructed and give the desired constraints, provided that a finite number of frequencies, given a priori, are chosen in a fixed range. This method is based on the holomorphicity of the solutions with respect to the frequency and on the regularity theory for the PDE under consideration. This theory finds applications to several hybrid imaging inverse problems, where the unknown coefficients have to be imaged from internal measurements. In order to perform the reconstruction, we often need to find suitable boundary conditions such that the corresponding solutions satisfy certain non-zero constraints, depending on the particular problem under consideration. The multiple frequency approach introduced in this thesis represents a valid alternative to the use of complex geometric optics solutions to construct such boundary conditions. Several examples are discussed.

We investigate the simulations of the the Schrödinger equation using the onedimensional quantum lattice Boltzmann (QLB) scheme and the irregular behaviour of solution. We isolate error due to approximation of the Schrödinger solution with the non-relativistic limit of the Dirac equation and numerical error in solving the Dirac equation. Detailed analysis of the original scheme showed it to be first order accurate. By discretizing the Dirac equation consistently on both sides we derive a second order accurate QLB scheme with the same evolution algorithm as the original and requiring only a one-time unitary transformation of the initial conditions and final output. We show that initializing the scheme in a way that is consistent with the non-relativistic limit supresses the oscillations around the Schrödinger solution. However, we find the QLB scheme better suited to simulation of relativistic quantum systems governed by the Dirac equation and apply it to the Klein paradox. We reproduce the quantum tunnelling results of previous research and show second order convergence to the theoretical wave packet transmission probability. After identifying and correcting the error in the multidimensional extension of the original QLB scheme that produced asymmetric solutions, we expand our second order QLB scheme to multiple dimensions. Next we use the QLB scheme to simulate Klein tunnelling of massless charge carriers in graphene, compare with theoretical solutions and study the dependence of charge transmission on the incidence angle, wave packet and potential barrier shape. To do this we derive a representation of the Dirac-like equation governing charge carriers in graphene for the one-dimensional QLB scheme, and derive a two-dimensional second order graphene QLB scheme for more accurate simulation of wave packets. We demonstrate charge confinement in a graphene device using a configuration of multiple smooth potential barriers, thereby achieving a high ratio of on/off current with potential application in graphene field effect transistors for logic devices. To allow simulation in magnetic or pseudo-magnetic fields created by deformation of graphene, we expand the scheme to include vector potentials. In addition, we derive QLB schemes for bilayer graphene and the non-linear Dirac equation governing Bose-Einstein condensates in hexagonal optical lattices.