We are interested in a nonlinear filtering problem motivated by an information-based approach for modelling the dynamic evolution of a portfolio of credit risky securities. We solve this problem by `change of measure method\\\' and show the existence of the density of the unnormalized conditional distribution which is a solution to the Zakai equation. Zakai equation is a linear SPDE which, in general, cannot be solved analytically. We apply Galerkin method to solve it numerically and show the convergence of Galerkin approximation in mean square. Lastly, we design an adaptive Galerkin filter with a basis of Hermite polynomials and we present numerical examples to illustrate the effectiveness of the proposed method. The work is closely related to the paper Frey and Schmidt (2010).

We are interested in a nonlinear filtering problem motivated by an information-based approach for modelling the dynamic evolution of a portfolio of credit risky securities. We solve this problem by `change of measure method\\\'' and show the existence of the density of the unnormalized conditional distribution which is a solution to the Zakai equation. Zakai equation is a linear SPDE which, in general, cannot be solved analytically. We apply Galerkin method to solve it numerically and show the convergence of Galerkin approximation in mean square. Lastly, we design an adaptive Galerkin filter with a basis of Hermite polynomials and we present numerical examples to illustrate the effectiveness of the proposed method. The work is closely related to the paper Frey and Schmidt (2010).

No description available.

Dans cette thèse, la Méthode des Points Matériels (MPM) est étendue à l’approximation de Galerkin Discontinue (DG) et appliquée aux problèmes hyperboliques en mécanique des solides. La méthode résultante (DGMPM) a pour objectif de suivre précisément les ondes dans des solides subissant de fortes déformations et dont les modèles constitutifs dépendent de l’histoire du chargement. A la croisée des méthodes de types éléments finis et volumes finis, la DGMPM s’appuie sur une grille de calcul arbitraire dans laquelle des flux sont calculés au moyen de solveurs de Riemann approximés sur les arêtes entre les éléments. L’intérêt de ce type de solveurs est qu’ils permettent l’introduction de la structure caractéristique des solutions des équations aux dérivées partielles hyperboliques directement dans le schéma numérique. Les analyses de stabilité et de convergence ainsi que l’illustration de la méthode sur des simulations de problèmes unidimensionnels et bidimensionnels montrent que le schéma numérique permet d’améliorer le suivi des ondes par rapport à la MPM. Par ailleurs, un deuxième objectif poursuivi dans cette thèse consiste à caractériser la réponse des solides élastoplastiques à des sollicitations dynamiques en deux dimensions en vue d’améliorer la résolution numérique de ces problèmes. Bien qu’un certain nombre de travaux aient déjà été menés dans cette direction, les problèmes étudiés se limitent à des cas particuliers. Un cadre unifié pour l’étude de la propagation d’ondes simples dans les solides élastoplastiques en déformations et contraintes plane est proposé dans cette thèse. Les trajets de chargement suivis à l’intérieur de ces ondes simples sont de plus analysés. / In this thesis, the material point method (MPM) is extended to the discontinuous Galerkin approximation (DG) and applied to hyperbolic problems in solid mechanics. The resulting method (DGMPM) aims at accurately following waves in finite-deforming solids whose constitutive models may depend on the loading history. Merging finite volumes and finite elements methods, the DGMPM takes advantage of an arbitrary computational grid in which fluxes are evaluated at element faces by means of approximate Riemann solvers. This class of solvers enables the introduction of the characteristic structure of the solutions of hyperbolic partial differential equations within the numerical scheme. Convergence and stability analyses, along with one and two-dimensional numerical simulations,demonstrate that this approach enhances the MPM ability to track waves. On the other hand, a second purpose has been followed: it consists in identifying the response of two-dimensional elastoplastic solids to dynamic step-loadings in order to improve numerical results on these problems. Although some studies investigated similar questions, only particular cases have been treated. Thus,a generic framework for the study of the propagation of simple waves in elastic-plastic solids under plane stress and plane strain problems is proposed in this thesis. The loading paths followed inside those simple waves are further analyzed.

O objetivo desta tese é investigar aspectos práticos que facilitem a aplicação da teoria de controle H1 não linear em projetos de sistemas de controle. A primeira contribuição deste trabalho é a proposta do uso de funções ponderação com dinâmica no projeto de controladores H1 não lineares. Essas funções são usadas no projeto de controladores H1 lineares para rejeição de perturbações, ruídos, atenuação de erro de rastreamento, dentre outras especificações. O maior obstáculo para aplicação prática da teoria de controle H1 não linear é a dificuldade para resolver simultaneamente as duas equações de Hamilton-Jacobi-Isaacs relacionadas ao problema de realimentação de estados e injeção da saída. Não há métodos sistematicos para resolver essas duas equações diferenciais parciais não lineares, equivalentes µas equações de Riccati da teoria de controle H1 linear. A segunda contribuição desta tese é um método para obter a injeção da saída transformando a equação de Hamilton-Jacobi-Isaacs em uma sequencia de equações diferenciais parciais lineares, que são resolvidas usando o método de Galerkin. Controladores H1 não lineares para um sistema de levitação magnética são obtidos usando o método clássico de expansão em série de Taylor e o método de proposto para comparação. / The purpose of this thesis is to investigate practical aspects to facilitate the ap- plication of nonlinear H1 theory in control systems design. Firstly, it is shown that dynamic weighting functions can be used to improve the performance and robustness of the nonlinear H1 controller such as in the design of H1 controllers for linear plants. The biggest bottleneck to the practical applications of nonlinear H1 control theory has been the di±culty in solving the Hamilton-Jacobi-Isaacs equations associated with the design of a state feedback and an output injection gain. There is no systematic numerical approach for solving this ¯rst order, nonlinear partial di®erential equations, which reduces to Riccati equations in the linear context. In this work, successive ap- proximation and Galerkin approximation methods are combined to derive an algorithm that produces an output injection gain. Design of nonlinear H1 controllers obtained by the well established Taylor approximation and by the proposed Galerkin approxi- mation method applied to a magnetic levitation system are presented for comparison purposes.

We study a two-scale coupled system consisting of a macroscopic elliptic equation and a microscopic parabolic equation. This system models the interplay between a gas and liquid close to equilibrium within a porous medium with distributed microstructures. We use formal homogenization arguments to derive the target system. We start by proving well-posedness and inverse estimates for the two-scale system. We follow up by proposing a Galerkin scheme which is continuous in time and discrete in space, for which we obtain well-posedness, a priori error estimates and convergence rates. Finally, we propose a numerical error reduction strategy by refining the grid based on residual error estimators.

O objetivo desta tese é investigar aspectos práticos que facilitem a aplicação da teoria de controle H1 não linear em projetos de sistemas de controle. A primeira contribuição deste trabalho é a proposta do uso de funções ponderação com dinâmica no projeto de controladores H1 não lineares. Essas funções são usadas no projeto de controladores H1 lineares para rejeição de perturbações, ruídos, atenuação de erro de rastreamento, dentre outras especificações. O maior obstáculo para aplicação prática da teoria de controle H1 não linear é a dificuldade para resolver simultaneamente as duas equações de Hamilton-Jacobi-Isaacs relacionadas ao problema de realimentação de estados e injeção da saída. Não há métodos sistematicos para resolver essas duas equações diferenciais parciais não lineares, equivalentes µas equações de Riccati da teoria de controle H1 linear. A segunda contribuição desta tese é um método para obter a injeção da saída transformando a equação de Hamilton-Jacobi-Isaacs em uma sequencia de equações diferenciais parciais lineares, que são resolvidas usando o método de Galerkin. Controladores H1 não lineares para um sistema de levitação magnética são obtidos usando o método clássico de expansão em série de Taylor e o método de proposto para comparação. / The purpose of this thesis is to investigate practical aspects to facilitate the ap- plication of nonlinear H1 theory in control systems design. Firstly, it is shown that dynamic weighting functions can be used to improve the performance and robustness of the nonlinear H1 controller such as in the design of H1 controllers for linear plants. The biggest bottleneck to the practical applications of nonlinear H1 control theory has been the di±culty in solving the Hamilton-Jacobi-Isaacs equations associated with the design of a state feedback and an output injection gain. There is no systematic numerical approach for solving this ¯rst order, nonlinear partial di®erential equations, which reduces to Riccati equations in the linear context. In this work, successive ap- proximation and Galerkin approximation methods are combined to derive an algorithm that produces an output injection gain. Design of nonlinear H1 controllers obtained by the well established Taylor approximation and by the proposed Galerkin approxi- mation method applied to a magnetic levitation system are presented for comparison purposes.