Dans une première partie, on s'intéresse à un système d'EDP stochastiques variant selon deux échelles de temps, et plus particulièrement à l'approximation de la composante lente à l'aide d'un schéma numérique efficace. On commence par montrer un principe de moyennisation, à savoir la convergence de la composante lente du système vers la solution d'une équation dite moyennée. Ensuite on prouve qu'un schéma numérique de type Euler fournit une bonne approximation d'un coefficient inconnu apparaissant dans cette équation moyennée. Finalement, on construit et on analyse un schéma de discrétisation du système à partir des résultats précédents, selon la méthodologie dite HMM (Heterogeneous Multiscale Method). On met en évidence l'ordre de convergence par rapport au paramètre d'échelle temporelle et aux différents paramètres du schéma numérique- on étudie les convergences au sens fort (approximation des trajectoires) et au sens faible (approximation des lois). Dans une seconde partie, on étudie une méthode d'approximation de solutions d'EDP paraboliques, en combinant une approche semi-lagrangienne et une discrétisation de type Monte-Carlo. On montre d'abord dans un cas simplifié que la variance dépend des pas de discrétisation- enfin on fournit des simulations numériques de solutions, afin de mettre en avant les applications possibles d'une telle méthode. / In a first part, we are interested in the behavior of a system of Stochastic PDEs with two time-scales- more precisely, we focus on the approximation of the slow component thanks to an efficient numerical scheme. We first prove an averaging principle, which states that the slow component converges to the solution of the so-called averaged equation. We then show that a numerical scheme of Euler type provides a good approximation of an unknown coefficient appearing in the averaged equation. Finally, we build and we analyze a discretization scheme based on the previous results, according to the HMM methodology (Heterogeneous Multiscale Method). We precise the orders of convergence with respect to the time-scale parameter and to the parameters of the numerical discretization- we study the convergence in a strong sense - approximation of the trajectories - and in a weak sense - approximation of the laws. In a second part, we study a method for approximating solutions of parabolic PDEs, which combines a semi-lagrangian approach and a Monte-Carlo discretization. We first show in a simplified situation that the variance depends on the discretization steps. We then provide numerical simulations of solutions, in order to show some possible applications of such a method.

Cette thèse traite de la solution numérique de deux types de problèmes stochastiques. Premièrement, nous nous intéressons aux EDS fortement oscillantes, c'est-à-dire, les systèmes composés de variables ergodiques évoluant rapidement par rapport aux autres. Nous proposons un algorithme basé sur des résultats d'homogénéisation. Il est défini par un schéma d'Euler appliqué aux variables lentes couplé avec un estimateur à pas décroissant pour approcher la limite ergodique des variables rapides. Nous prouvons la convergence forte de l'algorithme et montrons que son erreur normalisée satisfait un résultat du type théorème limite centrale généralisé. Nous proposons également une version extrapolée de l'algorithme ayant une meilleure complexité asymptotique en satisfaisant les mêmes propriétés que la version originale. Ensuite, nous étudions la solution des EDS de type McKean-Vlasov (EDSPR-MKV) associées à la solution de certains problèmes de contrôle sous un environnement formé d'un grand nombre de particules ayant des interactions du type champ-moyen. D'abord, nous présentons un nouvel algorithme, basé sur la méthode de cubature sur l'espace de Wiener, pour approcher faiblement la solution d'une EDS du type McKean-Vlasov. Il est déterministe et peut être paramétré pour atteindre tout ordre de convergence souhaité. Puis, en utilisant ce nouvel algorithme, nous construisons deux schémas pour résoudre les EDSPR-MKV découplées et nous montrons que ces schémas ont des convergences d'ordres un et deux. Enfin, nous considérons le problème de réduction de la complexité de la méthode présentée tout en respectant la vitesse de convergence énoncée. / This Ph.D. thesis deals with the numerical solution of two types of stochastic problems. First, we investigate the numeric solution to strongly oscillating SDEs, i.e. systems in which some ergodic state variables evolve quickly with respect to the remaining ones. We propose an algorithm that uses homogenization results and consists of an Euler scheme for the slow scale variables coupled with a decreasing step estimator for the ergodic averages of the fast variables. We prove the strong convergence of the algorithm as well as a generalized central limit theorem result for the normalized error distribution. In addition, we propose an extrapolated version applicable under stronger regularity assumptions and which satisfies the same properties of the original algorithm with lower asymptotic complexity. Then, we treat the problem of solving decoupled Forward Backward Stochastic Differential equations of McKean-Vlasov type (MKV-FBSDE) which appear in some stochastic control problems in an environment of a large number of particles with mean field interactions. As a first step, we propose a new algorithm, based on the cubature method on Wiener spaces, to weakly approach the solution of a McKean-Vlasov SDE. It is deterministic and can be parametrized to obtain any given order of convergence. Using this first forward approximation algorithm, we construct two procedures to solve the decoupled MKV-FBSDE and show that they converge with orders one and two under appropriate regularity conditions. Finally, we consider the problem of reducing the complexity of the presented method while preserving the presented convergence rates.

[pt] O desenvolvimento consistente do método convencional dos elementos de contorno (CBEM), com a adição de conceitos da versão simplificada do método híbrido dos elementos de contorno (HBEM), proveniente do potencial variacional de Hellinger-Reissner, conduz-se a um processo computacionalmente mais econômico, sem a necessidade de ter sua precisão numérica reduzida para problemas de grande escala, podendo ser bidimensional ou tridimensional, de potencial ou elasticidade. Conseguiu-se mostrar que as matrizes de potencial duplo e simples do CBEM, H e G, respectivamente, cuja avaliação numérica requer a manipulação de integrais singulares e impróprias, podem ser obtidas de maneira expedita, eliminando-se quase toda a integração numérica, com exceção de algumas integrais regulares. Uma importante característica da formulação proposta, que advém da base variacional do HBEM, é a facilidade da obtenção de resultados em pontos internos, de maneira direta e sem a utilização de qualquer integral de contorno, já que a solução fundamental é a própria solução do problema. O presente trabalho pertence a um projeto cujo resultado final deve ser um código computacional para problemas de grande escala (milhões de graus de liberdade). Nesta fase, alguns exemplos numéricos foram testados para avaliar a aplicabilidade do método expedito, o seu esforço computacional e a convergência do resultado para as variáveis envolvidas no método. Para isso, foram implementados algoritmos para problemas bidimensionais de potencial e elasticidade - usando elementos lineares, quadráticos e cúbicos - e tridimensionais - usando elementos triangulares e quadrilaterais, lineares e quadráticos nos dois casos. Os códigos computacionais foram implementados focando na solução de problemas de grande escala. Espera-se que numa etapa final o projeto possa ser bem mais eficaz, com a incorporação de procedimentos do método fast multipole. / [en] The consistent development of the conventional boundary elements method (CBEM) by adding the concepts of the hybrid boundary element simplified method (HBEM) , from the Hellinger-Reissner variational potential leads to a computationally less intensive procedure, although not necessarily less accurate for large scale, two-dimensional or three-dimensional problems of potential and elasticity. It was shown that both single-layer and double-layer potential matrices, G and H, respectively, are obtained in an expeditious way that vanish almost any numerical integration, except for a few regular integrals, even G and H evaluation requires the handling of singular and improper integrals. The proposed formulation comes from the HBEM variational base and its evaluation at internal points is straightforward without the application of any boundary integral, since the fundamental solution is the analytical one. This work belongs to a project that aims a computer code for large-scale problems (millions of degrees of freedom). At this stage, some numerical examples were analyzed to evaluate the applicability of the method expeditious its computational effort and convergence of the results for the variables involved in the method. It was developed by the algorithms implementation for potential and elasticity problems. In the case of two-dimensional were employed linear, quadratic and cubic elements and to the three-dimensional case were employed triangular, quadrilateral, linear and quadratic elements in both cases. The computational codes were always implemented focused on solving largescale problems. It is expected that in a final stage of the project with the incorporation procedure of the method fast multipole, it can be more efficiently.

Dans cette thèse, nous étudions la diffraction électromagnétique par des réseau et surfaces rugueuse aléatoire. Le méthode C est une méthode exacte développée pour ce but. Il est basé sur équations de Maxwell sous forme covariante écrite dans un système de coordonnées non orthogonal. Le méthode C conduisent à résoudre le problème de valeur propre. Le champ diffusé est expansé comme une combinaison linéaire des solutions propres satisfaisant à la condition d’onde sortant.Nous nous concentrons sur l’aspect numérique de la méthode C, en essayant de développer une application efficace de cette méthode exacte. Pour les réseaux, nous proposons une nouvelle version de la méthode C qui conduit `a un système différentiel avec les conditions initiales. Nous montrons que cette nouvelle version de la méthode C peut être utilisée pour étudier les réseaux de multicouches avec un médium homogène.Nous vous proposons un algorithme QR parallèle conçu spécifiquement pour la méthode C pour résoudre le problème de valeurs propres. Cet algorithme QR parallèle est une variante de l’algorithme QR sur la base de trois tech- niques: “décalage rapide”, poursuite de renflement parallèle et de dégonflage parallèle agressif précoce (AED). / We study the electromagnetic diffraction by gratings and random rough surfaces. The C-method is an exact method developed for this aim. It is based on Maxwell’s equations under covariant form written in a nonorthogonal coordinate system. The C-method leads to an eigenvalue problem, the solution of which gives the diffracted field.We focus on the numerical aspect of the C-method, trying to develop an efficient application of this exact method. For gratings, we have developed a new version of C-method which leads to a differential system with initial conditions. This new version of C-method can be used to study multilayer gratings with homogeneous medium.We implemented high performance algorithms to the original versions of C-method. Especially, we have developed a specifically designed parallel QR algorithm for the C- method and spectral projection method to solve the eigenvalue problem more efficiently. Experiments have shown that the computation time can be reduced significantly.

O estudo de métodos de modelagem de escoamentos aerodinâmicos em regime transônico é de grande importância para a engenharia aeronáutica. O maior desafio no tratamento desses escoamentos está na sua característica não linear devido aos efeitos de compressibilidade e formação de ondas de choque. Tais efeitos não lineares influenciam no desempenho de superfícies aerodinâmicas em geral, bem como são responsáveis pelo aparecimento de fenômenos danosos para a resposta aeroelástica de aeronaves. O equacionamento para esses tipos de escoamentos pode ser obtido via as equações básicas da mecânica dos fluidos. No entanto, apenas soluções numéricas de tais equações são possíveis de ser obtidas de forma prática no presente momento. Para o caso específico do tratamento de problemas transônicos, as equações de Euler formam um conjunto de equações diferenciais a derivadas parciais capazes de capturar os efeitos não lineares de escoamentos compressíveis, porém os efeitos da viscosidade não são levados em consideração. O objetivo desse trabalho é implementar uma rotina computacional capaz de resolver numericamente escoamentos em regime transônico em torno de aerofólios. Para isso as equações de Euler não lineares são utilizadas e o campo de fluido ao redor de um perfil aerodinâmico é discretizado pelo método das diferenças finitas. Uma malha estruturada do tipo C discretizando o fluido ao redor de um aerofólio NACA0012 é considerada. A metodologia para solução numérica é baseada no método explícito de MacCormack de segunda ordem de precisão no tempo e espaço. Baseados na aproximação upwind, termos de dissipação artificial com coeficientes não lineares também são adicionados ao método. A solução do escoamento transônico estacionário ao redor do aerofólio NACA0012 é obtida e as principais propriedades do escoamento são apresentadas. Observa-se a formação de ondas de choque através de contornos de número de Mach ao redor do aerofólio. Gráficos das distribuições de pressão no intra e extradorso do aerofólio são mostrados, onde se identificam aos efeitos da brusca variação de pressão devido as ondas de choque. Os resultados são validados com valores de distribuição de pressão para o mesmo aerofólio encontradas na literatura técnica. Os resultados obtidos combinam bem com os fornecidos em códigos computacionais para solução do mesmo problema aerodinâmico / The study of aerodynamic modeling methods for the transonic flow regime is of great importance in aeronautical engineering. Major challenge on the treatment of those flows is on their nonlinear features due to compressibility effects and shock waves (appearance). Such nonlinear effects present a strong influence on aerodynamic performance, as well as they are responsible for harmful aeroelastic response phenomena in aircraft. Equations for transonic flows can be obtained from the basic fluid mechanic equations. However, only numerical methods are able to attain practical solutions for those set of differential equations at the present moment. For the specific case of treating transonic flow problems, the nonlinear Euler equations provide a set of partial differential equations with features to capture nonlinear effects of typical compressible flows, despite of not accounting for viscous flows effects. The aim of this work is to implement a computational routine for the numerical solution of transonic flows around airfoils. The Euler equations are used and the flow field around a aerodynamic profile is discretized by finite difference method. A C-type structured mesh is used to discretize the flow around a NACA0012 airfoil. The methodology for numerical solution is based on the explicit MacCormack method which has second order accuracy in time and space. Based on the upwind approximation, artificial dissipation with nonlinear coefficients is incorporated to the method. The steady transonic flow around the NACA0012 airfoil numerical solution is assessed and the main flow properties are presented. Shock wave structure can also be observed by means of the Mach number contours around the airfoil. Pressure distributions on upper and lower surfaces for different flow conditions are also shown, thereby allowing the observation of the effects of the abrupt pressure change due to shock waves. The results are validated using data presented in the technical literature. The present code solutions agree well with the solution obtained in other computational codes used for the same problem

Este trabalho apresenta uma nova metodologia para elastografia virtual em imagens simuladas de ultrassom utilizando métodos numéricos e métodos de visão computacional. O objetivo é estimar o módulo de elasticidade de diferentes tecidos tendo como entrada duas imagens da mesma seção transversal obtidas em instantes de tempo e pressões aplicadas diferentes. Esta metodologia consiste em calcular um campo de deslocamento das imagens com um método de fluxo óptico e aplicar um método iterativo para estimar os módulos de elasticidade (análise inversa) utilizando métodos numéricos. Para o cálculo dos deslocamentos, duas formulações são utilizadas para fluxo óptico: Lucas-Kanade e Brox. A análise inversa é realizada utilizando duas técnicas numéricas distintas: o Método dos Elementos Finitos (MEF) e o Método dos Elementos de Contorno (MEC), sendo ambos implementados em Unidades de Processamento Gráfico de uso geral, GpGPUs ( \"General Purpose Graphics Units\" ). Considerando uma quantidade qualquer de materiais a serem determinados, para a implementação do Método dos Elementos de Contorno é empregada a técnica de sub-regiões para acoplar as matrizes de diferentes estruturas identificadas na imagem. O processo de otimização utilizado para determinar as constantes elásticas é realizado de forma semi-analítica utilizando cálculo por variáveis complexas. A metodologia é testada em três etapas distintas, com simulações sem ruído, simulações com adição de ruído branco gaussiano e phantoms matemáticos utilizando rastreamento de ruído speckle. Os resultados das simulações apontam o uso do MEF como mais preciso, porém computacionalmente mais caro, enquanto o MEC apresenta erros toleráveis e maior velocidade no tempo de processamento. / This thesis presents a new methodology for computational elastography applied to simulated ultrasound images, using numerical methods and comptuter vision methods. The aim is to estimate the elastic moduli of diferent tissues using two diferent images of the same cross section acquired in diferent times and pressure conditions. The proposed methodology consists in evaluate the displacement field using optical flow techniques and then apply an inverse analysis using a numerical method. In order to evaluate the displacement field, two distinct formulations for optical flow are used: Lucas-Kanade and Brox. For the inverse analysis problem, the Finite Element Method and the Boundary Element Method are used, both implemented in general purpose graphic units, GpGPUs. Considering a number of materials that may be present in the images, the multiresgions boundary element method is used in order to couple diferent matrices for diferent materials. The optimization process is evaluated using complex variable method. The methodology is validated in three diferent steps: noiseless simulations; additive white gaussian noise simulations; and ultrasound mathematical phantom with speckle tracking. The results show that the Finite Element Method presents more accurate estimatives but a high computational cost, while the Boundary Element Method presents tolerable errors but a better processing time.

Este trabalho apresenta uma contribuição ao emprego do Método dos Elementos Finitos Generalizados (MEFG) na análise tridimensional não-linear de sólidos. A análise numérica em campo não-linear, com modelos de dano e plasticidade, é original. O MEFG é uma formulação não-convencional do Método dos Elementos Finitos (MEF), que resulta da incorporação a este último de conceitos e técnicas dos denominados métodos sem malha, especialmente o enriquecimento da aproximação inicial (partição de unidade) por funções convenientes. Apresenta-se uma breve revisão bibliográfica dos métodos sem malha e do método dos elementos finitos generalizados, bem como suas principais características. Apresenta-se, com base no MEFG, a formulação de elementos tetraédricos e hexaédricos. Três modelos constitutivos são considerados visando análises não-lineares: o de plasticidade (perfeita ou com encruamento isótropo linear) com critério de plastificação de von Mises; o de dano frágil em concreto sob carregamento monótono crescente (modelo de Mazars) e o de dano e plasticidade acoplados (modelo de Lemaitre), próprio para materiais metálicos. São apresentados detalhes do código computacional, baseado no MEFG e nos modelos constitutivos acima mencionados, bem como resultados de análises numéricas. Esses resultados ressaltam algumas das vantagens do MEFG aplicado à análise não-linear, tais como: o enriquecimento da aproximação inicial limitado a regiões de interesse no domínio, como por exemplo, as que exibem elevados gradientes de deformação e tensão; uma definição mais precisa da distribuição de grandezas como a variável de dano e a tensão equivalente de von Mises, evitando a necessidade de alterações na malha; e a superação do travamento volumétrico associado a modelos de plasticidade / This work presents a contribution to the generalized finite element method (GFEM) employment in three-dimensional nonlinear analysis of solids. The nonlinear numerical analysis conduced with damage and plasticity models is original. GFEM is a nonconventional formulation of finite element method (FEM) which results from the addition to the latter of concepts and techniques of the so called Meshless methods, specially the enrichment of the initial approximations (partition of unity) by customized functions. A brief review of Meshless methods and generalized finite element method bibliography is presented, as well as their main features. Based on GFEM, the formulation of tetrahedral and hexahedral elements is shown. Three material laws are considered aiming nonlinear analysis: plasticity (perfectly plastic or linear isotropic hardening), with von Mises yield criterion; brittle damage on concrete under monotonic increasing loading (Mazars model) and damage coupled with plasticity (Lemaitre model), a suitable model for metals. Details of the computational code, based on GFEM and material laws mentioned above, are presented, as well as results of numerical analysis. These results emphasize some of the advantages of GFEM applied to nonlinear analysis, such as: enrichment of the basic approximations limited to some regions of interest in the domain, for instance, those exhibiting high strain and stress gradients; an accurated definition of the distributions of quantities like damage variable and von Mises equivalent stress, avoiding remeshing; and overcoming of volumetric locking associated to plasticity models

Bei der Modellierung elektrischer Schaltungen ergeben sich Algebro-Differentialgleichungen (DAEs) mit proper formuliertem Hauptterm. Diese Gleichungen müssen z.B. bei der transienten Schaltungssimulation numerisch gelöst werden. Bei den klassischen Ansätzen der Linearen Mehrschrittverfahren oder der Runge-Kutta Verfahren ergeben sich Nachteile, die durch Verwendung von Allgemeinen Linearen Verfahren vermieden werden können. Sowohl Lineare Mehrschrittverfahren als auch Runge-Kutta Verfahren sind als Spezialfälle in dieser allgemeineren Klasse enthalten. Darüberhinaus sind aber neue Verfahren mit verbesserten Eigenschaften möglich. In dieser Arbeit werden DAEs der Schaltungssimulation eingehend studiert und Allgemeine Lineare Verfahren für solche Gleichungen untersucht. Die Verfahrenskonstruktion und Implementierungsfragen werden ausführlich diskutiert. Diese Arbeit erscheint im Logos Verlag Berlin (www.logos-verlag.de, ISBN 3-8325-1353-1). / Modelling electrical circuits leads to differential algebraic equations (DAEs) having a properly stated leading term. These equations need to be solved numerically, e.g. in case of a transient analysis of the given circuit. Classical methods such as linear multistep methods or Runge-Kutta schemes suffer from disadvantages that can be overcome by studying general linear schemes. Both Runge-Kutta methods and linear multistep schemes belong to this class as special cases, but there is plenty of room for new methods with improved properties. This work presents both a detailed study of DAEs in the framework of integrated circuit design and a thorough analysis of general linear methods for these kind of equations. The construction and implementation of general linear methods for DAEs is discussed in detail. This work is published by Logos Verlag Berlin (www.logos-verlag.de, ISBN 3-8325-1353-1).

Diese Arbeit stellt eine gründliche Analyse der Schwerpunktsbewegung von Exzitonen in Halbleiter-Quantengräben und -Quantendrähten dar. Dabei wurde die k.p-Kopplung der schweren und leichten Löcher im Valenzband sowie das Coulomb-Potential voll berücksichtigt. Die Optimierung der Schwerpunktstransformation auf der Basis eines Ansatzes für die Abhängigkeit des Grundzustands des Exzitons vom Schwerpunktsimpuls Q ermöglichte numerische Ergebnisse hoher Qualität. Es zeigt sich nämlich, daß in einer Subbandentwicklung die Enveloppe des Grundzustands des Exzitons in guter Näherung unabhängig vom Schwerpunktsimpuls ist. So konnten erstmalig Multiband-Exziton-Berechnungen in Quantendrähten mit voller Berücksichtigung der Coulomb-Wechselwirkung durchgeführt werden. Die in dieser Arbeit dargestellten Untersuchungen zeigen interessante physikalische Effekte auf, wie beispielsweise eine nichtmonotone Zunahme der Bindungsenergie des exzitonischen Grundzustands mit wachsendem Q und einen zur entsprechenden Kontinuumskante weitestgehend parallelen Verlauf der Dispersion des exzitonischen Grundzustands. Die Optimierung der Schwerpunktstransformation führt außerdem zu einem analytischen Ausdruck für eine mittlere Masse, die relevant für den exzitonischen Grundzustand ist. / This thesis presents a thorough investigation of the center-of-mass dispersion properties of excitons in semiconductor quantum wells and quantum wires. The k.p coupling of heavy and light holes as well as the Coulomb coupling are taken fully into account. High-quality numerical calculations of the exciton center-of-mass dispersion are achieved by optimizing the center-of-mass transformation, making use of an Ansatz for the dependence of the groundstate exciton upon the center-of-mass momentum Q. Indeed, the envelope in the subband expansion of the groundstate exciton is to a good approximation independent of Q. This technique made possible for the first time multiband calculations in quantum wires that take the Coulomb coupling fully into account. Various physically interesting effects are found and investigated, like, e.g., the non-monotonous increase of the exciton groundstate binding energy with Q or the fact that the exciton groundstate energy follows the exciton continuum edge rather closely. The center-of-mass optimization leads also to an analytical expression for an estimate of the exciton groundstate center-of-mass mass.

Ziel dieser Dissertation ist die Untersuchung der asymptotischen Stabilität numerischer Verfahren für Index-2-Algebro-Differentialgleichungen. Es werden Anfangswertaufgaben für quasilineare Algebro-Differentialgleichungen (ADGln). Die meisten anwendungsrelevanten Aufgaben können damit behandelt werden. Zuerst werden einige Stabilitätsbegrife und Aussagen vorgestellt, die das Fundament für den Rest der Arbeit darstellen. Dies erstreckt sich sowohl auf den kontinuierlichen als auch auf den diskreten Fall. Insbesondere werden Kontraktivitätskonzepte eingeführt und Beziehungen zwischen der Kontraktivität der ADGl und derer der Anwendung eines numerischen Verfahrens. Die eingeführte Kontraktivitätsbegriffe erweitern oder verallgemeinern die bereits bekannten Konzepte. Als wichtigste Aussage in dem Kontraktivitätskontext geht ein Theorem hervor, das allgemeine Bedingungen aufstellt, damit die Anwendung eines IRK(DAE)-Verfahrens auf eine ADGl stabil ist. Bekannte Aussagen für gewöhnliche und Algebro-Differntialgleichungen können als Sonderfälle dieses Ergebnisses gesehen werden. Im weiteren Verlauf der Arbeit wird anhand von neuartigen Index-2-Entkopplungs- und Indexreduktionstechniken die Stabilität von Diskretisierungsverfahren untersucht. Die durchgeführte Analyse erbringt neue Ergebnisse, die eine Verbesserung des Kenntnissstandes in diesem Gebiet darstellen. Die erzielte Aussagen stellen hinreichende Bedingungen, damit ein BDF- oder IRK-Verfahren für eine ADGl das gleiche Stabilitätsverhalten wie für eine gewöhnliche Differentialgleichung besitzt. Diese Ergebnisse werden durch numerishce Beispiele veranschaulicht. Weiterhin stellt man fest, dass eine der gefundenen Voraussetzungen für die Kontraktivität der Anwendung eines algebraisch stabilen IRK(DAE)-Verfahrens, auf eine ebenfalls kontraktive ADGl, genügt. Dieses Ergebnis wurde durch die Anwendung der im ersten Teil dieser Arbeit erzielten Kontraktivitätsaussagen ermöglicht. Die Konsequenzen der soeben genannten Aussage für bestimmte Modelle der Schaltkreissimulation werden ebenfalls erläutert. Aus der oben genannten Analyse, ebenso wie aus der Fachliteratur, geht hervor, dass bei manchen ADGl-Aufgaben die Diskretisierungsverfahren Stabilitätsprobleme aufweisen. Um solche Probleme zu behandeln sind bereits einige Ansätze bekannt. Im letzten Teil der Arbeit werden zwei repräsentativen Ansätze betrachtet und ihre Aussichtschancen für Index-2-Aufgaben anhand eines kritischen Beispieles evaluiert. Des Weiteren wird eine Verallgemeinerung für vollimplizite lineare ADGln des Gear-Gupta-Leimkuhler-Ansatzes (GGL) vorgeschlagen. Der Rest der Arbeit beschäftigt sich mit der Stabilitätsuntersuchung der GGL-Formulierung und der auf sie angewandten numerischen Verfahren. Dafür werden Aussagen dieser Arbeit eingesetzt und man kommt zu der Schlussfolgerung, dass sowohl für die IRK(DAE)- als auch für die BDF-Verfahren die Integration der GGL-Formulierung, natürlich unter bestimmten Voraussetzungen, stabil ist. Dieses Ergebniss wird durch ein numerisches Beispiel belegt. Dabei handelt es um eine Gleichung, die mit einer direkten Anwendung eines Verfahrens Instabilitäten aufweist. Jedoch ist die Integration der entsprechenden GGL stabil. / The purpose of the present PhD work is the asymptotic stability investigation of numerical methods for index 2 differential algebraic equations. Initial value problems are considered for quasi linear differential algebraic equations (DAEs) that cover the most important applications. First some stability concepts and related results are presented, which represent the basis for further investigations. This background concerns both, the continuous and the discreet case. Especially contractivity concepts are introduced and the relationship between the asymptotic stability of the DAE and the numerical method applied to it is established. The new contractivity concepts extend or generalize the already known concepts. The most important result in this context is a theorem that establishes general conditions under which the application of an algebraic stable IRK(DAE) method to a DAE is contractive. Well-known assertions for ordinary and differential algebraic equations can be considered as special cases of this general result. Later on the stability of numerical discretizations applied to index-2 DAEs is investigated. This is made possible by the introduction of new decopling and index reduction techniques. The analysis makes new insights in the asymptotic of numerical methods for DAEs possible. The obtained results state sufficient conditions in order that a BDF or an IRK(DAE) method applying to DAEs shows the same asymptotic stability properties as for ODEs. These results are illustrated by some numerical examples. Moreover, it can be realized that one of the found conditions is sufficient in order to show contractivity of the application of an algebraic stable IRK(DAE) method, supposed the DAE is contractive. This assertion is possible based on the general theorem mentioned in the paragraph above. Further some consequences of the mentioned results for electric network models are shown. According to both, the above mentioned analysis and the specialized literature of this field, the application of numerical methods to some special DAEs shows asymptotic stability problems. A few approaches are known to manage such difficult equations. Two exponents of these techniques are considered and their chances of success for index-2 DAEs are evaluated with the application to a critical example. A generalization of the Gear-Gupta-Leimkuhler (GGL) approach is proposed for full implicit linear DAEs. This generalization is investigated in detail in the rest of the paper, concerning both the analytical and the numerical asymptotic stability of the GGL equation and the numerical methods applied to it correspondingly. The result is, that, if some conditions are fulfilled, IRK(DAE) and BDF methods for the GGL equation will produce stable solutions. This result is illustrated by a numerical example. The application of the methods directly to the considered DAE produces unstable solutions. However, the integration of the corresponding GGL formulation is stable. The obtained result opens new possibility for the numerical treatment of instabilities by differential algebraic equations.