Neste trabalho faz-se uma análise de fluxos unidimensionais usando as equações de Burgers e de Euler. Para esta última, é obtida a solução exata e uma aproximação desta via método numérico. A obtenção da solução exata é baseada na combinação de ondas simples (uma onda de choque, uma descontinuidade de contato e uma onda de expansão) e na validade das relações de salto para as equações de Euler. Os resultados assim obtidos são utilizados para verificar (certificar) os resultados numéricos. As equações de Euler são integradas no tempo através de um esquema simplificado de Runge-Kutta; considera-se também a adição explícita de termos dissipativos ao esquema de discretização espacial. Sã.o apresentadas comparações entre as soluções exata e numérica, além de comparações da solução numérica para diferentes valores dos coeficientes de dissipação. Analisa-se também as regiões de estabilidade de métodos de Runge-Kutta para uma equação modelo, cujas propriedades são semelhantes às das equações de Burgers e de Euler. Por fim , propõe-se o estudo da convergência de um esquema semelhante ao de Runge-Kutta; faz-se uma estimativa de erro a posteriori em espaços de Banach de dimensão infinita. Além disto, são calculadas algumas estimativas a priori para a equação de Burgers que são usadas, juntamente com idéias da teoria de "shadowing" para estabelecer estimativas relativas à validade de simulações numéricas para a equação de Burgers. Além disto, são mostrados alguns estudos computacionais relevantes sobre a separação espectral, os quais poderiam ser entendidos como uma forma de estimativas a posteriori. / In this work we analyze unidimensional flows using Burgers and Euler equations. We show how to obtain both an exact and an approximate solution to the Euler equations. The exact solution is obtained by simple waves interaction (a shock wave, a contact discontinuity and an expansion fan). It is based on the jump relations for the Euler equations. \Ve use the exact solution thus obtained as a mean of comparison to the approximate (numerical) solution. A simplified Runge-Kutta method is used to integrate the Euler equations in time. Dissipation terms are added to the spatial discretized equations. vVe study the effect of these terms by plotting the exact versus the approximate solution considering different dissipation coefficients. We choose a model equation, which has similar properties to the Burgers and Euler equations, in order to study the stability regions of the Runge-Kutta scheme. Finally, we propose a convergence analysis of a Runge-Kutta-like scheme and make a posteriori error estimates in infinite dimensional Banach spaces. Furthermore, we perform some a priori estimates for the Burgers' equation and use them together with ideas of the Shadowing Theory in order to establish estimates concerning the validity of numerical simulations for Burgers' equation. As well we show a few relevant computer studies on the spectral separation which could be regarded as a form of a posteriori estimates.

Seguindo as ideias de s. smale, estudamos a eficiencia probabilistica de algoritmos numericos para equacoes diferenciais ordinarias. especial atencao e dada a dois exemplos classicos: os algoritmos de runge-kutta de dois e de quatro estagios, sendo a sua eficiencia estimada em termos de medidas gaussianas. em ambos os casos, sao obtidas estimativas detalhadas que levam a uma expressao para a media do erro global. / Following the ideas of S. Smale, we study the probabilistic efficiency of numerical algorithms in ordinary differential equations. Special attention is directed to two classical examples: the algorithms of Runge-Kutta of two and four stages with their efficiency estimated in terms of gaussian measures. In both these cases detailed estimates are given. leading to an expression for the mean global error.

As aplicações da mecânica vibratória vêm crescendo significativamente na análise de sistemas de suspensões e estruturas de veículos, dentre outras. Desta forma, o presente trabalho desenvolve técnicas para a simulação e o controle de uma suspensão de automóvel utilizando modelos dinâmicos com um, dois e três graus de liberdade. Na obtenção das equações do movimento para o sistema massa-mola-amortecedor, o modelo matemático utilizado tem como base a equação de Lagrange e a segunda lei de Newton, com condições iniciais apropriadas. A solução numérica destas equações é obtida através do método de Runge-Kutta de 4ª ordem, utilizando o software MATLAB. Para controlar as vibrações do sistema utilizou-se três métodos diferentes de controle: clássico, LQR e alocação de pólos. O sistema assim obtido satisfaz as condições de estabilidade e de desempenho e é factível para aplicações práticas, pois os resultados obtidos comparam adequadamente com dados analíticos, numéricos ou experimentais encontrados na literatura, indicando que técnicas de controle como o clássico podem ser simples e eficientes.

La résolution numérique du système d'équations pleinement compressibles en vue de son utilisation pour des applications en Prévision Numérique du Temps (PNT) soulève de nombreuses questions. L'une d'elles porte sur le choix des schémas de discrétisation temporelle à mettre en oeuvre afin de résoudre ce système de la manière la plus efficace possible, pour permettre la continuelle amélioration qualitative des prévisions. Jusqu'alors, les schémas de discrétisation temporelle basés sur des techniques semi-implicites (SI) étaient les plus couramment employés PNT, compte tenu de leur robustesse et de leur grande propriété de stabilité. Mais avec l'émergence des machines massivement parallèles à mémoire distribuée, l'efficacité de ces techniques est actuellement remise en question, car leur confortable plage de stabilité est obtenue au prix de l'inversion d'un problème elliptique tri-dimensionnel très gourmand en communications. Ce travail thèse vise à explorer d'autres méthodes de discrétisation temporelle, en remplacemant des méthodes SI, s'appuyant sur des approches de type Horizontalement Explicite et Verticalement Implicite (HEVI). D'une part, ces approches s'affranchissent de la contrainte numérique imposée sur le pas de temps par la propagation verticales des ondes rapides supportées par le système, grâce au traitement implicite des processus verticaux. D'autre part, elles exploitent le paradigme de programmation voulant que chaque colonne verticale du modèle numérique soit traitée par un unique processeur. Ainsi, le traitement implicite de cette direction n'engendre aucunes communications entre les processeurs. Cependant, bien que ces ap- proches HEVI apparaissent comme une solution attractive, rien ne garanti que leurs efficacités puissent être aussi compétitives que celles des sché- mas SI. Pour ce faire, ces schémas HEVI doivent permettre l'utilisation de pas de temps raisonnables pour une application en PNT. L'objectif de ce travail de thèse est d'élaborer un schéma de discrétisation temporelle HEVI le plus efficace possible pour une utilisation en PNT, c'est à dire, un schéma qui autorise le plus long pas de temps possible. Dans cette optique, deux voies ont été explorées : la première, issue des méthodes à pas de temps fractionné, a permis de revisiter et d'améliorer un schéma de discrétisation temporelle déjà proposé mais dont l'examen n'a jamais été approfondi dans la littérature ; il s'agit du schéma d'avance temporelle saute-moutons trapézo\"idal. Il a été mis en évidence que l'ajout d'un simple filtre temporel d'usage commun en PNT, améliore grandement la stabilité de ce schéma, lui permettant ainsi à moindre coût de rivaliser en terme de stabilité avec le schéma Runge-Kutta explicite d'ordre 3. La seconde voie, plus récente, c'est avérée la plus prometteuse. Elle repose sur l'utilisation des méthodes Runge-Kutta Implicite-Explicite (RK-IMEX) HEVI. Au cours l'étude, il a été tout d'abord mis en évidence certains problèmes de stabi- lité des schémas initialement suggérés dans la littérature en présence des processus d'advection. Puis, une nouvelle classe de schéma RK-IMEX HEVI s'appuyant sur un traitement temporel spécifique des termes d'ajustement horizontaux a été proposé / The use non-hydrostatic fully compressible modelling system in the perspective of Numerical Weather Prediction (NWP) raises many challenging questions, among which the choice time discretization scheme. It is commonly acknowledge that the ideal time marching algorithms to integrate the fully compressible system should both overcome the stability constraint imposed on time-step by the fast propagating waves supported by the system, and be scalable enough for efficiently computing on massively parallel computer machine. The assumed poor scalability property of Semi-implicit (SI) time schemes, currently favoured in NWP, is quite a drawback as they require global communications to solve a full three-dimensional elliptic problem. Because it is considered as the best compromise between stability, accuracy and scalability the properties of various classes of Horizontally Explicit Vertically Implicit (HEVI) schemes have been deeply explore in this work in a view of solving the fully system in mass-based coordinate. This class of time discretization approach eliminates all the problems linked to the implicit treatment of horizontal high-frequency forcings by coupling multi-step or multi-stage explicit methods for the horizontal propagation of fast waves to an implicit scheme for the treatment of vertically prop- agating elastic disturbances. The limitation in time-steps compared to SI schemes would be compensated by a much more economical algorithm per time-step. However, it is not firmly established that the efficiency of such a HEVI schemes could compete with one of the semi-implicit schemes. The main objective of this Phd thesis work is to elaborate an efficient HEVI time scheme allowing usable time-step for NWP applications. For this purpose, the so-called explicit time-splitting technique and the recently suggested Runge-Kutta IMEX (RK-IMEX) schemes have been explored un- der HEVI approach. Firstly, the superiority in term of stability of the RK-IMEX methods in respect with the time-splitting approach has been con- firmed. However, in presence of advection processes some unstable numerical behaviour of these schemes has been pointed out. To circumvent this problem a new class of RK-IMEX HEVI schemes has been proposed. This new class of HEVI time schemes reveals to be very attractive since they provide both good stability and accuracy properties. Secondly, in a side aspect of the HEVI approach, the stability impact of the temporal treatment of the terrain following coordinate non-linear metric terms has been demonstrated. Numerical analyses on simplified framework indicate that there might be a benefit to deal with these specific terms in the implicit part of the HEVI schemes. All the theoretical studies have been confirmed by nu- merical testing through the use of a Cartesian vertical plane fully compressible model cast in a mass-based coordinate.

In this thesis, we compute approximate solutions to initial value problems of first-order linear ODEs using five explicit Runge-Kutta methods, namely the forward Euler method, Heun's method, RK4, RK5, and RK8. This thesis aims to compare the accuracy, stability, and efficiency properties of the five explicit Runge-Kutta methods. For accuracy, we carry out a convergence study to verify the convergence rate of the five explicit Runge-Kutta methods for solving a first-order linear ODE. For stability, we analyze the stability of the five explicit Runge-Kutta methods for solving a linear test equation. For efficiency, we carry out an efficiency study to compare the efficiency of the five explicit Runge-Kutta methods for solving a system of first-order linear ODEs, which is the main focus of this thesis. This system of first-order linear ODEs is a semi-discretization of a two-dimensional wave equation.

In many applications, ordinary differential equations can be additively partitioned \[y'=f(y)=\sum_{m=1}^{N}\f{}{m}(y).] It can be advantageous to discriminate between the different parts of the right-hand side according to stiffness, nonlinearity, evaluation cost, etc. In 2015, Sandu and G\"{u}nther \cite{sandu2015gark} introduced Generalized Additive Runge-Kutta (GARK) methods which are given by \begin{eqnarray*} Y_{i}^{\{q\}} & = & y_{n}+h\sum_{m=1}^{N}\sum_{j=1}^{s^{\{m\}}}a_{i,j}^{\{q,m\}}f^{\{m\}}\left(Y_{j}^{\{m\}}\right)\\ & & \text{for } i=1,\dots,s^{\{q\}},\,q=1,\dots,N\\ y_{n+1} & = & y_{n}+h\sum_{m=1}^{N}\sum_{j=1}^{s^{\{m\}}}b_{j}^{\{m\}}f^{\{m\}}\left(Y_{j}^{\{m\}}\right)\end{eqnarray*} with the corresponding generalized Butcher tableau \[\begin{array}{c|ccc} \c{}{1} & \A{1,1} & \cdots & \A{1,N}\\\vdots & \vdots & \ddots & \vdots\\ \c{}{N} & \A{N,1} & \cdots & \A{N,N}\\\hline & \b{}{1} & \cdots & \b{}{N}\end{array}\] The diagonal blocks $\left(\A{q,q},\b{}{q},\c{}{q}\right)$ can be chosen for example from standard Runge-Kutta methods, and the off-diagonal blocks $\A{q,m},\:q\neq m,$ act as coupling coefficients between the underlying methods. The case when $N=2$ and both diagonal blocks are implicit methods (IMIM) is examined. This thesis presents order conditions and simplifying assumptions that can be used to choose the off-diagonal coupling blocks for IMIM methods. Error analysis is performed for stiff problems of the form \begin{eqnarray*}\dot{y} & = & f(y,z)\\ \epsilon\dot{z} & = & g(y,z)\end{eqnarray*} with small stiffness parameter $\epsilon.$ As $\epsilon\to 0,$ the problem reduces to an index 1 differential algebraic equation provided $g_{z}(y,z)$ is invertible in a neighborhood of the solution. A tree theory is developed for IMIM methods applied to the reduced problem. Numerical results will be presented for several IMIM methods applied to the Van der Pol equation.

A tailored integration scheme is developed to treat stiff reaction-diffusion prob- lems. The construction adapts a stiff solver, namely VODE, to treat reaction im- plicitly together with explicit treatment of diffusion. The second-order Runge-Kutta- Chebyshev (RKC) scheme is adjusted to integrate diffusion. Spatial operator is de- scretised by second-order finite differences on a uniform grid. The overall solution is advanced over S fractional stiff integrations, where S corresponds to the number of RKC stages. The behavior of the scheme is analyzed by applying it to three simple problems. The results show that it achieves second-order accuracy, thus, preserving the formal accuracy of the original RKC. The presented development sets the stage for future extensions, particularly, to multidimensional reacting flows with detailed chemistry.

The purpose of this thesis is to solve unsteady two-dimensional compressible Navier-Stokes equations for a moving mesh using implicit explicit (IMEX) Runge- Kutta scheme. The moving mesh is implemented in the equations using Arbitrary Lagrangian Eulerian (ALE) formulation. The inviscid part of the equation is explicitly solved using second-order Godunov method, whereas the viscous part is calculated implicitly. We simulate subsonic compressible flow over static NACA-0012 airfoil at different angle of attacks. Finally, the moving mesh is examined via oscillating the airfoil between angle of attack = 0 and = 20 harmonically. It is observed that the numerical solution matches the experimental and numerical results in the literature to within 20%.

The purpose of this paper is solving the European option pricing problem under the Black–Scholes model. Our approach is to use the so-called stochastic Runge–Kutta (SRK) numericalscheme to find the corresponding expectation of the functional to the stochastic differentialequation under the Black–Scholes model. Several numerical solutions were made to study howquickly the result converges to the theoretical value. Then, we study the order of convergenceof the SRK method with the help of MATLAB.

In this thesis, our research focus on a weak second order stochastic Runge–Kutta method applied to a system of stochastic differential equations known as the Gatheral Model. We approximate numerical solutions to this system and investigate the rate of convergence of our method. Both call and put options are priced using Monte-Carlo simulation to investigate the order of convergence. The numerical results show that our method is consistent with the theoretical order of convergence of the Monte-Carlo simulation. However, in terms of the Runge-Kutta method, we cannot accept the consistency of our method with the theoretical order of convergence without further research.