We consider the constrained convex minimization problem min x2C f(x) we will present gradient projection method which generates a sequence fxkg according to the formula xk+1 = PC(xk

In this thesis, high order stochastic Runge-Kutta methods are developed for the numerical solution of (Stratonvich) stochastic differential equations and numerical results are presented. The problems associated with non-communativity of stochastic differential equation systems are addressed and stochastic Runge-Kutta methods particularly suited for such systems are derived. The thesis concludes with a discussion on various implementation issues, along with numerical results from variable stepsize implementation of a stochastic embedded pair of Runge-Kutta methods.

The theory of solving polynomial systems by homotopy continuation method has been proposed by Garcia, Zangwill and Drexler, and the most typical method in this category is total degree homotpy. The numerical implementation of tracking homotopy curves can be taken as two parts: prediction and correction. In this thesis we compare the performance of several prediction methods in the total degree homotopy, including Runge-Kutta method, Adams-Bashforth method and cubic Hermite method. In addition, we design an adaptive stepsize control algorithm in path tracking, which is based on the information obtained during Newton correction process. The numerical experiment shows that the stepsize control algorithm is quite efficient and reliable in path tracking. In the end we employ the algorithm for solving eigenvalue problems by random product homotopy method

We consider the constrained convex minimization problem: min_x∈C f(x) we will present gradient projection method which generates a sequence x^k according to the formula x^(k+1) = P_c(x^k − £\_k∇f(x^k)), k= 0, 1, ¡P ¡P ¡P , our ideal is rewritten the formula as a xed point algorithm: x^(k+1) = T_(£\k)x^k, k = 0, 1, ¡P ¡P ¡P is used to solve the minimization problem. In this paper, we present the gradient projection method(GPM) and different choices of the stepsize to discuss the convergence of gradient projection method which converge to a solution of the concerned problem.

Modern numerical ordinary differential equation initial-value problem (ODE-IVP) solvers compute a piecewise polynomial approximate solution to the mathematical problem. Evaluating the mathematical problem at this approximate solution defines the defect. Corless and Corliss proposed rigorous defect control of numerical ODE-IVP. This thesis automates rigorous defect control for explicit, first-order, nonlinear ODE-IVP. Defect control is residual-based backward error analysis for ODE, a special case of Wilkinson's backward error analysis. This thesis describes a complete software implementation of the Corless and Corliss algorithm and extensive numerical studies. Basic time-stepping software is adapted to defect control and implemented. Advances in software developed for validated computing applications and advances in programming languages supporting operator overloading enable the computation of a tight rigorous enclosure of the defect evaluated at the approximate solution with Taylor models. Rigorously bounding a norm of the defect, the Corless and Corliss algorithm controls to mathematical certainty the norm of the defect to be less than a user specified tolerance over the integration interval. The validated computing software used in this thesis happens to compute a rigorous supremum norm. The defect of an approximate solution to the mathematical problem is associated with a new problem, the perturbed reference problem. This approximate solution is often the product of a numerical procedure. Nonetheless, it solves exactly the new problem including all errors. Defect control accepts the approximate solution whenever the sup-norm of the defect is less than a user specified tolerance. A user must be satisfied that the new problem is an acceptable model. / Thesis / Master of Science (MSc) / Many processes in our daily lives evolve in time, even the weather. Scientists want to predict the future makeup of the process. To do so they build models to model physical reality. Scientists design algorithms to solve these models, and the algorithm implemented in this project was designed over 25 years ago. Recent advances in mathematics and software enabled this algorithm to be implemented. Scientific software implements mathematical algorithms, and sometimes there is more than one software solution to apply to the model. The software tools developed in this project enable scientists to objectively compare solution techniques. There are two forces at play; models and software solutions. This project build software to automate the construction of the exact solution of a nearby model. That's cool.

Submitted by Pablo Aguiar De Maio (pabloamaio@outlook.com) on 2015-09-10T19:50:43Z No. of bitstreams: 1 Pablo Aguiar De Maio - Dissertação - Um método de linearização local com passo adaptativo para solução numérica de equações diferenciais estocásticas com ruído aditivo.pdf: 2233029 bytes, checksum: d3ed48936d09fde216e44fb4d688b47d (MD5) / Approved for entry into archive by Janete de Oliveira Feitosa (janete.feitosa@fgv.br) on 2015-09-25T12:16:10Z (GMT) No. of bitstreams: 1 Pablo Aguiar De Maio - Dissertação - Um método de linearização local com passo adaptativo para solução numérica de equações diferenciais estocásticas com ruído aditivo.pdf: 2233029 bytes, checksum: d3ed48936d09fde216e44fb4d688b47d (MD5) / Approved for entry into archive by Marcia Bacha (marcia.bacha@fgv.br) on 2015-09-28T16:51:14Z (GMT) No. of bitstreams: 1 Pablo Aguiar De Maio - Dissertação - Um método de linearização local com passo adaptativo para solução numérica de equações diferenciais estocásticas com ruído aditivo.pdf: 2233029 bytes, checksum: d3ed48936d09fde216e44fb4d688b47d (MD5) / Made available in DSpace on 2015-09-28T16:51:50Z (GMT). No. of bitstreams: 1 Pablo Aguiar De Maio - Dissertação - Um método de linearização local com passo adaptativo para solução numérica de equações diferenciais estocásticas com ruído aditivo.pdf: 2233029 bytes, checksum: d3ed48936d09fde216e44fb4d688b47d (MD5) Previous issue date: 2015-07-31 / In this work we present a new numerical method with adaptive stepsize based on the local linearization approach, to integrate stochastic differential equations with additive noise. We also propose a computational scheme that allows efficient implementation of this method, properly adapting the algorithm of Padé with scaling-squaring strategy to compute the exponential of matrices involved. To introduce the construction of this method, we briefly explain what stochastic differential equations are, the mathematics that is behind them, their relevance to the modeling of various phenomena, and the importance of using numerical methods to evaluate this kind of equations. A succinct study of numerical stability is also presented on the following pages. With this dissertation, we intend to introduce the necessary basis for the construction of the new method/scheme. At the end, several numerical experiments are performed to demonstrate, in a practical way, the effectiveness of the proposed method, comparing it with other methods commonly used. / Neste trabalho apresentamos um novo método numérico com passo adaptativo baseado na abordagem de linearização local, para a integração de equações diferenciais estocásticas com ruído aditivo. Propomos, também, um esquema computacional que permite a implementação eficiente deste método, adaptando adequadamente o algorítimo de Padé com a estratégia “scaling-squaring” para o cálculo das exponenciais de matrizes envolvidas. Antes de introduzirmos a construção deste método, apresentaremos de forma breve o que são equações diferenciais estocásticas, a matemática que as fundamenta, a sua relevância para a modelagem dos mais diversos fenômenos, e a importância da utilização de métodos numéricos para avaliar tais equações. Também é feito um breve estudo sobre estabilidade numérica. Com isto, pretendemos introduzir as bases necessárias para a construção do novo método/esquema. Ao final, vários experimentos numéricos são realizados para mostrar, de forma prática, a eficácia do método proposto, e compará-lo com outros métodos usualmente utilizados.

Les problèmes inverses consistent à retrouver une donnée qui a été transformée ou perturbée. Ils nécessitent une régularisation puisque mal posés. En traitement d'images, la variation totale en tant qu'outil de régularisation a l'avantage de préserver les discontinuités tout en créant des zones lisses, résultats établis dans cette thèse dans un cadre continu et pour des énergies générales. En outre, nous proposons et étudions une variante de la variation totale. Nous établissons une formulation duale qui nous permet de démontrer que cette variante coïncide avec la variation totale sur des ensembles de périmètre fini. Ces dernières années les méthodes non-locales exploitant les auto-similarités dans les images ont connu un succès particulier. Nous adaptons cette approche au problème de complétion de spectre pour des problèmes inverses généraux. La dernière partie est consacrée aux aspects algorithmiques inhérents à l'optimisation des énergies convexes considérées. Nous étudions la convergence et la complexité d'une famille récente d'algorithmes dits Primal-Dual.

Tato disertační práce se zabývá analýzou stability a konvergence klasických numerických metod pro řešení obyčejných diferenciálních rovnic. Jsou představeny klasické jednokrokové metody, jako je Eulerova metoda, Runge-Kuttovy metody a nepříliš známá, ale rychlá a přesná metoda Taylorovy řady. V práci uvažujeme zobecnění jednokrokových metod do vícekrokových metod, jako jsou Adamsovy metody, a jejich implementaci ve dvojicích prediktor-korektor. Dále uvádíme generalizaci do vícekrokových metod vyšších derivací, jako jsou např. Obreshkovovy metody. Dvojice prediktor-korektor jsou často implementovány v kombinacích modů, v práci uvažujeme tzv. módy PEC a PECE. Hlavním cílem a přínosem této práce je nová metoda čtvrtého řádu, která se skládá z dvoukrokového prediktoru a jednokrokového korektoru, jejichž formule využívají druhých derivací. V práci je diskutována Nordsieckova reprezentace, algoritmus pro výběr proměnlivého integračního kroku nebo odhad lokálních a globálních chyb. Navržený přístup je vhodně upraven pro použití proměnlivého integračního kroku s přístupe vyšších derivací. Uvádíme srovnání s klasickými metodami a provedené experimenty pro lineární a nelineární problémy.