|
|
The Lanczos method in finite precision arithmetic
Šimonová, Dorota ; Tichý, Petr (advisor) ; Hnětynková, Iveta (referee)
In this thesis we consider the Lanczos algoritm and its behaviour in finite precision. Having summarized theoretical properties of the algorithm and its connection to orthogonal polynomials, we recall the idea of the Lanczos method for approximating the matrix eigenvalues. As the behaviour of the algorithm is strongly influenced by finite precision arithmetic, the linear independence of the Lanczos vectors is usually lost after a few iterations. We use the most im- portant results from analysis of the finite precision Lanczos algorithm according to Paige, Greenbaum, Strakos and others. Based on that, we study formulation and properties of the mathematical model of finite presicion Lanczos computati- ons suggested by Greenbaum. We carry out numerical experiments in Matlab, which support the theoretical results.
|
|
|
Line search in descent methods
Moravová, Adéla ; Tichý, Petr (advisor) ; Vlasák, Miloslav (referee)
In this thesis, we deal with descent methods for functional minimalization. We discuss three conditions for the choice of the step length (Armijo, Goldstein, and Wolfe condition) and four descent methods (The steepest descent method, Newton's method, Quasi-Newton's method BFGS and the conjugate gradient method). We discuss their convergence properties and their advantages and dis- advantages. Finally, we test these methods numerically in the GNU Octave pro- gramming system on three different functions with different number of variables. 1
|
|
|
Incomplete Cholesky factorization
Hoang, Phuong Thao ; Tůma, Miroslav (advisor) ; Tichý, Petr (referee)
The thesis is about the incomplete Cholesky factorization and its va- riants, which are important for preconditioning a system with symmetric and positive definite matrix. Our main focus is on solving these systems, which arise in many technical applications and natural sciences, using preconditioned Con- jugate Gradients. Besides many other ways we can apply Cholesky factorization approximately, incompletely. In this thesis we study existence of the incomplete Cholesky factorization and we evaluate behaviour and potential of different vari- ants of the generic algorithm. 1
|
|
|
The choice of the stopping criteria for Newton-like methods
Kurnas, Jakub ; Dolejší, Vít (advisor) ; Tichý, Petr (referee)
We formulate examples of partial differential equations which can be solved through their discretization and subsequent solution of derived algebraic system. A brief summary of Discontinuous Galerkin Discretization is given as well as definitions of algebraic and discretization errors. We derive the Newton method, which solves nonlinear algebraic systems by solving a sequence of linear problems, we modify the method and examine implementation options. We define stopping criteria for the Newton-like method using aforementioned errors and we explain how to keep accuracy of the solution of derived algebraic system and the original partial differential equation in balance. We present numerical experiments to illustrate theoretical background and mention several basic properties of the Newton- like method.
|
|
|
The choice of the step in trust region methods
Rapavý, Martin ; Tichý, Petr (advisor) ; Kučera, Václav (referee)
The main goal of this thesis is the choice of steps in trust region methods for finding a minimum of a given function. The step corresponds to the problem of finding a minimum of a model function on a trust region. We characterize a solu- tion of this problem (Moré-Sorensen theorem) and consider various techniques for approximating a solution of this problem (the Cauchy point method, the dogleg method, the conjugate gradients method). In the case of the first two techniques we prove convergence of the optimization method. Finally, the above techniques are tested numerically in MATLAB on properly chosen functions and initial data. We comment on advantages and disadvantages of considered algorithms. 1
|
|
|
Field of values of a matrix: Theory and computation
Vacek, Lukáš ; Tichý, Petr (advisor) ; Tůma, Miroslav (referee)
The field of values of a matrix A is a convex set in the complex plane assigned to A. It is important in matrix analysis, especially in invetigation of properties of nonnormal matrices and matrix polynomials, in study of the con- vergence of iterative methods applied to these matrices, in the estimation of ma- trix function norms, etc. This thesis summarizes theory about the field of values of a matrix, formulates open problems and explaines the main idea of the basic numerical method for its computation. In numerical experiments the standart algorithmic realization of method is compared with alternative approaches that use power method, Lanczos algorithm and Chebfun.
|
|
|
Numerical computation with functions using Chebfun
Lébl, Matěj ; Tichý, Petr (advisor) ; Hnětynková, Iveta (referee)
Goal of this work is to introduce Chebfun software and show ideas behind it. In the first chapter we summarize the theory of polynomial interpolation with focus on the Chebyshev interpolants. In the second chapter we introduce Chebfun software, its basic commands and principles of constructing interpolants. The third chapter is devoted to demonstrate theorems from the first chapter and to show practical applications of Chebfun when finding roots of a function and solving differential equations. Powered by TCPDF (www.tcpdf.org)
|
|
|
Methods for the solution of nonlinear equations
Havelková, Eva ; Kučera, Václav (advisor) ; Tichý, Petr (referee)
The aim of this bachelor thesis is to present an overview of elementary numerical methods for solving nonlinear algebraic equations in one variable. Firstly, related concepts from numerical mathematics and mathematical analysis are explained. The main part of the thesis provides a detailed description of chosen iterative methods as well as the proofs of their orders of convergence. The methods covered are namely the bisection method, fixed-point iteration, regula falsi method, Newton's method, secant method and methods that are based on quadratic interpolation. The practical part of the thesis presents results of numerical experiments that were carried out with Matlab software on various types of nonlinear equations. These results are compared with the theory introduced in the preceding parts. The contribution of this thesis is to provide a comprehensive overview and comparison of the characteristics of basic methods for solving nonlinear equations based on a variety of literature. Powered by TCPDF (www.tcpdf.org)
|
|
| |
|
|
Multilevel methods and adaptivity
Vacek, Petr ; Strakoš, Zdeněk (advisor) ; Tichý, Petr (referee)
After introduction of the model problem we derive its weak formulation, show the existence and the uniqueness of the solution, and present the Galerkin finite element method. Then we briefly describe some of the stationary iterative methods and their smoothing property. We present the most common multigrid schemes, i.e. two-grid correction scheme, V-cycle scheme, and the full multigrid algorithm. Then we perform numerical experiment showing the differences between the use of the direct and iterative coarsest grid solver in V-cycle scheme and experiment considering a perturbation of the correction vector simulating a fault of a computational device. Powered by TCPDF (www.tcpdf.org)
|
guest ::
