Alguns problemas elípticos assintoticamente lineares são considerados e é provada a existência de solução. Os principais resultados são estabelecidos de dois modos distintos e as provas são baseadas em resultados clássicos da teoria de pontos críticos, a saber: minimização, princípio variacional de Ekeland, grau topológico, teorema do ponto de sela e o teorema do passo da montanha / Some asymptotically linear elliptic problems are considered and solutions are proved to exist. The main results are proved in two different ways. The proofs rely on some classical results in Critical Point Theory such as minimization, Ekeland variational principle, topological degree, saddle point theorem and mountain pass theorem

<p>In recent years, there has been a great interest towards the higher order numerical methods for singularly perturbed problems. As compared to their lower order counterparts, they provide better accuracy with fewer mesh points. Construction and/or implementation of direct higher order methods is usually very complicated. Thus a natural choice is to use some convergence acceleration techniques, e.g., Richardson extrapolation, defect correction, etc. In this thesis, we will consider various classes of problems described by singularly perturbed ordinary and partial differential equations. For these problems, we design some novel numerical methods and attempt to increase their accuracy as well as the order of convergence. We also do the same for existing numerical methods in some instances. We &macr / nd that, even though the Richardson extrapolation technique always improves the accuracy, it does not perform equally well when applied to different methods for certain classes of problems. Moreover, while in some cases it improves the order of convergence, in other cases it does not. These issues are discussed in this thesis for linear and nonlinear singularly perturbed ODEs as well as PDEs. Extrapolation techniques are analyzed thoroughly in all the cases, whereas the limitations of the defect correction approach for certain problems is indicated at the end of the thesis</p>

<p>In recent years, there has been a great interest towards the higher order numerical methods for singularly perturbed problems. As compared to their lower order counterparts, they provide better accuracy with fewer mesh points. Construction and/or implementation of direct higher order methods is usually very complicated. Thus a natural choice is to use some convergence acceleration techniques, e.g., Richardson extrapolation, defect correction, etc. In this thesis, we will consider various classes of problems described by singularly perturbed ordinary and partial differential equations. For these problems, we design some novel numerical methods and attempt to increase their accuracy as well as the order of convergence. We also do the same for existing numerical methods in some instances. We &macr / nd that, even though the Richardson extrapolation technique always improves the accuracy, it does not perform equally well when applied to different methods for certain classes of problems. Moreover, while in some cases it improves the order of convergence, in other cases it does not. These issues are discussed in this thesis for linear and nonlinear singularly perturbed ODEs as well as PDEs. Extrapolation techniques are analyzed thoroughly in all the cases, whereas the limitations of the defect correction approach for certain problems is indicated at the end of the thesis</p>

Submitted by Izabel Franco (izabel-franco@ufscar.br) on 2016-09-23T20:10:37Z No. of bitstreams: 1 TeseWSR.pdf: 754474 bytes, checksum: f94a1d7e509de51f9a78d4e27289b7aa (MD5) / Approved for entry into archive by Marina Freitas (marinapf@ufscar.br) on 2016-09-26T20:46:07Z (GMT) No. of bitstreams: 1 TeseWSR.pdf: 754474 bytes, checksum: f94a1d7e509de51f9a78d4e27289b7aa (MD5) / Approved for entry into archive by Marina Freitas (marinapf@ufscar.br) on 2016-09-26T20:46:12Z (GMT) No. of bitstreams: 1 TeseWSR.pdf: 754474 bytes, checksum: f94a1d7e509de51f9a78d4e27289b7aa (MD5) / Made available in DSpace on 2016-09-26T20:46:19Z (GMT). No. of bitstreams: 1 TeseWSR.pdf: 754474 bytes, checksum: f94a1d7e509de51f9a78d4e27289b7aa (MD5) Previous issue date: 2015-03-06 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / The aim of this work is to present results of existence of solutions for a class of nonlinear asymmetryc elliptic problems. The asymmetry that we consider here has linear behavior on - infnity and superlinear on + infnity. To obtain these results we apply variational methods as linking theorems and topological methods like topological degree theory. / Neste trabalho apresentamos resultados de existência de soluções para uma classe de problemas elípticos não lineares assimétricos. A assimetria que consideramos aqui tem comportamento linear em - infinito e superlinear em + infinito. Para obter tais resultados aplicamos métodos variacionais como teoremas de linking e métodos topológicos como a teoria do grau topológico.

Made available in DSpace on 2015-05-15T11:46:09Z (GMT). No. of bitstreams: 1 arquivototal.pdf: 652579 bytes, checksum: a75e379418567b71eed4fc021cfeeae6 (MD5) Previous issue date: 2011-08-25 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / We start studing semi-stable solutions for the equation u = f(u) in a smooth and bounded domain of Rn, 2 n 4. The presented result is a L1 boundedness, which holds for all semi-stable positive solution and all non-linearity f. We also show a approach about the case u = f(u) in the unitary ball of Rn. The results obtained are Lq and Wk;q estimates for semi-stable radial solutions u 2 H1 0 , the proof of a boundedness if n 9 and, in case that g is increasing and convex, u 2 W3;2 in all dimension n. / Começamos estudamos soluções semi-estáveis para a equação u = f(u) em um domínio suave limitado do Rn, 2 n 4. O resultado apresentado é uma limitação L1 a qual vale para toda solução positiva semi-estável e toda nãolinearidade f. Mostramos também uma abordagem sobre o caso u = f(u) na bola unitária do Rn. Os resultados obtidos são estimativas em Lq e Wk;q para soluções semi-estáveis radiais u 2 H1 0 , a prova de uma limitação se n 9 e, no caso em que g é crescente e convexa, u 2 W3;2 em toda dimensão n.

Submitted by Erika Demachki (erikademachki@gmail.com) on 2017-03-27T17:45:29Z No. of bitstreams: 2 Dissertação - Gustavo Silvestre do Amaral Costa - 2017.pdf: 671324 bytes, checksum: fdf29c0b102f3ee24a198d5616ecd4b4 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2017-03-28T11:31:51Z (GMT) No. of bitstreams: 2 Dissertação - Gustavo Silvestre do Amaral Costa - 2017.pdf: 671324 bytes, checksum: fdf29c0b102f3ee24a198d5616ecd4b4 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2017-03-28T11:31:51Z (GMT). No. of bitstreams: 2 Dissertação - Gustavo Silvestre do Amaral Costa - 2017.pdf: 671324 bytes, checksum: fdf29c0b102f3ee24a198d5616ecd4b4 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2017-03-17 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / In this work we will discuss the existence of nonnegative solutions for a class of indefinite semilinear elliptic problems: (Pμ)   − u = λ1u+μg(x,u)+W(x)f(u), em u = 0 , sobre ∂ , where is a bounded smooth domain in RN, N ≥ 3, μ is a nonnegative parameter, λ1 is the first eigenvalue of the operator − under Dirichlet boundary conditions, W ∈ C(¯ ,R) is a weight function, f ∈ C(R,R), and g : ¯ ×R→R is a Carathéodory locally bounded function, i.e, for every s0 > 0, there is M := M(s0) > 0 such that |g(x,s)| ≤M for 0 ≤ |s| ≤ s0 and for almost every x ∈ ¯ . / Neste trabalho discutiremos a existência de soluções não negativas para os seguintes problemas semilineares elípticos indefinidos: (Pμ)   − u = λ1u+μg(x,u)+W(x)f(u), em u = 0 , sobre ∂ . onde é um domínio limitado suave de RN, N ≥ 3, λ1 é o primeiro autovalor de − , μ > 0, W ∈ C(¯ ,R) e f ∈ C(R,R), g : ×R→R é uma função Carathéodory localmente limitada, isto é, para todo s0 > 0 existe M(s0) > 0, tal que |g(x,s)| ≤ M(s0), para todo s ∈ [−s0,s0] e q.t.p em ¯ .

Alguns problemas elípticos assintoticamente lineares são considerados e é provada a existência de solução. Os principais resultados são estabelecidos de dois modos distintos e as provas são baseadas em resultados clássicos da teoria de pontos críticos, a saber: minimização, princípio variacional de Ekeland, grau topológico, teorema do ponto de sela e o teorema do passo da montanha / Some asymptotically linear elliptic problems are considered and solutions are proved to exist. The main results are proved in two different ways. The proofs rely on some classical results in Critical Point Theory such as minimization, Ekeland variational principle, topological degree, saddle point theorem and mountain pass theorem

Philosophiae Doctor - PhD / In recent years, there has been a great interest towards the higher order numerical methods for singularly perturbed problems. As compared to their lower order counterparts, they provide better accuracy with fewer mesh points. Construction and/or implementation of direct higher order methods is usually very complicated. Thus a natural choice is to use some convergence acceleration techniques, e.g., Richardson extrapolation, defect correction, etc. In this thesis, we will consider various classes of problems described by singularly perturbed ordinary and partial differential equations. For these problems, we design some novel numerical methods and attempt to increase their accuracy as well as the order of convergence. We also do the same for existing numerical methods in some instances. We find that, even though the Richardson extrapolation technique always improves the accuracy, it does not perform equally well when applied to different methods for certain classes of problems. Moreover, while in some cases it improves the order of convergence, in other cases it does not. These issues are discussed in this thesis for linear and nonlinear singularly perturbed ODEs as well as PDEs. Extrapolation techniques are analyzed thoroughly in all the cases, whereas the limitations of the defect correction approach for certain problems is indicated at the end of the thesis. / South Africa

La méthode des éléments finis inversés est une méthode sans troncature qui a été introduite pour résoudre des équations aux dérivées partielles en domaines non bornés. L’objective de cette thèse est d’analyser, d’adapter puis d’implémenter cette méthode pour résoudre quelques problèmes issus de la physique, notamment lorsque le domaine géométrique est l’espace R3 tout entier. Dans un premier temps, nous présentons de manière détaillée les aspects et les principes fondamentaux de la méthode. Ensuite, nous adapterons la méthode à des problèmes de type div-rot et de potentiels vecteurs posés dans R3. Après avoir analysé la convergence de la méthode, on montrera quelques résultats numériques obtenus avec un code tridimensionnel. On s’intéresse ensuite au problème de calcul de l’énergie magnétostatique dans des problèmes de micromagnétisme, où on développe avec succès une approche numérique utilisant les éléments finis inversés. Dans la dernière partie, on adapte la méthode à un problème provenant de la chimie quantique (modèle de continuum polarisable) pour lequel on prouve qu’elle donne des résultats numériques très prometteurs. La thèse comporte beaucoup de résultats numériques issus de codes tridimensionnels écrits ou co-écrits, notamment lorsque le domaine est l’espace tout entier. Elle comporte aussi des résultats théoriques liés à l’utilisation des espaces de Sobolev à poids comme cadre fonctionnel. On apporte en particulier une preuve constructive de quelques inégalités de type div-rot dans des domaines non bornés. / Inverted finite element method (IFEM) is a non runcature method which was introduced for solving partial differential equations in unbounded domains. The objective of this thesis is to analyze, to adapt and to implement IFEM for solving several problems arising in physics, especially when the domain is the whole space R3. We first give a presentation in which we detail the principles and the main features of the method. Then, we adapt IFEM for solving some div-curl systems and vector potential problems in the whole space. In a second part, we successfully develop an IFEM based approach for computing the stray-field energy in micromagnetism. In the last part, we are interested in the study of the polarizable continuum model arising in quantum chemistry. The manuscript contains a large number of numerical results obtained with some 3D codes, especially when the domain is the whole space R3. It also contains some theoretical results in relation with weighted Sobolev spaces. We give in particular a constructive proof of some div-curl inequalities in unbounded domains.

Neste trabalho, mostramos a existência e multiplicidade de soluções para a seguinte classe de equações elípticas quasilineares { - \'DELTA IND. \'NÜ\' POT. \'upsilon\' + \'|\'upsilon\'| POT. \'NÜ\' - 2 \'upsilon\' = f(x, u), \'upsilon\' \'DIFERENTE\' 0, \'upsilon\' \'PERTENCE A >>: Nu + jujN2 u = f(x; u); x 2 ; u 6= 0; u 2 W1;N( ); onde e um domnio em RN, N 2, N e o operador N-Laplaciano e f e uma func~ao que possui um crescimento crtico exponencial. Para obter nossos resultados utilizamos o Princpio Variacional de Ekeland, Teorema do Passo da Montanha, Categoria de Lusternik- Schnirelman, Ac~ao de Grupo e tecnicas baseadas na Teoria do G^enero. Palavras chaves: Problemas elpticos quasilineares, Metodo Variacional, N-Laplaciano, crescimento crtico exponencial, Princpio Variacional de Ekeland, Categoria de Lusternik- Schnirelman, Desigualdade de Trudinger-Moser / In this work, we show the existence and multiplicity of solutions for the following class of quasilinear elliptic equations { - \'DELTA\' IND. \'NÜ\' \'upsilon\'\' + |\'upsilon\'| POT. \'NÜ\' - 2 = f(x, \'upsilon\'), x \"IT BELONGS\' \'OMEGA\', \'upsilon\' \'DIFFERENT\' 0, \'upsilon\' \'IT BELONGS\' W POT. 1, \'NÜ\' ( OMEGA), where \'OMEGA\' is a domain in \' R POT. \'NÜ\' > OR = 2, \'DELTA\' IND. \'NÜ\' is the N-Laplacian operator and f is a function with exponential critical growth. To obtain our results we utilize the Ekeland Variational Principle, the Mountain Pass Theorem, Lusternik-Schnirelman of Category, Group Action and techniques based on Genus Theory