Dans le présent travail, des nouvelles méthodes de décomposition de domaine et des nouvelles implémentations pour des méthodes existantes sont développées. Une nouvelle méthode basée sur les méthodes antérieures de décomposition du domaine est formulée. Les méthodes classiques FETI plus FETI-2LM sont utilisées pour construire le nouveau Hybrid-FETI. L'idée de base est de développer un nouvel algorithme qui peut utiliser les deux méthodes en même temps en choisissant dans chaque interface l'état le plus adapté en fonction des caractéristiques du problème. En faisant cela, nous recherchons un code plus rapide et plus robuste qui peut fonctionner avec des configurations selon lesquelles les méthodes de base ne le géreront pas de manière optimale par lui-même. La performance est testée sur un problème de contact. La partie suivante implique le développement d'une nouvelle implémentation pour la méthode S-FETI, l'idée est de réduire l'utilisation de la mémoire de cette méthode, afin de pouvoir fonctionner dans des problèmes de taille plus important. Différentes variantes pour cette méthode sont également proposées, tout en cherchant la réduction des directions stockées chaque itération de la méthode itérative. Finalement, une extension de la méthode FETI-2LM à sa version en bloc comme dans S-FETI, est développée. Les résultats numériques pour les différents algorithmes sont présentés. / In this work new domain decomposition methods and new implementations for existing methods are developed. A new method based on previous domain decomposition methods is formulated. The classic FETI plus FETI-2LM methods are used to build the new Hybrid-FETI. The basic idea is to develop a new algorithm that can use both methods at the same time by choosing in each interface the most suited condition depending on the characteristics of the problem. By doing this we search to have a faster and more robust code that can work with configurations that the base methods will not handle it optimally by himself. The performance is tested on a contact problem. The following part involves the development of a new implementation for the S-FETI method, the idea is to reduce the memory usage of this method, to make it able to work in larger problem. Different variation for this method are also proposed, all searching the reduction of directions stored each iteration of the iterative method. Finally, an extension of the FETI-2LM method to his block version as in S-FETI, is developed. Numerical results for the different algorithms are presented.

Esta tesis aborda la paralelización de los métodos de Krylov con reinicio para problemas de valores propios y valores singulares (SVD). Estos métodos son de naturaleza iterativa y resultan adecuados para encontrar unos pocos valores propios o singulares de problemas dispersos. El procedimiento de ortogonalización suele ser la parte más costosa de este tipo de métodos, por lo que ha recibido especial atención en esta tesis, proponiendo y validando nuevos algoritmos para mejorar sus prestaciones paralelas. La implementación se ha realizado en el marco de la librería SLEPc, que proporciona una interfaz orientada a objetos para la resolución iterativa de problemas de valores propios o singulares. SLEPc está basada en la librería PETSc, que dispone de implementaciones paralelas de métodos iterativos para la resolución de sistemas lineales, precondicionadores, matrices dispersas y vectores. Ambas librerías están optimizadas para su ejecución en máquinas paralelas de memoria distribuida y con problemas dispersos de gran dimensión. Esta implementación incorpora los métodos para valores propios de Arnoldi con reinicio explícito, de Lanczos (incluyendo variantes semiortogonales) con reinicio explícito, y versiones de Krylov-Schur (equivalente al reinicio implícito) para problemas no Hermitianos y Hermitianos (Lanczos con reinicio grueso). Estos métodos comparten una interfaz común, permitiendo su comparación de forma sencilla, característica que no está disponible en otras implementaciones. Las mismas técnicas utilizadas para problemas de valores propios se han adaptado a los métodos de Golub-Kahan-Lanczos con reinicio explícito y grueso para problemas de valores singulares, de los que no existe ninguna otra implementación paralela con paso de mensajes. Cada uno de los métodos se ha validado mediante una batería de pruebas con matrices procedentes de aplicaciones reales. Las prestaciones paralelas se han medido en máquinas tipo cluster, comprobando una buena escalabilidad inc / Tomás Domínguez, A. (2009). Implementación paralela de métodos de Krylov con reinicio para problemas de valores propios y singulares [Tesis doctoral no publicada]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/5082 / Palancia

The SHIRA method of Mehrmann and Watkins belongs among the structure preserving Krylov subspace methods for solving skew-Hamiltonian eigenvalue problems. It can also be applied to Hamiltonian eigenproblems by considering a suitable transformation. Structure induced shift-and-invert techniques are employed to steer the algorithm towards the interesting region of the spectrum. However, the shift cannot be altered in the middle of the computation without discarding the information that has been accumulated so far. This paper shows how SHIRA can be combined with ideas from Ruhe's Rational Krylov algorithm to yield a method that permits an adjustment of shift after every step of the computation, adding greatly to the flexibility of the algorithm. We call this new method rational SHIRA. A numerical example is presented to demonstrate its efficiency.

Computational methods for reducing the complexity of Finite Element (FE) models in structural dynamics are usually based on modal analysis. Classical approaches such as modal truncation, static condensation (Craig-Bampton, Guyan), and component mode synthesis (CMS) are available in many CAE tools such as ANSYS. In other disciplines, different techniques for Model Order Reduction (MOR) have been developed in the previous 2 decades. Krylov subspace methods are one possible choice and often lead to much smaller models than modal truncation methods given the same prescribed tolerance threshold. They have become available to ANSYS users through the tool mor4ansys. A disadvantage is that neither modal truncation nor CMS nor Krylov subspace methods preserve properties important to control design. System-theoretic methods like balanced truncation approximation (BTA), on the other hand, are directed towards reduced-order models for use in closed-loop control. So far, these methods are considered to be too expensive for large-scale structural models. We show that recent algorithmic advantages lead to MOR methods that are applicable to FE models in structural dynamics and that can easily be integrated into CAE software. We will demonstrate the efficiency of the proposed MOR method based on BTA using a control system including as plant the FE model of a machine tool.

Title: Krylov Subspace Methods - Analysis and Application Author: Tomáš Gergelits Department: Department of Numerical Mathematics Supervisor: prof. Ing. Zdeněk Strakoš, DrSc., Department of Numerical Mathematics Abstract: Convergence behavior of Krylov subspace methods is often studied for linear algebraic systems with symmetric positive definite matrices in terms of the condition number of the system matrix. As recalled in the first part of this thesis, their actual convergence behavior (that can be in practice also substantially affected by rounding errors) is however determined by the whole spectrum of the system matrix, and by the projections of the initial residual to the associated invariant subspaces. The core part of this thesis investigates the spectra of infinite dimensional operators −∇ · (k(x)∇) and −∇ · (K(x)∇), where k(x) is a scalar coefficient function and K(x) is a symmetric tensor function, preconditioned by the Laplace operator. Subsequently, the focus is on the eigenvalues of the matrices that arise from the discretization using conforming finite elements. Assuming continuity of K(x), it is proved that the spectrum of the preconditi- oned infinite dimensional operator is equal to the convex hull of the ranges of the diagonal function entries of Λ(x) from the spectral decomposition K(x) =...

Viele Resultate über MR- und OR-Verfahren zur Lösung linearer Gleichungssysteme bleiben (in leicht modifizierter Form) gültig, wenn der betrachtete Operator nicht invertierbar ist. Neben dem für reguläre Probleme charakteristischen Abbruchverhalten, kann bei einem singulären Gleichungssystem auch ein so genannter singulärer Zusammenbruch auftreten. Für beide Fälle werden verschiedene Charakterisierungen angegeben. Die Unterrauminverse, eine spezielle verallgemeinerte Inverse, beschreibt die Näherungen eines MR-Unterraumkorrektur-Verfahrens. Für Krylov-Unterräume spielt die Drazin-Inverse eine Schlüsselrolle. Bei Krylov-Unterraum-Verfahren kann a-priori entschieden werden, ob ein regulärer oder ein singulärer Abbruch auftritt. Wir können zeigen, dass ein Krylov-Verfahren genau dann für beliebige Startwerte eine Lösung des linearen Gleichungssystems liefert, wenn der Index der Matrix nicht größer als eins und das Gleichungssystem konsistent ist. Die Berechnung stationärer Zustandsverteilungen zeitstetiger Markov-Ketten mit endlichem Zustandsraum stellt eine praktische Aufgabe dar, welche die Lösung eines singulären linearen Gleichungssystems erfordert. Die Eigenschaften der Übergangs-Halbgruppe folgen aus einfachen Annahmen auf rein analytischem und matrixalgebrischen Wege. Insbesondere ist die erzeugende Matrix eine singuläre M-Matrix mit Index 1. Ist die Markov-Kette irreduzibel, so ist die stationäre Zustandsverteilung eindeutig bestimmt.

The discretization of the stationary diffusion equation with random parameters by the Stochastic Finite Element Method requires the solution of a highly structured but very large linear system of equations. Depending on the stochastic properties of the diffusion coefficient together with the stochastic discretization we consider three solver cases. If the diffusion coefficient is given by a stochastically linear expansion, e.g. a truncated Karhunen-Loeve expansion, and tensor product polynomial stochastic shape functions are employed, the Galerkin matrix can be transformed to a block-diagonal matrix. For the solution of the resulting sequence of linear systems we study Krylov subspace recycling methods whose success depends on the ordering and grouping of the linear systems as well as the preconditioner. If we use complete polynomials for the stochastic discretization instead, we show that decoupling of the Galerkin matrix with respect to the stochastic degrees of freedom is impossible. For a stochastically nonlinear diffusion coefficient, e.g. a lognormal random field, together with complete polynomials serving as stochastic shape functions, we introduce and test the performance of a new Kronecker product preconditioner, which is not exclusively based on the mean value of the diffusion coefficient.

The main theme of this work is effectiveness and efficiency of Krylov subspace methods and Krylov subspace recycling. While solving long, slowly changing sequences of large linear systems, such as the ones that arise in engineering, there are many issues we need to consider if we want to make the process reliable (converging to a correct solution) and as fast as possible. This thesis is built on three main components. At first, we target bilinear and quadratic form estimation. Bilinear form $c^TA^{-1}b$ is often associated with long sequences of linear systems, especially in optimization problems. Thus, we devise algorithms that adapt cheap bilinear and quadratic form estimates for Krylov subspace recycling. In the second part, we develop a hybrid recycling method that is inspired by a complex CFD application. We aim to make the method robust and cheap at the same time. In the third part of the thesis, we optimize the implementation of Krylov subspace methods on Graphic Processing Units (GPUs). Since preconditioners based on incomplete matrix factorization (ILU, Cholesky) are very slow on the GPUs, we develop a preconditioner that is effective but well suited for GPU implementation. / Ph. D.

Nous nous intéressons, dans cette thèse, à l'étude des méthodes d'extrapolation polynômiales et à l'application de ces méthodes dans l'accélération de méthodes de points fixes pour des problèmes donnés. L'avantage de ces méthodes d'extrapolation est qu'elles utilisent uniquement une suite de vecteurs qui n'est pas forcément convergente, ou qui converge très lentement pour créer une nouvelle suite pouvant admettreune convergence quadratique. Le développement de méthodes cycliques permet, deplus, de limiter le coût de calculs et de stockage. Nous appliquons ces méthodes à la résolution des équations de Navier-Stokes stationnaires et incompressibles, à la résolution de la formulation Kohn-Sham de l'équation de Schrödinger et à la résolution d'équations elliptiques utilisant des méthodes multigrilles. Dans tous les cas, l'efficacité des méthodes d'extrapolation a été montrée.Nous montrons que lorsqu'elles sont appliquées à la résolution de systèmes linéaires, les méthodes d'extrapolation sont comparables aux méthodes de sous espaces de Krylov. En particulier, nous montrons l'équivalence entre la méthode MMPE et CMRH. Nous nous intéressons enfin, à la parallélisation de la méthode CMRH sur des processeurs à mémoire distribuée et à la recherche de préconditionneurs efficaces pour cette même méthode. / In this thesis, we study polynomial extrapolation methods. We discuss the design and implementation of these methods for computing solutions of fixed point methods. Extrapolation methods transform the original sequance into another sequence that converges to the same limit faster than the original one without having explicit knowledge of the sequence generator. Restarted methods permit to keep the storage requirement and the average of computational cost low. We apply these methods for computing steady state solutions of incompressible flow problems modelled by the Navier-Stokes equations, for solving the Schrödinger equation using the Kohn-Sham formulation and for solving elliptic equations using multigrid methods. In all cases, vector extrapolation methods have a useful role to play. We show that, when applied to linearly generated vector sequences, extrapolation methods are related to Krylov subspace methods. For example, we show that the MMPE approach is mathematically equivalent to CMRH method. We present an implementation of the CMRH iterative method suitable for parallel architectures with distributed memory. Finally, we present a preconditioned CMRH method.

Mit dem Begriff Wellenleiterquantenelektrodynamik (WQED) wird gemeinhin die Physik des quantisierten und in eindimensionalen Wellenleitern geführten Lichtes in Wechselwirkung mit einzelnen Emittern bezeichnet. In dieser Arbeit untersuche ich Effekte der WQED für einzelne Dreiniveausysteme (3NS) bzw. Paare von Zweiniveausystemen (2NS), die in den Wellenleiter eingebettet sind. Hierzu bediene ich mich hauptsächlich numerischer Methoden und betrachte die Modellsysteme im Rahmen der Drehwellennäherung. Ich untersuche die Dynamik der Streuung einzelner Photonen an einzelnen, in den Wellenleiter eingebetteten 3NS. Dabei analysiere ich den Einfluss dunkler bzw. nahezu dunkler Zustände der 3NS auf die Streuung und zeige, wie sich mit Hilfe stationärer elektrischer Treibfelder gezielt auf die Streuung einwirken lässt. Ich quantifiziere Verschränkung zwischen dem Lichtfeld im Wellenleiter und den Emittern mit Hilfe der Schmidt-Zerlegung und untersuche den Einfluss der Form der Einhüllenden eines Einzelphotonpulses auf die Ausbeute der Verschränkungserzeugung bei der Streuung des Photons an einem einzelnen Lambda-System im Wellenleiter. Hier zeigt sich, dass die Breite der Einhüllenden im k-Raum und die Emissionszeiten der beiden Übergänge des 3NS die maßgeblichen Parameter darstellen. Abschließend ergründe ich die Emissionsdynamik zweier im Abstand L in den Wellenleiter eingebetteter 2NS. Diese Dynamik wird insbesondere durch kavitätsartige und polaritonische Zustände des Systems aus Wellenleiter und Emitter ausschlaggebend beeinflusst. Bei der kollektiven Emission der 2NS treten - abhängig vom Abstand L - Sub- bzw. Superradianz auf. Dabei nimmt die Intensität dieser Effekte mit längerem Abstand L zu. Diese Eigenart lässt sich auf die Eindimensionalität des Wellenleiters zurückführen. / The field of waveguide quantum electrodynamics (WQED) deals with the physics of quantised light in one-dimensional (1D) waveguides coupled to single emitters. In this thesis, I investigate WQED effects for single three-level systems (3LS) and pairs of two-level systems (2LS), respectively, which are embedded in the waveguide. To this end, I utilise numerical techniques and consider all model systems within the rotating wave approximation. I investigate the dynamics of single-photon scattering by single, embedded 3LS. In doing so, I analyse the influence of dark and almost-dark states of the 3LS on the scattering dynamics. I also show, how stationary electrical driving fields can control the outcome of the scattering. I quantify entanglement between the waveguide''s light field and single emitters by utilising the Schmidt decomposition. I apply this formalism to a lambda-system embedded in a 1D waveguide and study the generation of entanglement by scattering single-photon pulses with different envelopes on the emitter. I show that this entanglement generation is mainly determined by the photon''s width in k-space and the 3LS''s emission times. Finally, I explore the emission dynamics of a pair of 2LS embedded by a distance L into the waveguide. These dynamics are primarily governed by bound states in the continuum and by polaritonic atom-photon bound-states. For collective emission processes of the two 2LS, sub- and superradiance appear and depend strongly on the 2LS''s distance: the effects increase for larger L. This is an exclusive property of the 1D nature of the waveguide.