71 |
Tensor product methods in numerical simulation of high-dimensional dynamical problemsDolgov, Sergey 20 August 2014 (has links)
Quantification of stochastic or quantum systems by a joint probability density or wave function is a notoriously difficult computational problem, since the solution depends on all possible states (or realizations) of the system.
Due to this combinatorial flavor, even a system containing as few as ten particles may yield as many as $10^{10}$ discretized states.
None of even modern supercomputers are capable to cope with this curse of dimensionality straightforwardly, when the amount of quantum particles, for example, grows up to more or less interesting order of hundreds.
A traditional approach for a long time was to avoid models formulated in terms of probabilistic functions,
and simulate particular system realizations in a randomized process.
Since different times in different communities, data-sparse methods came into play.
Generally, they aim to define all data points indirectly, by a map from a low amount of representers,
and recast all operations (e.g. linear system solution) from the initial data to the effective parameters.
The most advanced techniques can be applied (at least, tried) to any given array, and do not rely explicitly on its origin.
The current work contributes further progress to this area in the particular direction: tensor product methods for separation of variables.
The separation of variables has a long history, and is based on the following elementary concept: a function of many variables may be expanded as a product of univariate functions.
On the discrete level, a function is encoded by an array of its values, or a tensor.
Therefore, instead of a huge initial array, the separation of variables allows to work with univariate factors with much less efforts.
The dissertation contains a short overview of existing tensor representations: canonical PARAFAC, Hierarchical Tucker, Tensor Train (TT) formats, as well as the artificial tensorisation, resulting in the Quantized Tensor Train (QTT) approximation method.
The contribution of the dissertation consists in both theoretical constructions and practical numerical algorithms for high-dimensional models, illustrated on the examples of the Fokker-Planck and the chemical master equations.
Both arise from stochastic dynamical processes in multiconfigurational systems, and govern the evolution of the probability function in time.
A special focus is put on time propagation schemes and their properties related to tensor product methods.
We show that these applications yield large-scale systems of linear equations,
and prove analytical separable representations of the involved functions and operators.
We propose a new combined tensor format (QTT-Tucker), which descends from the TT format (hence TT algorithms may be generalized smoothly), but provides complexity reduction by an order of magnitude.
We develop a robust iterative solution algorithm, constituting most advantageous properties of the classical iterative methods from numerical analysis and alternating density matrix renormalization group (DMRG) techniques from quantum physics.
Numerical experiments confirm that the new method is preferable to DMRG algorithms.
It is as fast as the simplest alternating schemes, but as reliable and accurate as the Krylov methods in linear algebra.
|
72 |
The Eyring-Kramers formula for Poincaré and logarithmic Sobolev inequalities / Die Eyring-Kramer-Formel für Poincaré- und logarithmische Sobolev-UngleichungenSchlichting, André 25 October 2012 (has links)
The topic of this thesis is a diffusion process on a potential landscape which is given by a smooth Hamiltonian function in the regime of small noise. The work provides a new proof of the Eyring-Kramers formula for the Poincaré inequality of the associated generator of the diffusion. The Poincaré inequality characterizes the spectral gap of the generator and establishes the exponential rate of convergence towards equilibrium in the L²-distance. This result was first obtained by Bovier et. al. in 2004 relying on potential theory.
The presented approach in the thesis generalizes to obtain also asymptotic sharp estimates of the constant in the logarithmic Sobolev inequality. The optimal constant in the logarithmic Sobolev inequality characterizes the convergence rate to equilibrium with respect to the relative entropy, which is a stronger distance as the L²-distance and slightly weaker than the L¹-distance. The optimal constant has here no direct spectral representation.
The proof makes use of the scale separation present in the dynamics. The Eyring-Kramers formula follows as a simple corollary from the two main results of the work: The first one shows that the associated Gibbs measure restricted to a basin of attraction has a good Poincaré and logarithmic Sobolev constants providing the fast convergence of the diffusion to metastable states. The second main ingredient is a mean-difference estimate. Here a weighted transportation distance is used. It contains the main contribution to the Poincaré and logarithmic Sobolev constant, resulting from exponential long waiting times of jumps between metastable states of the diffusion.
|
73 |
Grandes d´eviations de matrices aléatoires et équation de Fokker-Planck libre / Large deviations of random matrices and free Fokker-Planck equationGroux, Benjamin 09 December 2016 (has links)
Cette thèse s'inscrit dans le domaine des probabilités et des statistiques, et plus précisément des matrices aléatoires. Dans la première partie, on étudie les grandes déviations de la mesure spectrale de matrices de covariance $XX^*$, où $X$ est une matrice aléatoire rectangulaire à coefficients i.i.d. ayant une queue de probabilité en $exp(-at^{alpha})$, $alpha in ]0,2[$. On établit un principe de grandes déviations analogue à celui de Bordenave et Caputo, de vitesse $n^{1+alpha/2}$ et de fonction de taux explicite faisant intervenir la convolution libre rectangulaire. La démonstration repose sur un résultat de quantification de la liberté asymptotique dans le modèle information-plus-bruit. La seconde partie de cette thèse est consacrée à l'étude du comportement en temps long de la solution de l'équation de Fokker-Planck libre en présence du potentiel quartique $V(x) = frac14 x^4 + frac{c}{2} x^2$ avec $c ge -2$. On montre que quand $t to +infty$, la solution $mu_t$ de cette équation aux dérivées partielles converge en distance de Wasserstein vers la mesure d'équilibre associée au potentiel $V$. Ce résultat fournit un premier exemple de convergence en temps long de la solution de l'équation des milieux granulaires en présence d'un potentiel non convexe et d'une interaction logarithmique. Sa démonstration utilise notamment des techniques de probabilités libres. / This thesis lies within the field of probability and statistics, and more precisely of random matrix theory. In the first part, we study the large deviations of the spectral measure of covariance matrices XX*, where X is a rectangular random matrix with i.i.d. coefficients having a probability tail like $exp(-at^{alpha})$, $alpha in (0,2)$. We establish a large deviation principle similar to Bordenave and Caputo's one, with speed $n^{1+alpha/2}$ and explicit rate function involving rectangular free convolution. The proof relies on a quantification result of asymptotic freeness in the information-plus-noise model. The second part of this thesis is devoted to the study of the long-time behaviour of the solution to free Fokker-Planck equation in the setting of the quartic potential $V(x) = frac14 x^4 + frac{c}{2} x^2$ with $c ge -2$. We prove that when $t to +infty$, the solution $mu_t$ to this partial differential equation converge in Wasserstein distance towards the equilibrium measure associated to the potential $V$. This result provides a first example of long-time convergence for the solution of granular media equation with a non-convex potential and a logarithmic interaction. Its proof involves in particular free probability techniques.
|
74 |
On the diffusion in inhomogeneous systemsHeidernätsch, Mario 29 May 2015 (has links)
Ziel dieser Arbeit ist die Untersuchung des Einflusses der stochastischen Interpretation der Langevin Gleichung mit zustandsabhängigen Diffusionskoeffizienten auf den Propagator des zugehörigen stochastischen Prozesses bzw. dessen Mittelwerte. Dies dient dem besseren Verständnis und der Interpretation von Messdaten von Diffusion in inhomogenen Systemen und geht einher mit der Frage der Form der Diffusionsgleichung in solchen Systemen. Zur Vereinfachung der Fragestellung werden in dieser Arbeit nur Systeme untersucht die vollständig durch einen ortsabhängigen Diffusionskoeffizienten und Angabe der stochastischen Interpretation beschrieben werden können.
Dazu wird zunächst für mehrere experimentell relevante eindimensionale Systeme der jeweilige allgemeine Propagator bestimmt, der für jede denkbare stochastische Interpretation gültig ist. Der analytisch bestimmte Propagator wird dann für zwei exemplarisch ausgewählte stochastische Interpretationen, hier für die Itô und Klimontovich-Hänggi Interpretation, gegenübergestellt und die Unterschiede identifiziert. Für Mittelwert und Varianz der Prozesse werden die drei wesentlichen stochastischen Interpretationen verglichen, also Itô, Stratonovich und Klimontovich-Hänggi Interpretation. Diese systematische Untersuchung von inhomogenen Diffusionsprozessen kann zukünftig helfen diese Art von, in genau einer stochastischen Interpretation, driftfreien Systemen einfacher zu identifizieren.
Ein weiterer wesentlicher Teil der Arbeit erweitert die Frage auf mehrdimensionale inhomogene anisotrope Systeme. Dies wird z.B. bei der Untersuchung von Diffusion in Flüssigkristallen mit inhomogenem Direktorfeld relevant. Obwohl hier, im Gegensatz zu eindimensionalen Systemen, der Propagator nicht allgemein berechnet werden kann, wird dennoch der Einfluss der Inhomogenität auf Messgrößen, wie die mittlere quadratische Verschiebung oder die Verteilung der Diffusivitäten, bestimmt. Anhand eines Beispiels wird auch der Einfluss der stochastischen Interpretation auf diese Messgrößen demonstriert. / The aim of this thesis is to investigate the influence of the stochastic interpretation of the Langevin equation with state-dependent diffusion coefficient on the propagator of the related stochastic process, or its averages, respectively. This helps to obtain a deeper understanding and to interpret measurement data of diffusion in inhomogeneous systems and is accompanied with the question of the proper form of the diffusion equation in such systems. To simplify the question, in this thesis only systems are considered which can be fully described by a spatially dependent diffusion coefficient and a given stochastic interpretation.
Therefore, for several experimentally relevant one-dimensional systems, the respective general propagator is determined, which is valid for any possible stochastic interpretation. Then, the propagator for two exemplary stochastic interpretations, here the Itô and Klimontovich-Hänggi interpretation, are compared and the differences are identified. For mean and variance of the processes three major interpretations are compared, namely the Itô, the Stratonovich and the Klimontovich-Hänggi interpretation. This systematic research on inhomogeneous diffusion process may help in future to identify these kind of, in exactly one stochastic interpretation, drift-free systems more easily.
Another important part of this thesis extends this question to multidimensional inhomogeneous anisotropic systems. This is of high relevance, for instance, for the research of diffusion in liquid crystalline systems with an inhomogeneous director field. Although, in contrast to one-dimensional systems, the propagator may not be calculated generally, the influence of the inhomogeneity on measurement data like the mean squared displacement or the distribution of diffusivities is determined. Based on one example, also the influence of the stochastic interpretation on these quantities is demonstrated.
|
75 |
One-Dimensional Velocity Distributions of Fast Ions under RF Heating Including Doppler Shift in TokamaksBähner, Lukas January 2022 (has links)
The goal of nuclear fusion research is to create a clean and virtually limitless energy source. In order to that, a plasma must be heated to hundreds of millions degrees Celsius. A commonly used heating mechanism is ion cyclotron resonance heating, where antennas emit radio waves into the plasma. The wave can resonate with the ions at their cyclotron frequency, which leads to wave absorption. In order to investigate and improve the heating, one can perform computer simulations. FEMIC is a finite element model for ICRH that calculates the wave field created by the antennas. However, this code does not take into account how the wave modifies the velocity distribution of the plasma. Therefore, a time-independent Fokker-Planck solver is implemented that computes the fast ion distribution due to the incident wave field calculated with FEMIC. The novelty of this code is to include Doppler shift, which influences where ions resonate and how they are heated. / Målet med fusionsforskningen är att skapa en ren energikälla som kan producera obegränsade mängder energi. För detta krävs att ett plasma värms till hundratals miljoner grader Celsius. En vanlig teknik för att värma plasmat är joncyklotronuppvärmning, där en antenn emitterar radiovågor som propagerar in i plasmat. Om vågen är i resonans med jonernas cyklotronrörelse leder detta till att vågen absorberas av jonerna. För att studera och utveckla denna uppvärmningsteknik kan man använda datorsimuleringar. FEMIC är en kod baserad på den finita elementmetoden som beräknar vågfälten som skapas av antennen. Med denna kod kan vi dock inte beräkna hur vågen påverkar jonernas fördelningsfunktioner. Därför har en Fokker-Planck-lösare implementerats som kan beräkna fördelningen av snabba joner som accelererats av vågfältet från FEMIC. Det nya i denna modell är att koden tar hänsyn till Dopplerskiftet, vilket påverkar var jonerna är i resonans med vågen och hur de värms upp.
|
76 |
Existence et stabilité de solutions fortes en théorie cinétique des gaz / Existence and stability of strong solutions in kinetic theoryTristani, Isabelle 22 June 2015 (has links)
Cette thèse est centrée sur l’étude d’équations issues de la théorie cinétique des gaz. Dans tous les problèmes qui y sont explorés, une analyse des problèmes linéaires ou linéarisés associés est réalisée d’un point de vue spectral et du point de vue des semi-groupes. A cela s’ajoute une analyse de la stabilité non linéaire lorsque le modèle est non linéaire. Plus précisément, dans une première partie, nous nous intéressons aux équations de Fokker-Planck fractionnaire et Boltzmann sans cut-off homogène en espace et nous prouvons un retour vers l’équilibre des solutions de ces équations avec un taux exponentiel dans des espaces de type L1 à poids polynomial. Concernant l’équation de Landau inhomogène en espace, nous développons une théorie de Cauchy de solutions perturbatives dans des espaces de type L2 avec différents poids (polynomiaux ou exponentiels) et nous prouvons également la stabilité exponentielle de ces solutions.Nous démontrons ensuite pour l’équation de Boltzmann inélastique inhomogène avec terme diffusif le même type de résultat dans des espaces L1 à poids polynomial dans un régime de faible inélasticité. Pour finir, nous étudions dans un cadre général et uniforme des modèles qui convergent vers l’équation de Fokker-Planck du point de vue de l’analyse spectrale et des semi-groupes. / The topic of this thesis is the study of models coming from kinetic theory. In all the problems that are addressed, the associated linear or linearized problem is analyzed from a spectral point of view and from the point of view of semigroups. Tothat, we add the study of the nonlinear stability when the equation is nonlinear. More precisely, to begin with, we treat the problem of trend to equilibrium for the fractional Fokker-Planck and Boltzmann without cut-off equations, proving an exponential decay to equilibrium in spaces of type L1 with polynomial weights. Concerning the inhomogeneous Landau equation, we develop a Cauchy theory of perturbative solutions in spaces of type L2 with various weights such as polynomial and exponential weights and we also prove the exponential stability of these solutions. Then, we prove similar results for the inhomogeneous inelastic diffusively driven Boltzmann equation in a small inelasticity regime in L1 spaces with polynomial weights. Finally, we study in the same and uniform framework from the spectral analysis point of view with a semigroup approach several Fokker-Planck equations which converge towards the classical one.
|
Page generated in 0.0847 seconds