Diese Dissertation widmet sich der Untersuchung singulärer stochastischer partieller Differentialgleichungen (engl. SPDEs). Wir entwickeln Erweiterungen der bisherigen Lösungstheorien, zeigen fundamentale Beziehungen zwischen verschiedenen Ansätzen und präsentieren Anwendungen in der Finanzmathematik und der mathematischen Physik. Die Theorie parakontrollierter Systeme wird für diskrete Räume formuliert und eine schwache Universalität für das parabolische Anderson Modell bewiesen. Eine fundamentale Relation zwischen Hairer's modellierten Distributionen und Paraprodukten wird bewiesen: Wir zeigen das sich der Raum modellierter Distributionen durch Paraprodukte beschreiben lässt. Dieses Resultat verallgemeinert die Fourierbeschreibung von Hölderräumen mittels Littlewood-Paley Theorie. Schließlich wird die Existenz von Lösungen der stochastischen Schrödingergleichung auf dem ganzen Raum bewiesen und eine Anwendung Hairer's Theorie zur Preisermittlung von Optionen aufgezeigt. / This thesis is concerned with the study of singular stochastic partial differential equations (SPDEs). We develop extensions to existing solution theories, present fundamental interconnections between different approaches and give applications in financial mathematics and mathematical physics. The theory of paracontrolled distribution is formulated for discrete systems, which allows us to prove a weak universality result for the parabolic Anderson model. This thesis further shows a fundamental relation between Hairer's modelled distributions and paraproducts: The space of modelled distributions can be characterized completely by using paraproducts. This can be seen a generalization of the Fourier description of Hölder spaces. Finally, we prove the existence of solutions to the stochastic Schrödinger equation on the full space and provide an application of Hairer's theory to option pricing.

Disertační práce se zabývá modelováním znečištění ovzduší, jeho transportních a disperzních procesů ve spodní části atmosféry a zejména numerickými metodami, které slouží k řešení těchto modelů. Modelování znečištění ovzduší je velmi důležité pro předpověď kontaminace a pomáhá porozumět samotnému procesu a eliminaci následků. Hlavním tématem práce jsou metody pro řešení modelů popsaných parciálními diferenciálními rovnicemi, přesněji advekčně-difúzní rovnicí. Polovina práce je zaměřena na známou metodu přímek a je zde ukázáno, že tato metoda je vhodná k řešení určitých konkrétních problémů. Dále bylo navrženo a otestováno řešení paralelizace metody přímek, jež ukazuje, že metoda má velký potenciál pro akceleraci na současných grafických kartách a tím pádem i zvětšení přesnosti výpočtu. Druhá polovina práce se zabývá poměrně mladou metodou ELLAM a její aplikací pro řešení atmosférických advekčně-difúzních rovnic. Byla otestována konkrétní forma metody ELLAM společně s navrženými adaptacemi. Z výsledků je zřejmé, že v mnoha případech ELLAM překonává současné používané metody.

No description available.

Grey-box models are constructed by combining model components that are derived from first principles with components that are identified empirically from data. In this thesis a grey-box modelling method for describing distributed parameter systems is presented. The method combines partial differential equations with a multi-layer perceptron network in order to incorporate prior knowledge about the system while identifying unknown dynamics from data. A gradient-based optimization scheme which relies on the reverse mode of automatic differentiation is used to train the network. The method is presented in the context of modelling the dynamics of a chemical reaction in a fluid. Lastly, the grey-box modelling method is evaluated on a one-dimensional and two-dimensional instance of the reaction system. The results indicate that the grey-box model was able to accurately capture the dynamics of the reaction system and identify the underlying reaction. / Hybridmodeller konstrueras genom att kombinera modellkomponenter som härleds från grundläggande principer med modelkomponenter som bestäms empiriskt från data. I den här uppsatsen presenteras en metod för att beskriva distribuerade parametersystem genom hybridmodellering. Metoden kombinerar partiella differentialekvationer med ett neuronnätverk för att inkorporera tidigare känd kunskap om systemet samt identifiera okänd dynamik från data. Neuronnätverket tränas genom en gradientbaserad optimeringsmetod som använder sig av bakåt-läget av automatisk differentiering. För att demonstrera metoden används den för att modellera kemiska reaktioner i en fluid. Metoden appliceras slutligen på ett en-dimensionellt och ett två-dimensionellt exempel av reaktions-systemet. Resultaten indikerar att hybridmodellen lyckades återskapa beteendet hos systemet med god precision samt identifiera den underliggande reaktionen.

No description available.

Piezoelektrika haben ein breit gefächertes Anwendungsspektrum in Industrie, Alltag und Forschung. Dies erfordert ein genaues Wissen über das Materialverhalten der betrachteten piezoelektrischen Elemente, was mit dem Lösen von simulationsgestützten inversen Parameteridentifikationsproblemen einhergeht. Die vorliegende Arbeit befasst sich mit der optimalen Versuchsplanung (OED) für dieses Problem. Piezoelektrische Materialien weisen die Eigenschaft auf, sich als Reaktion auf angelegte Potentiale oder Kräfte mechanisch oder elektrisch zu verändern (direkter und indirekter piezoelektrischer Effekt). Um eine Spannung anzulegen und den indirekten piezoelektrischen Effekt auszunutzen, werden Elektroden aufgebracht, deren Konfiguration einen erheblichen Einfluss auf mögliche Systemantworten hat. Daher werden das Potential, die Anzahl und die Größe der Elektroden zunächst im zweidimensionalen Fall optimiert. Das piezoelektrische Verhalten basiert im betrachteten Kleinsignalbereich auf zeitabhängigen, linearen partiellen Differentialgleichungen. Die Herleitung sowie Existenz und Eindeutigkeit der Lösungen werden gezeigt. Zur Berechnung der elektrischen Ladung und der Impedanz, die für das Materialidentifikationsproblem und damit für die Versuchsplanung relevant sind, werden zeit- und frequenzabhängige Simulationen auf Basis der Finite Elemente Methode (FEM) mit dem FEM Simulationstool FEniCS durchgeführt. Es wird auf Nachteile bei der Berechnung der Ableitungen eingegangen und erste adjungierte Gleichungen formuliert. Die Modellierung des Problems der optimalen Versuchsplanung erfolgt hauptsächlich durch die Kontrolle des Potentials der Dirichlet Randbedingungen des Randwertproblems. Anhand mehrerer numerischer Beispiele werden die resultierenden Konfigurationen gezeigt. Weitere Ansätze zur Elektrodenmodellierung, z.B. durch Kontrolle der Materialeigenschaften, werden ebenfalls vorgestellt. Schließlich wird auf mögliche Erweiterungen des vorgestellten OED Problems hingewiesen. / Piezoelectrics have a wide range of applications in industry, everyday life and research. This requires an accurate knowledge of the material behavior, which implies the solution of simulation-based inverse identification problems. This thesis focuses on the optimal design of experiments addressing this problem. Piezoelectric materials exhibit the property of mechanical or electrical changes in response to applied potentials or forces (direct and indirect piezoelectric effect). To apply voltage and to exploit the indirect piezoelectric effect, electrodes are attached whose configura- tion have a significant influence on possible system responses. Therefore, the potential, the number and the size of the electrodes are initially optimized in the two-dimensional case. The piezoelectric behavior in the considered small signal range is based on a time dependent linear partial differential equation system. The derivation as well as the exis- tence, uniqueness and regularity of the solutions of the equations are shown. Time- and frequency-dependent simulations based on the finite element method (FEM) with the FEM simulation tool FEniCS are performed to calculate the electric charge and the impedance, which are relevant for the material identification problem and thus for the experimental design. Drawbacks in the derivative calculations are pointed out and a first set of adjoint equations is formulated. The modeling of the optimal experimental design (OED) prob- lem is done mainly by controlling the potential of the Dirichlet boundary conditions of the boundary value problem. Several numerical examples are used to show the resulting configurations and to address the difficulties encountered. Further electrode modeling ap- proaches for example by controlling the material properties are then discussed. Finally, possible extensions of the presented OED problem are pointed out.

<p dir="ltr">In this thesis, we consider the inverse problem of reconstructing the shape, position, and size of an unknown scattering object. We will talk about different methods used for nondestructive testing in scattering theory. We will consider qualitative reconstruction methods to understand and determine important information about the support of unknown scattering objects. We will also discuss the material properties of the system and connect them to certain crucial aspects of the region of interest, as well as develop useful techniques to determine physical information using inverse scattering theory. </p><p><br></p><p dir="ltr">In the first part of the analysis, we consider the transmission eigenvalue (TE) problem associated with the scattering of a plane wave for an isotropic scatterer. In particular, we examine the transmission eigenvalue problem with two conductivity boundary parameters. In previous studies, this eigenvalue problem was analyzed with one conductive boundary parameter, whereas we will consider the case of two parameters. We will prove the existence and discreteness of the transmission eigenvalues. In addition, we will study the dependence of the TE's on the physical parameters and connect the first transmission eigenvalue to the physical parameters of the problem by a monotone-type argument. Lastly, we will consider the limiting procedure as the second boundary parameter vanishes at the boundary of the scattering region and provide numerical examples to validate the theory presented in Chapter 2. </p><p><br></p><p dir="ltr">The connection between transmission eigenvalues and the system's physical parameters provides a way to do testing in a nondestructive way. However, to understand the region of interest in terms of its shape, size, and position, one needs to use different techniques. As a result, we consider reconstructing extended scatterers using an analogous method to the Direct Sampling Method (DSM), a new sampling method based on the Landweber iteration. We will need a factorization of the far-field operator to analyze the corresponding imaging function for the new Landweber direct sampling method. Then, we use the factorization and the Funk--Hecke integral identity to prove that the new imaging function will accurately recover the scatterer. The method studied here falls under the category of qualitative reconstruction methods, where an imaging function is used to retrieve the scatterer. We prove the stability of our new imaging function as well as derive a discrepancy principle for recovering the regularization parameter. The theoretical results are verified with numerical examples to show how the reconstruction performs by the new Landweber direct sampling method.</p><p><br></p><p dir="ltr">Motivated by the work done with the transmission eigenvalue problem with two conductivity parameters, we also study the direct and inverse problem for isotropic scatterers with two conductive boundary conditions. In such a problem, one analyzes the behavior of the scattered field as one of the conductivity parameters vanishes at the boundary. Consequently, we prove the convergence of the scattered field dealing with two conductivity parameters to the scattered field dealing with only one conductivity parameter. We consider the uniqueness of recovering the coefficients from the known far-field data at a fixed incident direction for multiple frequencies. Then, we consider the inverse shape problem for recovering the scatterer for the measured far-field data at a fixed frequency. To this end, we study the direct sampling method for recovering the scatterer by studying the factorization for the far-field operator. The direct sampling method is stable concerning noisy data and valid in two dimensions for partial aperture data. The theoretical results are verified with numerical examples to analyze the performance using the direct sampling method. </p>

<p>The need to carry out Uncertainty quantification (UQ) is ubiquitous in science and engineering. However, carrying out UQ for real-world problems is not straightforward and they require a lot of computational budget and resources. The objective of this thesis is to develop computationally efficient approaches based on machine learning to carry out UQ. Specifically, we addressed two problems.</p> <p><br></p> <p>The first problem is, it is difficult to carry out Uncertainty propagation (UP) in systems governed by elliptic PDEs with spatially varying uncertain fields in coefficients and boundary conditions. Here as we have functional uncertainties, the number of uncertain parameters is large. Unfortunately, in these situations to carry out UP we need to solve the PDE a large number of times to obtain convergent statistics of the quantity governed by the PDE. However, solving the PDE by a numerical solver repeatedly leads to a computational burden. To address this we proposed to learn the surrogate of the solution of the PDE in a data-free manner by utilizing the physics available in the form of the PDE. We represented the solution of the PDE as a deep neural network parameterized function in space and uncertain parameters. We introduced a physics-informed loss function derived from variational principles to learn the parameters of the network. The accuracy of the learned surrogate is validated against the corresponding ground truth estimate from the numerical solver. We demonstrated the merit of using our approach by solving UP problems and inverse problems faster than by using a standard numerical solver.</p> <p><br></p> <p>The second problem we focused on in this thesis is related to inverse problems. State of the art approach to solving inverse problems involves posing the inverse problem as a Bayesian inference task and estimating the distribution of input parameters conditioned on the observed data (posterior). Markov Chain Monte Carlo (MCMC) methods and variational inference methods provides us ways to estimate the posterior. However, these inference techniques need to be re-run whenever a new set of observed data is given leading to a computational burden. To address this, we proposed to learn a Bayesian inverse map i.e., the map from the observed data to the posterior. This map enables us to do on-the-fly inference. We demonstrated our approach by solving various examples and we validated the posteriors learned from our approach against corresponding ground truth posteriors from the MCMC method.</p>

No description available.

Ce travail présente quelques résultats sur les systèmes de particules en interaction pour l'interprétation probabiliste des équations aux dérivées partielles, avec des applications à des questions de dynamique moléculaire et de chimie quantique. On présente notamment une méthode particulaire permettant d'analyser le processus de la force biaisante adaptative, utilisé en dynamique moléculaire pour le calcul de différences d'énergies libres. On étudie également la sensibilité de dynamiques stochastiques par rapport à un paramètre, en vue du calcul des forces dans l'approximation de Born-Oppenheimer pour rechercher l'état quantique fondamental de molécules. Enfin, on présente un schéma numérique basé sur un système de particules pour résoudre des lois de conservation scalaires, avec un terme de diffusion anormale se traduisant par une dynamique de sauts sur les particules / This work presents some results on stochastically interacting particle systems and probabilistic interpretations of partial differential equations with applications to molecular dynamics and quantum chemistry. We present a particle method allowing to analyze the adaptive biasing force process, used in molecular dynamics for the computation of free energy differences. We also study the sensitivity of stochastic dynamics with respect to some parameter, aiming at the computation of forces in the Born-Oppenheimer approximation for determining the fundamental quantum state of molecules. Finally, we present a numerical scheme based on a particle system for the resolution of scalar conservation laws with an anomalous diffusion term, corresponding to a jump dynamics on the particles