Méthode itérative

Résumé

En analyse numérique, une méthode itérative est un procédé algorithmique utilisé pour résoudre un problème, par exemple la recherche d’une solution d’un système d'équations ou d’un problème d’optimisation. En débutant par le choix d’un point initial considéré comme une première ébauche de solution, la méthode procède par itérations au cours desquelles elle détermine une succession de solutions approximatives raffinées qui se rapprochent graduellement de la solution cherchée. Les points générés sont appelés des itérés. Comparaison Les méthodes itératives contrastent avec les méthodes directes qui résolvent le problème en une seule étape (par exemple la solution d'un système linéaire Ax = b obtenue en calculant la matrice inverse de A). Les méthodes itératives se substituent avantageusement aux autres lorsque :

celles-ci sont inapplicables, coûteuses ou simplement inconnues ;
le problème est mal conditionné ou comprend un grand nombre de variables, car les solutions su

Source officielle

https://fr.wikipedia.org/wiki/M%C3%A9thode_it%C3%A9rative

À propos de ce résultat

Cette page est générée automatiquement et peut contenir des informations qui ne sont pas correctes, complètes, à jour ou pertinentes par rapport à votre recherche. Il en va de même pour toutes les autres pages de ce site. Veillez à vérifier les informations auprès des sources officielles de l'EPFL.

Publications associées (87)

Publications associées

Chargement

Personnes associées (12)

Personnes associées

Chargement

Unités associées (6)

Unités associées

Chargement

Concepts associés (25)

Concepts associés

Chargement

Cours associés (65)

Cours associés

Chargement

Séances de cours associées (199)

Séances de cours associées

Chargement

Multi space reduced basis preconditioners for parametrized partial differential equations

Niccolò Dal Santo

The multiquery solution of parametric partial differential equations (PDEs), that is, PDEs depending on a vector of parameters, is computationally challenging and appears in several engineering contexts, such as PDE-constrained optimization, uncertainty quantification or sensitivity analysis. When using the finite element (FE) method as approximation technique, an algebraic system must be solved for each instance of the parameter, leading to a critical bottleneck when we are in a multiquery context, a problem which is even more emphasized when dealing with nonlinear or time dependent PDEs. Several techniques have been proposed to deal with sequences of linear systems, such as truncated Krylov subspace recycling methods, deflated restarting techniques and approximate inverse preconditioners; however, these techniques do not satisfactorily exploit the parameter dependence. More recently, the reduced basis (RB) method, together with other reduced order modeling (ROM) techniques, emerged as an efficient tool to tackle parametrized PDEs. In this thesis, we investigate a novel preconditioning strategy for parametrized systems which arise from the FE discretization of parametrized PDEs. Our preconditioner combines multiplicatively a RB coarse component, which is built upon the RB method, and a nonsingular fine grid preconditioner. The proposed technique hinges upon the construction of a new Multi Space Reduced Basis (MSRB) method, where a RB solver is built at each step of the chosen iterative method and trained to accurately solve the error equation. The resulting preconditioner directly exploits the parameter dependence, since it is tailored to the class of problems at hand, and significantly speeds up the solution of the parametrized linear system. We analyze the proposed preconditioner from a theoretical standpoint, providing assumptions which lead to its well-posedness and efficiency. We apply our strategy to a broad range of problems described by parametrized PDEs: (i) elliptic problems such as advection-diffusion-reaction equations, (ii) evolution problems such as time-dependent advection-diffusion-reaction equations or linear elastodynamics equations (iii) saddle-point problems such as Stokes equations, and, finally, (iv) Navier-Stokes equations. Even though the structure of the preconditioner is similar for all these classes of problems, its fine and coarse components must be accurately chosen in order to provide the best possible results. Several comparisons are made with respect to the current state-of-the-art preconditioning and ROM techniques. Finally, we employ the proposed technique to speed up the solution of problems in the field of cardiovascular modeling.

EPFL2018

Chargement

This thesis focuses on non-parametric covariance estimation for random surfaces, i.e.~functional data on a two-dimensional domain. Non-parametric covariance estimation lies at the heart of functional data analysis, andconsiderations of statistical and computational efficiency often compel the use of separability of the covariance, when working with random surfaces. We seek to provide efficient alternatives to this ambivalent assumption.In Chapter 2, we study a setting where the covariance structure may fail to be separable locally -- either due to noise contamination or due to the presence of a non-separable short-range dependent signal component. That is, the covariance is an additive perturbation of a separable component by a non-separable but banded component. We introduce non-parametric estimators hinging on shifted partial tracing -- a novel concept enjoying strong denoising properties. We illustrate the usefulness of the proposed methodology on a data set of mortality surfaces.In Chapter 3, we propose a distinctive decomposition of the covariance, which allows us to understand separability as an unconventional form of low-rankness. From this perspective, a separable covariance has rank one. Allowing for a higher rank suggests a structured class in which any covariance can be approximated up to an arbitrary precision. The key notion of the partial inner product allows us to generalize the power iteration method to general Hilbert spaces and estimate the aforementioned decomposition from data. Truncation and retention of the leading terms automatically induces a non-parametric estimator of the covariance, whose parsimony is dictated by the truncation level. Advantages of this approach, allowing for estimation beyond separability, are demonstrated on the task of classification of EEG signals.While Chapters 2 and 3 propose several generalizations of separability in the densely sampled regime, Chapter 4 deals with the sparse regime, where the latent surfaces are observed only at few irregular locations. Here, a separable covariance estimator based on local linear smoothers is proposed, which is the first non-parametric utilization of separability in the sparse regime. The assumption of separability reduces the intrinsically four-dimensional smoothing problem into several two-dimensional smoothers and allows the proposed estimator to retain the classical minimax-optimal convergence rate for two-dimensional smoothers. The proposed methodology is used for a qualitative analysis of implied volatility surfaces corresponding to call options, and for prediction of the latent surfaces based on information from the entire data set, allowing for uncertainty quantification. Our quantitative results show that the proposed methodology outperforms the common approach of pre-smoothing every implied volatility surface separately.Throughout the thesis, we put emphasis on computational aspects, since those are the main reason behind the immense popularity of separability. We show that the covariance structures of Chapters 2 and 3 come with no (asymptotic) computational overhead relative to assuming separability. In fact, the proposed covariance structures can be estimated and manipulated with the same asymptotic costs as the separable model. In particular, we develop numerical algorithms that can be used for efficient inversion, as required e.g.~for prediction. All the methods are implemented in R and available on~GitHub.

EPFL2022

Chargement

Expectation propagation in the large data limit

Guillaume Philippe Ivan Joseph Dehaene

Expectation propagation (EP) is a widely successful algorithm for variational inference. EP is an iterative algorithm used to approximate complicated distributions, typically to find a Gaussian approximation of posterior distributions. In many applications of this type, EP performs extremely well. Surprisingly, despite its widespread use, there are very few theoretical guarantees on Gaussian EP, and it is quite poorly understood. To analyse EP, we first introduce a variant of EP: averaged EP, which operates on a smaller parameter space. We then consider averaged EP and EP in the limit of infinite data, where the overall contribution of each likelihood term is small and where posteriors are almost Gaussian. In this limit, we prove that the iterations of both averaged EP and EP are simple: they behave like iterations of Newton's algorithm for finding the mode of a function. We use this limit behaviour to prove that EP is asymptotically exact, and to obtain other insights into the dynamic behaviour of EP, e.g. that it may diverge under poor initialization exactly like Newton's method. EP is a simple algorithm to state, but a difficult one to study. Our results should facilitate further research into the theoretical properties of this important method.

Wiley2018

Chargement

Multi space reduced basis preconditioners for parametrized partial differential equations

Niccolò Dal Santo

EPFL2018

EPFL2022