Méthodes de projection-minimisation pour les problèmes linéaires
Methods for solving some parameter problems of algebra.
Metoda konjugovaných gradientů jako dobrodružství jdoucí přes staletí
Metoda konjugovaných gradientů a Lanczosova metoda tvoří historický a metodologický základ tzv. metod krylovovských podprostorů pro numerickou aproximaci řešení lineárních rovnic a částečnou aproximaci spektra lineárních operátorů. Ačkoliv jsou v obecném povědomí spojovány především s numerickým řešením velmi rozsáhlých soustav lineárních algebraických rovnic a aproximací vlastních čísel velkých matic, je přirozené uvažovat jejich formulaci v kontextu operátorů na Hilbertových prostorech (konečné...
Metodi di Krylov per sistemi lineari di Equazioni Differenziali Ordinarie
Metodi polinomiali per il calcolo di funzioni di matrici non simmetriche e di grandi dimensioni
Metodi probabilistici per il calcolo di autovalori ed autovettori
Métodos para la actualización de los factores de Q y R de una matriz.
Recientemente se han propuesto varios métodos para modificar los factores Q y R de una matriz una vez que se ha eliminado (o añadido) una fila o una columna. Normalmente la descripción de estos métodos se efectúa en el contexto de una determinada aplicación; quizá sea ésta la causa de su escasa difusión.
Minimax Estimators for a Bounded Location Parameter.
Minimization of the spectral norm of the SOR operator in a mixed case.
Minimization properties and short recurrences for Krylov subspace methods.
Minimum Relative Error Approximations for 1/t.
Mixed finite element approximation of an MHD problem involving conducting and insulating regions : the 2D case
We show that the Maxwell equations in the low frequency limit, in a domain composed of insulating and conducting regions, has a saddle point structure, where the electric field in the insulating region is the Lagrange multiplier that enforces the curl-free constraint on the magnetic field. We propose a mixed finite element technique for solving this problem, and we show that, under mild regularity assumption on the data, Lagrange finite elements can be used as an alternative to edge elements.
Mixed Finite Element approximation of an MHD problem involving conducting and insulating regions: the 2D case
We show that the Maxwell equations in the low frequency limit, in a domain composed of insulating and conducting regions, has a saddle point structure, where the electric field in the insulating region is the Lagrange multiplier that enforces the curl-free constraint on the magnetic field. We propose a mixed finite element technique for solving this problem, and we show that, under mild regularity assumption on the data, Lagrange finite elements can be used as an alternative to edge elements.
Mixed Finite Elements for Second Order Elliptic Problems in Three Variables.
Mixed precision GMRES-based iterative refinement with recycling
With the emergence of mixed precision hardware, mixed precision GMRES-based iterative refinement schemes for solving linear systems have recently been developed. However, in certain settings, GMRES may require too many iterations per refinement step, making it potentially more expensive than the alternative of recomputing the LU factors in a higher precision. In this work, we incorporate the idea of Krylov subspace recycling, a well-known technique for reusing information across sequential invocations,...
Model analysis of BPX preconditioner based on smoothed aggregation
We prove nearly uniform convergence bounds for the BPX preconditioner based on smoothed aggregation under the assumption that the mesh is regular. The analysis is based on the fact that under the assumption of regular geometry, the coarse-space basis functions form a system of macroelements. This property tends to be satisfied by the smoothed aggregation bases formed for unstructured meshes.
Moderate deviations in a random graph and for the spectrum of Bernoulli random matrices.
Modification of the alternating direction implicit method and connections with von Neumann's method
Modified block-approximate factorization strategies.
Modified Gram-Schmidt for solving linear least squares problems is equivalent to Gaussian elimination for the normal equations