An augmented subspace Conjugate Gradient

نویسنده

  • Jocelyne Erhel
چکیده

Many scientiic applications require to solve successively linear systems Ax = b with diierent right-hand sides b and a symmetric positive deenite matrix A. The Conjugate Gradient method applied to the rst system generates a Krylov subspace which can be eeciently recycled thanks to orthogonal projections in subsequent systems. A modiied Conjugate Gradient method is then applied with a speciic initial guess and initial descent direction and a modiied descent direction during the iterations. This paper gives new theoretical results for this method and proposes a new version which seems robust as far as loss of orthogonality is concerned. Numerical experiments show the eecacy of our method even for quite diierent right-hand sides. Gradient Conjugu e avec sous-espace augment e R esum e : Dans beaucoup d'applications scientiiques, il faut r esoudre successivement des syst emes lin eaires du type Ax = b avec dii erents seconds membres b et une matrice sym etrique d eenie positive A. Lorsqu'on applique la m ethode du Gradient Conjugu e au premier syst eme, on g en ere un sous-espace de Krylov que l'on peut recycler eecacement dans les syst emes suivants gr^ ace a des projections orthogonales. On applique alors une m ethode de Gradient Conjugu e modii ee, o u on d eenit une solution et une direction de descente initiales sp eciiques et o u on modiie la direction de descente durant les it erations. Ce rapport pr esente de nouveaux r esultats th eoriques pour cette m ethode et propose une nouvelle version qui semble robuste par rapport a la perte d'orthogonalit e. Les essais num eriques montrent l'eecacit e de notre m ethode m^ eme pour des seconds membres tr es dii erents.

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

Krylov-subspace recycling via the POD-augmented conjugate-gradient algorithm

This work presents a new Krylov-subspace-recycling method for efficiently solving sequences of linear systems of equations characterized by a non-invariant symmetric-positive-definite matrix. As opposed to typical truncation strategies used in recycling such as deflation, we propose a truncation method inspired by goal-oriented proper orthogonal decomposition (POD) from model reduction. This id...

متن کامل

An Augmented Conjugate Gradient Method for Solving Consecutive Symmetric Positive Definite Linear Systems

Many scientific applications require one to solve successively linear systems Ax = b with different right-hand sides b and a symmetric positive definite matrix A. The conjugate gradient method applied to the first system generates a Krylov subspace which can be efficiently recycled thanks to orthogonal projections in subsequent systems. A modified conjugate gradient method is then applied with ...

متن کامل

A Class of Nested Iteration Schemes for Generalized Coupled Sylvester Matrix Equation

Global Krylov subspace methods are the most efficient and robust methods to solve generalized coupled Sylvester matrix equation. In this paper, we propose the nested splitting conjugate gradient process for solving this equation. This method has inner and outer iterations, which employs the generalized conjugate gradient method as an inner iteration to approximate each outer iterate, while each...

متن کامل

An Adaptive POD-Krylov Reduced-Order Model for Structural Optimization

1. Abstract We present an adaptive proper orthogonal decomposition (POD)-Krylov reduced-order model (ROM) for structural optimization. At each step of the optimization loop, we compute approximate solutions to the structural state and sensitivity equations using a novel POD-augmented conjugate gradient (CG) algorithm. This algorithm consists of three stages. In the first two stages, the solutio...

متن کامل

First International Conference and Exhibition Digital Signal Processing ( DSP ' 98 ) Subspace Adaptive Algorithm For Blind Separation Of Convolutive Mixtures By Conjugate Gradient Method

In this paper, a new subspace adaptive algorithm, for blind separation of convolutive mixture, is proposed. This algorithm can be decomposed into two steps: At rst, the convolutive mixture will be reduced to an instantaneous mixture (memoryless mixture), using a second-order statistics criterion based on subspace approach. The second step consists on the separation of the residual instantaneous...

متن کامل

Extrapolation vs. projection methods for linear systems of equations

It is shown that the four vector extrapolation methods, minimal polynomial extrapolation, reduced rank extrapolation, modified minimal polynomial extrapolation, and topological epsilon algorithm, when applied to linearly generated vector sequences, are Krylov subspace methods, and are equivalent to some well known conjugate gradient type methods. A unified recursive method that includes the con...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 1997