A framework of conjugate direction methods for symmetric linear systems in optimization
From MaRDI portal
Publication:2342136
Abstract: In this paper we introduce a parameter dependent class of Krylov-based methods, namely CD, for the solution of symmetric linear systems. We give evidence that in our proposal we generate sequences of conjugate directions, extending some properties of the standard Conjugate Gradient (CG) method, in order to preserve the conjugacy. For specific values of the parameters in our framework we obtain schemes equivalent to both the CG and the scaled-CG. We also prove the finite convergence of the algorithms in CD, and we provide some error analysis. Finally, preconditioning is introduced for CD, and we show that standard error bounds for the preconditioned CG also hold for the preconditioned CD.
This paper deals with iterative Krylov-based methods for the solution of symmetric linear systems, arising in both numerical analysis and optimization contexts. The author addresses a parameter-dependent class of conjugate gradient (CG)-based methods, which can equivalently reduce to the CG for a suitable choice of the parameters. A general framework of iterative methods is presented, inspired by polarity for quadratic hypersurfaces, and based on the generation of conjugate directions. Some numerical results are included.
Recommendations
Cites work
- A nonmonotone truncated Newton-Krylov method exploiting negative curvature directions, for large scale unconstrained optimization
- A survey of truncated-Newton methods
- A truncated Newton method with non-monotone line search for unconstrained optimization
- Assessing a search direction within a truncated Newton method
- Automatic Preconditioning by Limited Memory Quasi-Newton Updating
- Exploiting negative curvature directions in linesearch methods for unconstrained optimization
- scientific article; zbMATH DE number 3648910 (Why is no real title available?)
- scientific article; zbMATH DE number 3881771 (Why is no real title available?)
- scientific article; zbMATH DE number 3686229 (Why is no real title available?)
- scientific article; zbMATH DE number 1049350 (Why is no real title available?)
- scientific article; zbMATH DE number 1953444 (Why is no real title available?)
- scientific article; zbMATH DE number 846277 (Why is no real title available?)
- scientific article; zbMATH DE number 5060482 (Why is no real title available?)
- scientific article; zbMATH DE number 961607 (Why is no real title available?)
- Iterative computation of negative curvature directions in large scale optimization
- Iterative Solution Methods
- Iterative solution of linear systems in the 20th century
- Lanczos conjugate-gradient method and pseudoinverse computation on indefinite and singular systems
- Methods of conjugate gradients for solving linear systems
- On A Class of Limited Memory Preconditioners For Large Scale Linear Systems With Multiple Right-Hand Sides
- Planar conjugate gradient algorithm for large-scale unconstrained optimization. I: Theory
- Planar conjugate gradient algorithm for large-scale unconstrained optimization. II: Application
- Preconditioning Newton-Krylov methods in nonconvex large scale optimization
- Predicting the Behavior of Finite Precision Lanczos and Conjugate Gradient Computations
- The Lanczos and Conjugate Gradient Algorithms
- Trust Region Methods
Cited in
(4)- Conjugate direction methods and polarity for quadratic hypersurfaces
- Polarity and conjugacy for quadratic hypersurfaces: a unified framework with recent advances
- scientific article; zbMATH DE number 624931 (Why is no real title available?)
- Delayed gradient methods for symmetric and positive definite linear systems
This page was built for publication: A framework of conjugate direction methods for symmetric linear systems in optimization
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2342136)
