Preconditioning a product of matrices arising in trust region subproblems
In solving large scale optimization problems, we find it advantageous to use iterative methods to solve the sparse linear systems that arise. In the ETR software for solving equality constrained optimization problems, we use a conjugate gradient method to approximately solve the trust region subproblems. To speed up the convergence of the conjugate gradient routine, we need to precondition matrices of the form Z{sup T} W Z, which are not explicitly stored. Four preconditioners were implemented and the results for each are given.
- Research Organization:
- Sandia National Labs., Livermore, CA (United States)
- Sponsoring Organization:
- USDOE, Washington, DC (United States)
- DOE Contract Number:
- AC04-94AL85000
- OSTI ID:
- 221914
- Report Number(s):
- SAND--96-8221; ON: DE96009180
- Country of Publication:
- United States
- Language:
- English
Similar Records
Efficient proximal subproblem solvers for a nonsmooth trust-region method
A fast eigenvalue approach for solving the trust region subproblem with an additional linear inequality
Polynomial preconditioning for conjugate gradient methods
Journal Article
·
Fri Jan 03 23:00:00 EST 2025
· Computational Optimization and Applications
·
OSTI ID:2505123
A fast eigenvalue approach for solving the trust region subproblem with an additional linear inequality
Journal Article
·
Thu Mar 15 00:00:00 EDT 2018
· Computational and Applied Mathematics
·
OSTI ID:22769384
Polynomial preconditioning for conjugate gradient methods
Technical Report
·
Mon Nov 30 23:00:00 EST 1987
·
OSTI ID:5576336