Search: onr:"swepub:oai:DiVA.org:liu-143315" >
A dense initializat...
A dense initialization for limited-memory quasi-Newton methods
-
- Brust, Johannes (author)
- University of California, Merced, CA, USA
-
- Burdakov, Oleg, 1953- (author)
- Linköpings universitet,Optimeringslära,Tekniska fakulteten
-
- Erway, Jennifer B. (author)
- Wake Forest University, Winston-Salem, NC, USA,Department of Mathematics
-
show more...
-
- Marcia, Roummel F. (author)
- University of California, Merced, CA, USA,Applied Mathematics
-
show less...
-
(creator_code:org_t)
- 2019-05-29
- 2019
- English.
-
In: Computational Optimization and Applications. - : Springer. - 0926-6003 .- 1573-2894. ; 74:1, s. 121-142
- Related links:
-
https://arxiv.org/ab...
-
show more...
-
https://liu.diva-por... (primary) (Raw object)
-
http://arxiv.org/pdf...
-
https://urn.kb.se/re...
-
https://doi.org/10.1...
-
show less...
Abstract
Subject headings
Close
- We consider a family of dense initializations for limited-memory quasi-Newton methods. The proposed initialization exploits an eigendecomposition-based separation of the full space into two complementary subspaces, assigning a different initialization parameter to each subspace. This family of dense initializations is proposed in the context of a limited-memory Broyden–Fletcher–Goldfarb–Shanno (L-BFGS) trust-region method that makes use of a shape-changing norm to define each subproblem. As with L-BFGS methods that traditionally use diagonal initialization, the dense initialization and the sequence of generated quasi-Newton matrices are never explicitly formed. Numerical experiments on the CUTEst test set suggest that this initialization together with the shape-changing trust-region method outperforms other L-BFGS methods for solving general nonconvex unconstrained optimization problems. While this dense initialization is proposed in the context of a special trust-region method, it has broad applications for more general quasi-Newton trust-region and line search methods. In fact, this initialization is suitable for use with any quasi-Newton update that admits a compact representation and, in particular, any member of the Broyden class of updates.
Subject headings
- NATURVETENSKAP -- Matematik -- Beräkningsmatematik (hsv//swe)
- NATURAL SCIENCES -- Mathematics -- Computational Mathematics (hsv//eng)
Keyword
- Large-scale nonlinear optimization
- limited-memory quasi-Newton methods
- trust-region methods
- quasi-Newton matrices
- shape-changing norm.
Publication and Content Type
- vet (subject category)
- art (subject category)
Find in a library
To the university's database