1973
DOI: 10.1093/imamat/11.1.73
|View full text |Cite
|
Sign up to set email alerts
|

Rank-one and Rank-two Corrections to Positive Definite Matrices Expressed in Product Form

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1

Citation Types

0
19
0

Year Published

1979
1979
2011
2011

Publication Types

Select...
4
3
1

Relationship

0
8

Authors

Journals

citations
Cited by 53 publications
(19 citation statements)
references
References 0 publications
0
19
0
Order By: Relevance
“…All the algorithms of this section have already been suggested by Goldfarb [14] using the Brodlie, Gourlay, and Greenstadt [1] factored form of the BFGS and DFP updates and the orthogonal decompositon update ideas of Gill, Golub, Murray, and Saunders [12]. Our purpose is to point out that they follow very naturally from the preceding derivations.…”
Section: Updating Cholesky Factorsmentioning
confidence: 95%
“…All the algorithms of this section have already been suggested by Goldfarb [14] using the Brodlie, Gourlay, and Greenstadt [1] factored form of the BFGS and DFP updates and the orthogonal decompositon update ideas of Gill, Golub, Murray, and Saunders [12]. Our purpose is to point out that they follow very naturally from the preceding derivations.…”
Section: Updating Cholesky Factorsmentioning
confidence: 95%
“…We can now calculate Sti) and yCi) from eqs. (25) and (26) and proceed with the evaluation of the matrix update that satisfies eq. (27).…”
Section: Matrix Updating Iterative Methodsmentioning
confidence: 99%
“…Evaluate the correction to the coefficient matrix. In the BFGS method the updated matrix can be expressed in product form (see [25]):…”
Section: Matrix Updating Iterative Methodsmentioning
confidence: 99%
“…This leaves the method for specification of {Bk} to be chosen. Many choices are available; most commonly Bk+X is obtained either by a Broyden update of Bk, (yk -P-ksk)sI (1)(2)(3)(4) Bk+x=Bk + T SkSk withyk = F(xk + sk) -F(xk) and sk = xk+x -xk, or from the exact Jacobian or a finite difference approximant, or by taking Bk+l = Bk [7]. The Broyden update is the unique solution to the Frobenius norm minimization problem:…”
mentioning
confidence: 99%
“…The algorithm outlined here is representative of a general technique for producing inexpensive updates of matrix factorizations. This is not the same as techniques for obtaining matrix updates in factored form as in the case of rank 1 and rank 2 updates [1], [12]. Here, the update is defined implicitly by the updated factors [5].…”
mentioning
confidence: 99%