1994
DOI: 10.1007/bf01183013
|View full text |Cite
|
Sign up to set email alerts
|

A new trust region algorithm for bound constrained minimization

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
74
0
19

Year Published

2001
2001
2014
2014

Publication Types

Select...
8

Relationship

1
7

Authors

Journals

citations
Cited by 116 publications
(93 citation statements)
references
References 8 publications
0
74
0
19
Order By: Relevance
“…Since the product of two strongly semismooth functions in again strongly semismooth, the first statement follows from Lemma 4.1 together with our general smoothness assumption on the mapping f . The remaining statements follow directly from the definition of the B-subdifferential, see Definition 2.3 (note, however, that, usually, (19), (20) contain more elements than those belonging to ∂ B G(x)).…”
Section: Affine-scaling Interior-point Newton Methodsmentioning
confidence: 99%
See 2 more Smart Citations
“…Since the product of two strongly semismooth functions in again strongly semismooth, the first statement follows from Lemma 4.1 together with our general smoothness assumption on the mapping f . The remaining statements follow directly from the definition of the B-subdifferential, see Definition 2.3 (note, however, that, usually, (19), (20) contain more elements than those belonging to ∂ B G(x)).…”
Section: Affine-scaling Interior-point Newton Methodsmentioning
confidence: 99%
“…Taking into account the definition of M (x * ) in (19), (20), it follows after some elementary calculations that the ith column vector A i of A := M (x * ) T is given by…”
Section: Local Convergence Analysismentioning
confidence: 99%
See 1 more Smart Citation
“…This is a Fortran double-precision code for solving nonlinear programming problems, based on augmented Lagrangian [19], trust region [13] and projected gradients combined with a mild active set strategy [6]. It is available at http://www.ime.unicamp.br/ ∼ martinez.…”
Section: A Simple Instance and An Alternative Formulationmentioning
confidence: 99%
“…The minimization of a nonlinear function subject to bounds on the variables has been the subject of intense previous work, along many possible avenues. Major classes of algorithms for bound-constrained problems include the ones based on: active or -active set methods (see, e.g., [1,13,32] and more recently [18] for a short review on active set methods); trust-region methods (see, e.g., [6,7,14,22,24]); interior-point methods (see, e.g., [5,10,19]); line-search projected gradient methods (see, e.g., [2] and the references therein; see also [3,25,35] for a limited memory BFGS method); and filter type methods (see [31]). The approach proposed and analyzed in this paper belongs to the trust-region class but also shares the flavor of projected gradient methods.…”
Section: Introductionmentioning
confidence: 99%