What Happens When We Trust a Region That is a Line?
dc.contributor.author | Vicente, Luís N. | en_US |
dc.date.accessioned | 2018-06-18T17:42:17Z | en_US |
dc.date.available | 2018-06-18T17:42:17Z | en_US |
dc.date.issued | 1995-04 | en_US |
dc.date.note | April 1995 | en_US |
dc.description.abstract | Line searches and trust regions are two techniques to globalize nonlinear optimization algorithms. We claim that the trust-region technique has built-in an appropriate regularization of ill-conditioned second-order approximation. The question we ask and then answer in this short paper supports this claim. We force the trust-region technique to act like a line search and we accomplish this by always choosing the step along the quasi-Newton direction. We obtain global first-order convergence as long as the condition number of the second-order approximation is uniformly bounded, a condition that is required in line searches but not in trust regions. | en_US |
dc.format.extent | 8 pp | en_US |
dc.identifier.citation | Vicente, Luís N.. "What Happens When We Trust a Region That is a Line?." (1995) <a href="https://hdl.handle.net/1911/101855">https://hdl.handle.net/1911/101855</a>. | en_US |
dc.identifier.digital | TR95-10 | en_US |
dc.identifier.uri | https://hdl.handle.net/1911/101855 | en_US |
dc.language.iso | eng | en_US |
dc.relation.HasVersion | Investigacao Operacional, 16, (1996), 173-179 | en_US |
dc.title | What Happens When We Trust a Region That is a Line? | en_US |
dc.type | Technical report | en_US |
dc.type.dcmi | Text | en_US |
Files
Original bundle
1 - 1 of 1