Article (Scientific journals)
A note on trust-region radius update
Walmag, J. M. B.; Delhez, Eric
2005In SIAM Journal on Optimization, 16 (2), p. 548-562
Peer Reviewed verified by ORBi
 

Files


Full Text
SIAM.pdf
Publisher postprint (320.52 kB)
Request a copy

All documents in ORBi are protected by a user license.

Send to



Details



Keywords :
nonlinear optimization; trust region; radius update; unconstrained optimization
Abstract :
[en] In classical trust-region optimization algorithms, the radius of the trust region is reduced, kept constant, or enlarged after, respectively, unsuccessful, successful, and very successful iterations. We propose here to re. ne the empirical rules used for this update by the definition of a new set of iterations that we call "too successful iterations." At such iterations, a large reduction of the objective function is obtained despite a crude local approximation of the objective function; the trust region is thus kept nearly constant instead of being enlarged. The new update rules preserve the strong convergence property of traditional trust-region methods. They can also be generalized to define a self-adaptive trust-region algorithm along the lines introduced by Hei [J. Comput. Math., 21 (2003), pp. 229-236]. Numerical experiments carried out on 70 unconstrained problems from the CUTEr collection demonstrate the positive impact of the modified update strategy on the efficiency and robustness of quasi-Newton variants of a trust-region solver, when BFGS or SR1 updates of the approximation of the Hessian matrix are carried at all iterations.
Disciplines :
Mathematics
Author, co-author :
Walmag, J. M. B.
Delhez, Eric ;  Université de Liège - ULiège > Département d'aérospatiale et mécanique > Mathématiques générales
Language :
English
Title :
A note on trust-region radius update
Publication date :
2005
Journal title :
SIAM Journal on Optimization
ISSN :
1052-6234
eISSN :
1095-7189
Publisher :
Society for Industrial & Applied Mathematics, Philadelphia, United States - Pennsylvania
Volume :
16
Issue :
2
Pages :
548-562
Peer reviewed :
Peer Reviewed verified by ORBi
Available on ORBi :
since 26 November 2009

Statistics


Number of views
80 (9 by ULiège)
Number of downloads
5 (4 by ULiège)

Scopus citations®
 
17
Scopus citations®
without self-citations
17
OpenCitations
 
12

Bibliography


Similar publications



Contact ORBi