TRUST REGIONS BASED ON CONIC FUNCTIONS IN LINEAR AND NONLINEAR-PROGRAMMING

Authors
NAZARETH JL
Citation
Jl. Nazareth, TRUST REGIONS BASED ON CONIC FUNCTIONS IN LINEAR AND NONLINEAR-PROGRAMMING, Numerical linear algebra with applications, 2(3), 1995, pp. 235-241
Citations number
14
Categorie Soggetti
Mathematics, General",Mathematics,Mathematics
Journal title
Numerical linear algebra with applications → ACNP
ISSN journal
10705325
Volume
2
Issue
3
Year of publication
1995
Pages
235 - 241
Database
ISI
SICI code
1070-5325(1995)2:3<235:TRBOCF>2.0.ZU;2-U
Abstract
An optimization method is developed based on ellipsoidal trust regions that are defined by conic functions. It provides a powerful unifying theory from which can be derived a variety of interesting and potentia lly useful optimization algorithms, in particular, conjugate-gradient- like algorithms for nonlinear minimization and Karmarkar-like interior -point algorithms for linear programming.