Huber approximation for the non-linear ℓ1 problem

Thumbnail
Author
Pınar, Mustafa Ç.
Hartmann, W. M.
Date
2006
Source Title
European Journal of Operational Research
Print ISSN
0377-2217
Publisher
Elsevier
Volume
169
Issue
3
Pages
1096 - 1107
Language
English
Type
Conference Paper
Item Usage Stats
78
views
73
downloads
Metadata
Show full item record
Abstract
The smooth Huber approximation to the non-linear ℓ1 problem was proposed by Tishler and Zang (1982), and further developed in Yang (1995). In the present paper, we use the ideas of Gould (1989) to give a new algorithm with rate of convergence results for the smooth Huber approximation. Results of computational tests are reported.
Keywords
Nonlinear programming
Non-differentiable optimization
Smoothing algorithms
Huber M-estimator
Permalink
http://hdl.handle.net/11693/27253
Published Version (Please cite this version)
http://dx.doi.org/10.1016/j.ejor.2004.10.029
Collections