A new multi-step gradient method for optimization problem

A new multi-step gradient method for optimization problem

1st Regional Conference on Applied and Engineering Mathematics (RCAEM-I) 2010 organized by Universiti Malaysia Perlis (UniMAP) and co-organized by Universiti Sains Malaysia (USM) & Universiti Kebangsaan Malaysia (UKM), 2nd - 3rd June 2010 at Eastern & Oriental Hotel, Penang.

Saved in:
Bibliographic Details
Main Authors: Mahboubeh, Farid, Wah, June Leong, Malik, Abu Hassan
Other Authors: mahboubeh_farid@yahoo.com
Format: Working Paper
Language:English
Published: Universiti Malaysia Perlis (UniMAP) 2010
Subjects:
Unconstrained optimization
Diagonal updating
Barzilai
Borwein method
Regional Conference on Applied and Engineering Mathematics (RCAEM)
Online Access:http://dspace.unimap.edu.my/xmlui/handle/123456789/10254
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Universiti Malaysia Perlis
Language: English
id my.unimap-10254
record_format dspace
spelling my.unimap-102542010-11-23T01:35:46Z A new multi-step gradient method for optimization problem Mahboubeh, Farid Wah, June Leong Malik, Abu Hassan mahboubeh_farid@yahoo.com Unconstrained optimization Diagonal updating Barzilai Borwein method Regional Conference on Applied and Engineering Mathematics (RCAEM) 1st Regional Conference on Applied and Engineering Mathematics (RCAEM-I) 2010 organized by Universiti Malaysia Perlis (UniMAP) and co-organized by Universiti Sains Malaysia (USM) & Universiti Kebangsaan Malaysia (UKM), 2nd - 3rd June 2010 at Eastern & Oriental Hotel, Penang. In this paper, we propose a new multi-step quasi- Newton method for solving unconstrained optimization problem. Our new gradient method is a in the frame of Barzilai and Borwein method. A multi step method is used to improve an accuracy of Hessian approximation which is in a diagonal matrix form. By incorporating a simple monotone strategy, the global convergence of our new method is achieved. Numerical results show the superiority of proposed method in compare with the BB method. 2010-11-23T01:35:46Z 2010-11-23T01:35:46Z 2010-06-02 Working Paper Vol.3(4), p.294-297 http://hdl.handle.net/123456789/10254 en Proceedings of the 1st Regional Conference on Applied and Engineering Mathematics (RCAEM-I) 2010 Universiti Malaysia Perlis (UniMAP) Institut Matematik Kejuruteraan
institution Universiti Malaysia Perlis
building UniMAP Library
collection Institutional Repository
continent Asia
country Malaysia
content_provider Universiti Malaysia Perlis
content_source UniMAP Library Digital Repository
url_provider http://dspace.unimap.edu.my/
language English
topic Unconstrained optimization
Diagonal updating
Barzilai
Borwein method
Regional Conference on Applied and Engineering Mathematics (RCAEM)
spellingShingle Unconstrained optimization
Diagonal updating
Barzilai
Borwein method
Regional Conference on Applied and Engineering Mathematics (RCAEM)
Mahboubeh, Farid
Wah, June Leong
Malik, Abu Hassan
A new multi-step gradient method for optimization problem
description 1st Regional Conference on Applied and Engineering Mathematics (RCAEM-I) 2010 organized by Universiti Malaysia Perlis (UniMAP) and co-organized by Universiti Sains Malaysia (USM) & Universiti Kebangsaan Malaysia (UKM), 2nd - 3rd June 2010 at Eastern & Oriental Hotel, Penang.
author2 mahboubeh_farid@yahoo.com
author_facet mahboubeh_farid@yahoo.com
Mahboubeh, Farid
Wah, June Leong
Malik, Abu Hassan
format Working Paper
author Mahboubeh, Farid
Wah, June Leong
Malik, Abu Hassan
author_sort Mahboubeh, Farid
title A new multi-step gradient method for optimization problem
title_short A new multi-step gradient method for optimization problem
title_full A new multi-step gradient method for optimization problem
title_fullStr A new multi-step gradient method for optimization problem
title_full_unstemmed A new multi-step gradient method for optimization problem
title_sort new multi-step gradient method for optimization problem
publisher Universiti Malaysia Perlis (UniMAP)
publishDate 2010
url http://dspace.unimap.edu.my/xmlui/handle/123456789/10254
_version_ 1643789795487383552

Similar Items