Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/27346
Title: The convergence of equilibrium algorithms with non-monotone line search technique
Authors: Gao, Z
Lam, WHK 
Wong, SC
Yang, H
Keywords: Convex combination algorithm
Equilibrium assignment
Global convergence
Non-monotone line search
Issue Date: 2004
Publisher: Elsevier
Source: Applied mathematics and computation, 2004, v. 148, no. 1, p. 1-13 How to cite?
Journal: Applied mathematics and computation 
Abstract: The slow-convergence characteristics of the regular convex combination algorithm (such as Frank-Wolfe method) are well known particularly when the optimal solution is being reached. In this paper, a new convex combination algorithm with non-monotone line search technique proposed for solving the equilibrium assignment problem, together with the proof of its global convergence. Moreover, it should be pointed out that the conditions which ensure the global convergence of the algorithm proposed in this paper are much milder than those suggested by Bonnans et al.; Grippo et al.; Panier and Tits; Xu et al. [SIAM J. Number. Anal. 29 (4) (1992) 1187; SIAM J. Number. Anal. 23 (4) (1986) 707; SIAM J. Number. Anal. 28 (4) (1991) 1183; Comput. Optimiz. Appl. 18 (2001) 221]. The new algorithm can be viewed as a generalization of the regular convex combination algorithm. Numerical results indicate that the proposed algorithm would lead to a considerable saving both in the number of line searches and in the number of function evaluations.
URI: http://hdl.handle.net/10397/27346
ISSN: 0096-3003
EISSN: 1873-5649
DOI: 10.1016/S0096-3003(02)00821-4
Appears in Collections:Journal/Magazine Article

Access
View full-text via PolyU eLinks SFX Query
Show full item record

SCOPUSTM   
Citations

8
Last Week
0
Last month
0
Citations as of Oct 24, 2017

WEB OF SCIENCETM
Citations

4
Last Week
1
Last month
0
Citations as of Oct 17, 2017

Page view(s)

48
Last Week
6
Last month
Checked on Oct 22, 2017

Google ScholarTM

Check

Altmetric



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.