Volume 51, Number 1, January-March 2017
|Page(s)||67 - 77|
|Published online||12 December 2016|
Some modified Yabe–Takano conjugate gradient methods with sufficient descent condition
1 School of Mathematics and Information science, Beifang
University of Nationalities, No. 204 North Wenchang Rd,Yinchuan, Ningxia 750021, P.R. China.
2 Network Information Technology Center, Beifang University of Nationalities, No. 204 North Wenchang Rd, Yinchuan, Ningxia 750021, P.R. China.
3 Department of Mathematics and Applied Mathematics, Huaihua University, No. 612 Yingfeng East Road, Huaihua, Hunan 418008, P.R. China.
Accepted: 2 April 2016
Descent condition is a crucial factor to establish the global convergence of nonlinear conjugate gradient method. In this paper, we propose some modified Yabe–Takano conjugate gradient methods, in which the corresponding search directions always satisfy the sufficient descent property independently of the convexity of the objective function. Differently from the existent methods, a new update strategy in constructing the search direction is proposed to establish the global convergence of the presented methods for the general nonconvex objective function. Numerical results illustrate that our methods can efficiently solve the test problems and therefore is promising.
Mathematics Subject Classification: 49M37 / 65K05 / 90C53
Key words: Yabe–Takano conjugate gradient method / global convergence / sufficient descent condition / conjugacy condition / numerical comparison
© EDP Sciences, ROADEF, SMAI 2016
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.