A one-parameter class of three-term conjugate gradient methods with an adaptive parameter choice. (1st November 2020)

• Record Type: Journal Article
• Authors: Yao, Shengwei; Ning, Liangshuo; Tu, Huonian; Xu, Jieqiong
• Is Part Of: Optimization methods and software. Volume 35:Number 6(2020)
• Page Start: 1051
• Page End: 1064
• Publication Date: 2020-11-01
• Subjects: Unconstrained optimization -- conjugate gradient method -- global convergence -- conjugacy condition; Mathematical optimization -- Periodicals; Algorithms -- Periodicals; 519.7
• DOI: 10.1080/10556788.2018.1510926
• Languages: English
• ISSNs: 1055-6788

Abstract

ABSTRACT: In this paper, a one-parameter class of three-term conjugate gradient methods is proposed for solving large-scale unconstrained optimization problem. Based on eigenvalue study, an adaptive choice for the embedded parameter and the corresponding three-term conjugate gradient method is obtained. In addition, the convergence properties of the obtained method are studied for both convex and general function. Numerical experiments show that the proposed method has efficient and robust performance.