Two modified DY conjugate gradient methods for unconstrained optimization problems

作者:

Highlights:

摘要

In this paper, we study the unconstrained optimization problems, and the two modified DY conjugate gradient methods (DDY1 method and DDY2 method) are proposed under the DY conjugate gradient method. By using the standard Wolfe line search, we prove the global convergence of the two methods. The search direction of DDY1 method is descent with the standard Wolfe line search. The search direction generated by the DDY2 method is sufficient descent, in which the property does not depend on any line search. Preliminary numerical results show that the two methods are effective.

论文关键词:Unconstrained optimization problem,Conjugate gradient method,Standard Wolfe line search,Global convergence,Sufficient descent property

论文评审过程:Received 1 August 2019, Revised 17 November 2019, Accepted 22 December 2019, Available online 13 January 2020, Version of Record 13 January 2020.

论文官网地址:https://doi.org/10.1016/j.amc.2019.125004