A class of hybrid conjugate gradient projection methods for equality constrained optimization problem
Zhao Hai-yan
Abstract
Zhao Hai-yan
Abstract
In this paper,a class of hybrid conjugate gradient methods were proposed,and they were applied to solve the linear equality constrained optimization problem combined with the Rosen gradient projection method.This new hybrid conjugate gradient projection method was effectivly improved the convergence rate of the Rosen gradient projection method,and the global convergence of the algorithm was proved under the Wolfe line search conditions.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
In this paper,a class of hybrid conjugate gradient methods were proposed,and they were applied to solve the linear equality constrained optimization problem combined with the Rosen gradient projection method.This new hybrid conjugate gradient projection method was effectivly improved the convergence rate of the Rosen gradient projection method,and the global convergence of the algorithm was proved under the Wolfe line search conditions.
Key concepts: Conjugate gradient method, Derivation of the conjugate gradient method, Proximal Gradient Methods, Nonlinear conjugate gradient method, Conjugate residual method, Gradient method, Mathematics, Projection (relational algebra)