A modified conjugate gradient algorithm with cyclic Barzilai-Borwein steplength for unconstrained optimization (Q421813): Difference between revisions
From MaRDI portal
Created a new Item |
Changed an Item |
||
Property / author | |||
Property / author: Yun-hai Xiao / rank | |||
Normal rank | |||
Property / review text | |||
From the authors' abstract: For solving large-scale unconstrained minimization problems, the nonlinear conjugate gradient method is welcome due to its simplicity, low storage, efficiency and nice convergence properties. Among all the methods in the framework, the conjugate gradient descent algorithm -- CG\_DESCENT is very popular, in which the generated directions descend automatically, and this nice property is independent of any line search used. In this paper, the authors generalize CG\_DESCENT with two Barzilai-Borwein steplength reused cyclically. It is shown that the resulting algorithm owns an attractive sufficient descent property and converges globally under some mild conditions. The proposed algorithm is tested by using a large set of unconstrained problems with high dimensions in CUTEr library. The numerical comparisons with the state-of-the-art algorithm CG\_DESCENT illustrate that the proposed method is effective, competitive, and promising. | |||
Property / review text: From the authors' abstract: For solving large-scale unconstrained minimization problems, the nonlinear conjugate gradient method is welcome due to its simplicity, low storage, efficiency and nice convergence properties. Among all the methods in the framework, the conjugate gradient descent algorithm -- CG\_DESCENT is very popular, in which the generated directions descend automatically, and this nice property is independent of any line search used. In this paper, the authors generalize CG\_DESCENT with two Barzilai-Borwein steplength reused cyclically. It is shown that the resulting algorithm owns an attractive sufficient descent property and converges globally under some mild conditions. The proposed algorithm is tested by using a large set of unconstrained problems with high dimensions in CUTEr library. The numerical comparisons with the state-of-the-art algorithm CG\_DESCENT illustrate that the proposed method is effective, competitive, and promising. / rank | |||
Normal rank | |||
Property / reviewed by | |||
Property / reviewed by: Nada I. Djuranović-Miličić / rank | |||
Normal rank | |||
Property / Mathematics Subject Classification ID | |||
Property / Mathematics Subject Classification ID: 65K05 / rank | |||
Normal rank | |||
Property / Mathematics Subject Classification ID | |||
Property / Mathematics Subject Classification ID: 90C06 / rank | |||
Normal rank | |||
Property / Mathematics Subject Classification ID | |||
Property / Mathematics Subject Classification ID: 90C30 / rank | |||
Normal rank | |||
Property / zbMATH DE Number | |||
Property / zbMATH DE Number: 6035220 / rank | |||
Normal rank | |||
Property / zbMATH Keywords | |||
Wolfe condition | |||
Property / zbMATH Keywords: Wolfe condition / rank | |||
Normal rank | |||
Property / zbMATH Keywords | |||
cuter library | |||
Property / zbMATH Keywords: cuter library / rank | |||
Normal rank | |||
Property / zbMATH Keywords | |||
global convergence | |||
Property / zbMATH Keywords: global convergence / rank | |||
Normal rank | |||
Property / zbMATH Keywords | |||
numerical examples | |||
Property / zbMATH Keywords: numerical examples / rank | |||
Normal rank | |||
Property / zbMATH Keywords | |||
large-scale unconstrained minimization problems | |||
Property / zbMATH Keywords: large-scale unconstrained minimization problems / rank | |||
Normal rank | |||
Property / zbMATH Keywords | |||
nonlinear conjugate gradient method | |||
Property / zbMATH Keywords: nonlinear conjugate gradient method / rank | |||
Normal rank | |||
Property / zbMATH Keywords | |||
conjugate gradient descent algorithm | |||
Property / zbMATH Keywords: conjugate gradient descent algorithm / rank | |||
Normal rank | |||
Property / zbMATH Keywords | |||
Barzilai-Borwein steplength | |||
Property / zbMATH Keywords: Barzilai-Borwein steplength / rank | |||
Normal rank | |||
Property / zbMATH Keywords | |||
algorithm | |||
Property / zbMATH Keywords: algorithm / rank | |||
Normal rank |
Revision as of 21:51, 29 June 2023
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | A modified conjugate gradient algorithm with cyclic Barzilai-Borwein steplength for unconstrained optimization |
scientific article |
Statements
A modified conjugate gradient algorithm with cyclic Barzilai-Borwein steplength for unconstrained optimization (English)
0 references
14 May 2012
0 references
From the authors' abstract: For solving large-scale unconstrained minimization problems, the nonlinear conjugate gradient method is welcome due to its simplicity, low storage, efficiency and nice convergence properties. Among all the methods in the framework, the conjugate gradient descent algorithm -- CG\_DESCENT is very popular, in which the generated directions descend automatically, and this nice property is independent of any line search used. In this paper, the authors generalize CG\_DESCENT with two Barzilai-Borwein steplength reused cyclically. It is shown that the resulting algorithm owns an attractive sufficient descent property and converges globally under some mild conditions. The proposed algorithm is tested by using a large set of unconstrained problems with high dimensions in CUTEr library. The numerical comparisons with the state-of-the-art algorithm CG\_DESCENT illustrate that the proposed method is effective, competitive, and promising.
0 references
Wolfe condition
0 references
cuter library
0 references
global convergence
0 references
numerical examples
0 references
large-scale unconstrained minimization problems
0 references
nonlinear conjugate gradient method
0 references
conjugate gradient descent algorithm
0 references
Barzilai-Borwein steplength
0 references
algorithm
0 references