An active set truncated Newton method for large-scale bound constrained optimization

被引:2
|
作者
Cheng, Wanyou [1 ]
Chen, Zixin [2 ,3 ]
Li, Dong-hui [4 ]
机构
[1] Dongguan Univ Technol, Coll Comp, Dongguan 523808, Peoples R China
[2] Dongguan Univ Technol, Dongguan 523808, Peoples R China
[3] Dongguan Univ Technol, City Coll, Dongguan 523808, Peoples R China
[4] S China Normal Univ, Sch Math Sci, Guangzhou 510631, Guangdong, Peoples R China
关键词
Bound constrained optimization; Conjugate gradient method; Global convergence; NONMONOTONE LINE SEARCH; ALGORITHM;
D O I
10.1016/j.camwa.2014.01.009
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
An active set truncated Newton method for large-scale bound constrained optimization is proposed. The active sets are guessed by an identification technique. The search direction consists of two parts: some of the components are simply defined; the other components are determined by the truncated Newton method. The method based on a nonmonotone line search technique is shown to be globally convergent. Numerical experiments are presented using bound constrained problems in the CUTEr test problem library. The numerical performance reveals that our method is effective and competitive with the famous algorithm TRON. (C) 2014 Elsevier Ltd. All rights reserved.
引用
收藏
页码:1016 / 1023
页数:8
相关论文
共 50 条