|
| 摘要: |
| 本文研究了大规模无约束优化问题,利用BFGS逼近搜索方向,提出了两种关于HSDY方法的自适应共轭梯度算法(HSDY1和HSDY2).新算法具有充分下降性和全局收敛性.数值实验表明,新方法比HSDY的计算性能更优. |
| 关键词: 共轭梯度法 BFGS算法 充分下降性 全局收敛性 |
| DOI： |
| 分类号:O224 |
| 基金项目:重庆市教委科学技术研究项目基金资助(KJQN 201800520). |
|
| HSDY MIXED CONJUGATE GRADIENT WAS MODIFIED BASED ON BFGS |
|
ZHU Zi-chang,LIU Li-ping
|
| Abstract: |
| In this paper, we are concerned with large-scale unconstrained optimization problems. We use BFGS to approximate the search direction, and propose two adaptive conjugate gradient algorithms(HSDY1 and HSDY2) for the HSDY method. We prove that the new algorithms have sufficient descent and global convergence. Numerical experiments are reported to show the computational performance of proposed algorithms are better than HSDY. |
| Key words: conjugate gradient methods BFGS algorithm sufficient descensibility global convergence |
