保守修正BFGS算法及全局收敛性 |
| |
引用本文: | 张继伟.保守修正BFGS算法及全局收敛性[J].厦门水产学院学报,2010(3):228-233. |
| |
作者姓名: | 张继伟 |
| |
作者单位: | 集美大学理学院,福建厦门361021 |
| |
基金项目: | 国家自然科学基金资助项目(10471036);集美大学科研基金资助项目(C60657) |
| |
摘 要: | 基于新拟牛顿方程,提出一类保守修正BFGS算法.该算法的特点是:即使当目标函数是非凸函数时,该算法仍然是全局收敛的.在适当的条件下,该算法具有局部超线性收敛性.初步的数值实验表明,该算法是有效的.
|
关 键 词: | BFGS算法 全局收敛性 超线性收敛性 |
A Cautious BFGS-type Method and Its Global Convergence |
| |
Authors: | ZHANG Ji-wei |
| |
Institution: | ZHANG Ji-wei (School of Sciences, Jimei University, Xiamen 361021, China) |
| |
Abstract: | A cautious BFGS-type algorithm based on a new secant equation was this version of BFGS algorithm was that this algorithm possessed global convergence even when the objective function was nonconvex. In the proper conditions, the cautious BFGS-type algorithm was superlinearly convergent. The numerical experiments showed that the cautious BFGS-type algorithm was promising. |
| |
Keywords: | BFGS algorithm global convergence superlinear convergence |
本文献已被 维普 等数据库收录! |
|