| 两类新的变参数下降算法及收敛性 |
| |
| 引用本文： | 刘金魁,杜祥林,王开荣.两类新的变参数下降算法及收敛性[J].应用数学学报,2010,33(2). |
| |
| 作者姓名： | 刘金魁 杜祥林 王开荣 |
| |
| 作者单位： | 1. 重庆三峡学院数学与计算科学学院,重庆,404000
2. 重庆大学数理学院,重庆,400030 |
| |
| 摘 要： | 本文对求解无约束优化问题给出两类新的变参数下降算法.在Wolfe线搜索下无需给定充分下降条件,即可证明它们的全局收敛性.大量数值试验表明它们是非常有效的和稳定的,能够广泛用于科学计算.
|
| 关 键 词： | 无约束优化 下降算法 Wolfe线搜索 充分下降性 全局收敛性 |
Convergence of Descent Methods with Variable Parameters |
| |
| LIU JINKUI,DU XIANGLIN,WANG KAIRONG.Convergence of Descent Methods with Variable Parameters[J].Acta Mathematicae Applicatae Sinica,2010,33(2). |
| |
| Authors: | LIU JINKUI DU XIANGLIN WANG KAIRONG |
| |
| Abstract: | In this paper, two classes of new descent methods with variable parameters are proposed to solve unconstrained optimization problems. And, under Wolfe line search conditions, we proved their global convergence without the given sufficient descent condition. Many numerical experiments show that they are efficient and robust, and they can be widely used in scientific computation. |
| |
| Keywords: | unconstrained optimization descent algorithm Wolfe line search sufficient descent property global convergence |
| 本文献已被 万方数据 等数据库收录！ |
|
