| 摘要: |
| 提出一种基于BFGS更新的Levenberg-Marquardt算法,该算法不仅具有全局收敛性和二次收敛速度,而且可以更有效地求解大规模优化问题.数值实验表明,该算法在求解大规模绝对值方程问题方面也是有效的. |
| 关键词: 广义Levenberg-Marquardt算法 BFGS更新 全局收敛性 绝对值方程 |
| DOI：10.13656/j.cnki.gxkx.20161121.003 |
| 投稿时间：2016-07-25 |
| 基金项目:国家自然科学基金资助项目(11161003)和广西杰出青年科学基金项目(2015GXNSFGA139001)资助。 |
|
| A Generalized BFGS Levenberg-Marquardt Algorithm |
|
JI Xianglin, WEI Zengxin
|
| (College of Mathematics and Information Science, Guangxi University, Nanning, Guangxi, 530004, China) |
| Abstract: |
| This paper proposes a modified Levenberg-Marquardt algorithm with BFGS update formula.Our algorithm converges globally to an optimal solution and the convergence rate is quadratic.Moreover,it has better efficiency for solving large-scale problems.Numerical results show that this algorithm is promising for solving large-scale absolute value equations problems. |
| Key words: generalized Levenberg-Marquardt algorithm BFGS update global convergence absolute value equations |
