计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (11): 64-65.

• 学术探讨 • 上一篇    下一篇

一种非标准共轭梯度法的收敛性及数值模拟

周光明   

  1. 湘潭大学
  • 收稿日期:2006-10-28 修回日期:1900-01-01 出版日期:2007-04-11 发布日期:2007-04-11
  • 通讯作者: 周光明

Convergence and Numerical Simulation on A Non-standard Conjugate Gradient Algorithm

Guang-Ming zhou   

  • Received:2006-10-28 Revised:1900-01-01 Online:2007-04-11 Published:2007-04-11
  • Contact: Guang-Ming zhou

摘要: 本文对无约束优化问题提出了一种新的非标准共轭梯度算法,该算法的搜索方向类似于曲线搜索算法的方向。证明了新算法的全局收敛性,并通过数值模拟验证了该算法是有效的和快速的。

关键词: 无约束优化, 共轭梯度算法, 收敛性, 数值模拟

Abstract: In this paper, a new non-standard conjugate gradient algorithm is proposed, which is dedicated to solving unconstrained optimization problem. Search direction of the new algorithm is similar to direction of curve search method. Global convergence of the algorithm is proved. Numerical simulations show that the new algorithm is efficient and fast.

Key words: unconstrained optimization, conjugate gradient algorithm, convergence, numerical simulation