节点文献
一种基于Wolfe搜索下的新共轭梯度法(英文)
A New Conjugate Gradient Methods based on the Wolfe Search
【摘要】 共轭梯度法在求解非线性最优化问题的一种重要的方法.基于前人的研究成果,提出一种新的梯度公式并将其进行修正,最后在Wolfe搜索下证明该梯度法是全局收敛的,然后通过数值试验显示该算法是有效的.
【Abstract】 The conjugate gradient method plays an important role in solving the nonlinear unconstrained optimization problems.In this paper a new conjugate gradient formula based on the former results is put forward,and it finally turns out to be globally convergent in the condition of Wolfe search.Then some numerical tests have been made to show that the algorithm is effective.
【关键词】 非线性最优化;
共轭梯度法;
全局收敛性;
【Key words】 unconstrained optimization; conjugate gradient method; global convergence;
【Key words】 unconstrained optimization; conjugate gradient method; global convergence;
- 【文献出处】 怀化学院学报 ,Journal of Huaihua University , 编辑部邮箱 ,2009年05期
- 【分类号】O224
- 【下载频次】44