新线搜索条件下的Dai-Yuan型共轭梯度法

Dai-Yuan Type Conjugate Gradient Method Under a New Line Search

对非线性无约束优化问题,文章构造一个新的算法,将一种新型线搜索方法与经典算法Dai-Yuan型共轭梯度法相结合.证明搜索方向是充分下降的,且算法具有全局收敛性.初步的数值实验结果表明,新线搜索下的Dai-Yuan型共轭梯度法是有效的.

For nonlinear unconstrained optimization problem,this paper constructs a new algorithm,which combines a new line search method and classical Dai-Yuan type conjugate gradient method.It is proved that the new direction has sufficient descent property and the algorithm is global convergent.Preliminary numeri⁃cal test results show that the Dai-Yuan type conjugate gradient method is effective under the new line search.