摘要
对于无约束最优化问题，本文基于局部收敛的Polak－Ribiere方法和全局收敛的Fletcher－Reeves方法，提出了一个新的共轭梯度算法，并采用精确线性搜索方法，得到了全局收敛的好性质。
In This Paper, I Have Presented A New Algorithm, Based on Polak-Ribiere Algorithm with Local Convergence and Fletcher-Reeves Algorithm with Globally Convergence.In Our Algorithm, Acurite Line Search Be Needed. It is Proved That the New Algorthm is Globally Convergent.
关键词
无约束最优化
共轭梯度算法
全局收敛
最佳化
Unconstrained Optimization, Conjugate Gradient Algorithm, Line search, Globally Convergence
