A SELF-ADAPTIVE TECHNIQUE FORA KIND OF NONLINEAR CONJUGATEGRADIENT METHODS  

作　　者：王丽平 

机构地区：[1]Department of Mathematics

出　　处：《Numerical Mathematics A Journal of Chinese Universities(English Series)》2004年第2期225-232,共8页

摘　　要：Conjugate gradient methods. are a class of important methods for unconstrained optimization, especially when the dimension is large. In 2001, Dai and Liao have proposed a new conjugate condition, based on it two nonlinear conjugate gradient methods are constructed. With trust region idea, this paper gives a self-adaptive technique for the two methods. The numerical results show that this technique works well for the given nonlinear optimization test problems.Conjugate gradient methods are a class of important methods for unconstrained optimization, especially when the dimension is large. In 2001, Dai and Liao have proposed a new conjugate condition, based on it two nonlinear conjugate gradient methods are constructed. With trust region idea, this paper gives a self-adaptive technique for the two methods. The numerical results show that this technique works well for the given nonlinear optimization test problems.

关 键 词：非线性最优化试验问题 非线性共轭梯度法 自适应技术 无约束最优化 下降方向 

分 类 号：O241[理学—计算数学]