摘要
将最速下降法与Newton法有机地结合起来,构造了无约束优化问题的一种组合迭代算法,并证明了算法的全局收敛性.该组合算法既继承了Newton法在极小点附近的快速收敛性,又解决了最速下降法难以求解的问题.
A hybrid iterative algorithm for unconstrained optimization problems is formulated by means of combining organically the steepest decent method and Newton method.This hybrid algorithm not only inherits the merit of Newton method that there is fast convergence at the adjacency of the minimum but also overcomes the difficulty happened with the latter method in problem solution.
出处
《兰州理工大学学报》
CAS
北大核心
2004年第5期143-145,共3页
Journal of Lanzhou University of Technology
基金
陕西省教委专项资助项目(01Jk057)
关键词
最速下降法
NEWTON法
全局收敛
组合迭代法
steepest decent method
Newton method
global convergence
hybrid iterative method
