首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于锥模型的非单调自适应信赖域算法
引用本文:冯琳,段复建.基于锥模型的非单调自适应信赖域算法[J].山西大学学报(自然科学版),2011,34(4):580-586.
作者姓名:冯琳  段复建
作者单位:桂林电子科技大学数学与计算科学学院,广西桂林,541004
基金项目:国家自然科学基金(1086100511061011); 广西自然科学基金(0991238)
摘    要:基于锥模型,结合提出的新的自适应技术,建立了一个求解无约束最优化问题的非单调自适应信赖域算法.当试探步不被接受时,采用非单调线搜索,减少了计算量.充分利用包含当前迭代点信息的新的自适应策略调节信赖域半径.在适当的条件下,证明了算法的全局收敛性和Q-二阶收敛性.数值试验表明该算法是有效的.

关 键 词:无约束最优化  锥模型  非单调信赖域方法  自适应  非单调线搜索  全局收敛性

A Non-monotone Self-adaptive Trust Region Method Based on the Conic Model
FENG Lin , DUAN Fu-jian.A Non-monotone Self-adaptive Trust Region Method Based on the Conic Model[J].Journal of Shanxi University (Natural Science Edition),2011,34(4):580-586.
Authors:FENG Lin  DUAN Fu-jian
Institution:FENG Lin,DUAN Fu-jian(School of Mathematics and Computational Science,Guilin University of Electronic Technology,Guilin 541004,China)
Abstract:Based on the conic model and a new self-adaptive technique,a non-monotone self-adaptive trust region algorithm for unconstrained optimization problems is got.When the trial step is not accepted,a non-monotonic line search is employed to reduce the cost.And the trust region radius is adjusted with a new self-adaptive adjustment strategy which makes full use of the information at the current point.Under mild conditions,the global convergence and Q-quadratic convergence of the new method are proved.Nemerical r...
Keywords:unconstrained optimization  conic model  non-monotonic trust-region method  self-adaptive  non-monotonic line search  global convergence  
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号