A new self-adaptive trust region method is presented.The trust radius in the new method is automatically determined with second order information,and this new strategy does not make computation complicated.Under general conditions,the local and global convergence results of the new method are proved.Numerical experiments show that the new method is very efficient.
【基金】
北京工业人学博士启动基金;
国家自然科学基金(10231060)
【更新日期】
2006-11-09
【分类号】
O221
【正文快照】
1问题的提出考虑无约束优化问题minf(x),(l)其中f:R“*R是一个二次连续可微函数。无约束优化信赖域方法最先由Powell[ll提出。从那时起,信赖域方法因为其强收敛性,强适性和稳定性等优点引起许多研究者的兴趣。信赖域方法是一种迭代方法,每次迭代,解信赖域子问题mino万d+ d