摘要
将Perry-Shanno无记忆拟牛顿方法与一类非单调搜索相结合,给出了一类求解无约束最优化问题的新算法.在目标函数为凸的条件下,证明了该算法的全局收敛性.
In this paper, the Perry-Shanno' s memoryless quasi-Newton method for unconstrained optimization is investigated. Nonmonotone Linesearch procedure is introduced, which is combined with Perry-Shanno' s memoryless quasi-Newton method. Under the convexity assumption on objective function, the globle convergence of this method is proved.
出处
《首都师范大学学报(自然科学版)》
2006年第6期10-14,共5页
Journal of Capital Normal University:Natural Science Edition
基金
北京市教委科研基金(KM200510028019)资助.
关键词
无记忆拟牛顿法
非单调线搜索
全局收敛性
memoryless quasi-Newton method, nonmonotone linesearch, globle convergence.