摘要
构造了求解无约束优化问题的新算法,该算法结合了一般的Newton算法的思想和过滤器线性搜索策略,一方面搜索方向由Newton算法产生;另一方面在接受新的迭代点时,采用过滤器线性搜索策略,确定步长,且新算法是全局收敛的.
In this paper, we have constructed a new algorithm for unconstrained optimization. The algorithm combines the idea of Newton method with the strategy of filter-linear-search algorithm. One hand, the searching process of direction is similar to Newton method; on the other hand, when the new point will be accepted, we adopt the strategy of filter-linear-search algorithm. In addition, the new algorithm has global convergence.
出处
《湖南城市学院学报（自然科学版）》
CAS
2010年第3期46-49,共4页
Journal of Hunan City University:Natural Science
基金
湖南省教育厅科研基金资助项目(08C206)
湖南城市学院科研基金资助项目(08C016)
关键词
过滤器算法
无约束优化
线性搜索
filter method
unconstrained optimization
linear search
