摘要
提出了一种非单调自适应不定折线信赖域算法,当B_k不正定时,运用Bunch-Parlett分解产生搜索路径来确定下降方向.与一般的非单调信赖域算法相比,新算法根据实际下降量与预估计下降量的比值按照变化的速率对信赖域半径进行调整,在研究方法上具有一定的创新.
A nonmonotone adaptive indefinite dogleg trust region algorithm is presented in this paper. WhenBk is indefinite, the descendent direction is generated mainly by employing Bunch-Parlett factorization. Compared with traditional nonmonotone TR algorithm, the trust region radius is updated at a variable rate according to the ratio of the actual reduction to the predicted reduction of the objective function in the new algorithm that has certain innovation in research mathods.
出处
《数学的实践与认识》
北大核心
2015年第10期268-274,共7页
Mathematics in Practice and Theory
基金
国家自然科学基金(51275333)
太原科技大学教研项目(201240)
关键词
非单调技术
不定折线
自适应信赖域算法
nonmonotone technique
indefinite Dogleg
adaptive trust region algorithm
