当前位置: X-MOL 学术SIAM J. Optim. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
On High-Order Multilevel Optimization Strategies
SIAM Journal on Optimization ( IF 2.6 ) Pub Date : 2021-01-20 , DOI: 10.1137/19m1255355
Henri Calandra , Serge Gratton , Elisa Riccietti , Xavier Vasseur

SIAM Journal on Optimization, Volume 31, Issue 1, Page 307-330, January 2021.
We propose a new family of multilevel methods for unconstrained minimization. The resulting strategies are multilevel extensions of high-order optimization methods based on $q$th-order Taylor models (with $q\geq 1$) that have been recently proposed in the literature. The use of high-order models, while decreasing the worst-case complexity bound, makes these methods computationally more expensive. Hence, to counteract this effect, we propose a multilevel strategy that exploits a hierarchy of problems of decreasing dimension, still approximating the original one, to reduce the global cost of the step computation. A theoretical analysis of the family of methods is proposed. Specifically, local and global convergence results are proved, and a worst-case complexity bound to reach first-order stationary points is also derived. A multilevel version of the well-known adaptive regularization by cubics (corresponding to $q=2$ in our setting) has been implemented, as well as a multilevel third-order method ($q=3$). Numerical experiments clearly highlight the relevance of the new multilevel approaches leading to considerable computational savings compared to their one-level counterparts.


中文翻译:

高阶多级优化策略

SIAM优化杂志,第31卷,第1期,第307-330页,2021年1月。
我们提出了一个新的多级方法家族,用于无约束最小化。所得策略是最近在文献中提出的基于$ q $阶泰勒模型($ q \ geq 1 $)的高阶优化方法的多级扩展。高阶模型的使用在降低最坏情况下的复杂性界限的同时,使这些方法的计算成本更高。因此,为了抵消这种影响,我们提出了一种多层次的策略,该策略利用了递减维数的问题的层次结构,仍然与原始维数近似,以减少步骤计算的整体成本。提出了一系列方法的理论分析。具体来说,证明了局部和全局收敛的结果,并且还得出了到达一阶平稳点的最坏情况下的复杂度。已经实现了三次方的自适应自适应正则化的多级版本(对应于我们设置中的$ q = 2 $),以及多级三阶方法($ q = 3 $)。数值实验清楚地表明了新的多级方法的相关性,与一级方法相比,可节省大量计算量。
更新日期:2021-03-21
down
wechat
bug