arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2304.02465math.OC
收藏

求解可分离凸优化问题:更快的预测-校正框架

Solving separable convex optimization problems: Faster prediction-correction framework

Tao Zhang, Yong Xia, Shiru Li

AI总结:

本文提出一个更快的预测-校正框架,用于求解可分离凸优化问题,在非遍历和逐点意义上分别达到 $O(1/t)$ 和 $O(1/t^2)$ 的收敛速率,并给出三种基于该框架的算法,分别处理两块、多块和极小极大问题。

摘要:

何和袁的预测-校正框架 [SIAM J. Numer. Anal. 50: 700-709, 2012] 能够以 $O(1/t)$ 的速率($t$ 表示迭代次数)在遍历(迭代的平均)和逐点意义上提供求解可分离凸优化问题的收敛算法。本文提出一个更快的预测-校正框架,在非遍历意义(最后一次迭代)上达到 $O(1/t)$ 的速率,在逐点意义上达到 $O(1/t^2)$ 的速率。基于该更快的预测-校正框架,我们给出三种更快的算法,它们在非遍历意义上对原始-对偶间隙达到 $O(1/t)$,在逐点意义上达到 $O(1/t^2)$。第一种算法在求解具有等式线性约束的两块可分离凸优化问题时,对偶变量更新两次。第二种算法以 Gauss-Seidel 方式求解具有线性等式约束的多块可分离凸优化问题。第三种算法以更大的步长求解极小极大问题。

英文摘要:

He and Yuan's prediction-correction framework [SIAM J. Numer. Anal. 50: 700-709, 2012] is able to provide convergent algorithms for solving separable convex optimization problems at a rate of $O(1/t)$ ($t$ represents iteration times) in both ergodic (the average of iteration) and pointwise senses. This paper presents a faster prediction-correction framework at a rate of $O(1/t)$ in the non-ergodic sense (the last iteration) and $O(1/t^2)$ in the pointwise sense. Based the faster prediction-correction framework, we give three faster algorithms which enjoy $O(1/t)$ in the non-ergodic sense of primal-dual gap and $O(1/t^2)$ in the pointwise sense. The first algorithm updates dual variable twice when solving two-block separable convex optimization with equality linear constraints. The second algorithm solves multi-block separable convex optimization problems with linear equality constraints in Gauss-Seidel way. The third algorithm solves minmax problems with larger step sizes.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑