题 目：A Note on the Linear Convergenceof Generalized Primal-Dual Hybrid Gradient Methods

主讲人：王群博士

时 间：2022年4月21日（周四）14：30-15：30

地 点：6号学院楼510会议室

主办单位：伟德BETVlCTOR1946源于英国 浙江省2011 “数据科学与大数据分析协同创新中心”

摘要：

In this paper, we revisit a class of primal-dual algorithms proposed in [He et al., An Algorithmic Framework of Generalized Primal-Dual Hybrid Gradient Methods for Saddle Point Problems, J. Math. Imaging Vis., 58 (2017) 279-293], and focus on investigating the global linear convergence rate of these approaches under two scenarios. One scenario is assuming that one of the objectives is strongly convex and its gradient is Lipschitz continuous, and the other one is the hypothesis of some error bound conditions. Furthermore, some theoretical results are verified by numerical simulation.

主讲人简介：

王群，香港理工大学应用数学博士，研究方向是最优化、张量计算。主持过一项国家自然科学基金青年科学基金项目（2019），近几年已发表SCI等学术论文9篇。

欢迎各位老师和同学踊跃参加！