报告题目：The Balanced Augmented Lagrangian Method for Convex Programming

Abstract: We consider the canonical convex minimization problem with both linear equality and inequality constraints, and reshape the classic augmented Lagrangian method (ALM) by better balancing its subproblems. As a result, iterations of the balanced ALM treat the objective function and the coefficient matrix in the constraints separately. Compared with the classic ALM (as well as primal-dual type methods), the balanced ALM may have much easier subproblems and thus can be implemented more easily. We also discuss splitting versions of the balanced ALM for separable convex minimization problems. We report some numerical results when the balanced ALM is applied to solve some applications in image processing and cloud computing.

报告人：袁晓明教授（香港大学）

时间：8月14日15:30

地点：海纳苑2幢205教室

报告人简介：袁晓明教授，香港大学数学系系主任，主要研究领域是优化问题算法及理论、最优控制问题、云计算与人工智能问题的建模与计算。2023年与华为云合作获得Franz Edelman Award（finalist）。研究成果发表在SIAM Review等杂志。先后于2017年、2018年、2021年入选 Clavivate Analytics高被引学者。

联系人：孔德兴