Some Recent Developments of ADMM: New Theory and Applications

讲座简介

In this talk, we present some recent developments of ADMM for linearly constrained optimization problems.  Specifically, 1) we provide a new convergence result of ADMM for infeasible convex problems. We prove that the primal sequence generated by ADMM is still convergent and it will converge to a solution to the convex problem with the least constraint violation.  2) we also discuss some new applications of ADMM, including the Markov decision processes and distributionally robust optimization.

时间

2021-12-01

下午 13:30 ~ 15:00

主讲人

陈彩华，南京大学副教授

地点

会议号码:  747 485 566

参会密码:  123456

参会链接：https://meeting.tencent.com/dm/X0oL8NzL7aPS