On the pointwise iteration-complexity of a dynamic regularized ADMM with over-relaxation stepsize

作者:

Highlights:

摘要

In this paper, we extend the improved pointwise iteration-complexity estimation of a dynamic regularized alternating direction method of multipliers (ADMM) for a new stepsize domain. In this complexity analysis, the stepsize parameter can be chosen in the interval (0,2) instead of interval (0,(1+5)/2). We illustrate, by means of a numerical experiment, that the enlargement of this stepsize domain can lead to better performance of the method in some applications. Our complexity study is established by interpreting this ADMM variant as an instance of a hybrid proximal extragradient framework applied to a specific monotone inclusion problem.

论文关键词:Alternating direction method of multipliers,Hybrid proximal extragradient framework,Pointwise iteration-complexity,Convex programming

论文评审过程:Received 5 November 2017, Revised 23 April 2018, Accepted 29 April 2018, Available online 26 May 2018, Version of Record 26 May 2018.

论文官网地址:https://doi.org/10.1016/j.amc.2018.04.066