The acceleration technique introduced by Nesterov for gradient descent is widely used in optimization but its principles are not yet fully understood. Recently, significant progress has been made to close this understanding gap through a continuous time dynamical systems perspective associated with gradient based methods for smooth and unconstrained problems. Here we extend this perspective to nonsmooth and constrained problems by deriving nonsmooth dynamical systems related to variants of the relaxed and accelerated alternating direction method of multipliers (ADMM). More specifically, we introduce two new accelerated ADMM variants, depending on two types of dissipation, and derive differential inclusions that model these algorithms in the continuous time limit. Through a nonsmooth Lyapunov analysis, we obtain rates of convergence for these dynamical systems in the convex and strongly convex settings that illustrate an interesting tradeoff between decaying versus constant damping strategies.
翻译:Nesterov为梯度下降引进的加速技术在优化中广泛使用,但其原则尚未完全理解。最近,通过持续的时间动态系统视角,与基于梯度的平滑和不受约束问题的方法相关联,在弥合这一理解差距方面取得了显著进展。在这里,我们通过生成与放松和加速交替的乘数方法(ADMM)变量相关的非平稳动态系统,将这一视角扩大到非平稳和受约束的问题。更具体地说,我们引入了两种新的加速的ADMM变体,这取决于两种类型的消散,并得出了在连续时限内模拟这些算法的差别包容。 通过非mooth Lyapunov 分析,我们获得了这些动态系统在二次曲线和强烈交错环境中的趋同率,这显示了衰减与常态阻断策略之间的有趣的平衡。