Year: 2015
Type: article
Abstract: We present a novel framework, namely, accelerated alternating direction method of multipliers (AADMM), for acceleration of linearized ADMM. The basic idea of AADMM is to incorporate a multistep acceleration scheme into linearized ADMM. We demonstrate that for solving a class of convex composite optimization with linear constraints, the rate of convergence of AADMM is better than that of linearized ADMM, in terms of their dependence on the Lipschitz constant of the smooth component. Moreover, AADMM is capable of dealing with the situation when the feasible region is unbounded, as long as the corresponding saddle point problem has a solution. A backtracking algorithm is also proposed for practical performance.
Source: SIAM Journal on Imaging Sciences
DOI: 10.1137/14095697x
Institution: University of Florida
Language: English
Subfield: Computational Theory and Mathematics
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Awards: CAREER: Reduced-order Methods for Big-Data Challenges in Nonlinear and Stochastic Optimization, G3387330441, Accelerated Algorithms for a Class of Saddle Point problems and Variational Inequalities, PFI-BIC: Innovation Transfer of the Portable Nuclear Moment Imaging Platform, Fast TV-Regularized Large-Scale and Ill-Conditioned Linear Inversion with Application to PPI, G600700408, Theory and Applications of Stochastic First-order Methods for Large-scale Stochastic Convex Optimization, G6671297155
OpenAlex ID: W1972552174
