A generic online acceleration scheme for optimization algorithms via relaxation and inertia. (4th March 2019)

• Record Type: Journal Article
• Authors: Iutzeler, F.; Hendrickx, J. M.
• Is Part Of: Optimization methods and software. Volume 34:Number 2(2019)
• Page Start: 383
• Page End: 405
• Publication Date: 2019-03-04
• Subjects: Applied optimization methods -- relaxation inertia -- acceleration; 65K10 -- 90C25 -- 65B99 -- 47H05; Mathematical optimization -- Periodicals; Algorithms -- Periodicals; 519.7
• DOI: 10.1080/10556788.2017.1396601
• Languages: English
• ISSNs: 1055-6788

Abstract

Abstract : We propose generic acceleration schemes for a wide class of optimization and iterative schemes based on relaxation and inertia. In particular, we introduce methods that automatically tune the acceleration coefficients online and establish their convergence. This is made possible by considering classes of fixed-point iterations over averaged operators which encompass gradient methods, ADMM (Alternating Direction Method of Multipliers), primal dual algorithms and so on.