BACKTRACKING STRATEGIES FOR ACCELERATED DESCENT METHODS WITH SMOOTH COMPOSITE OBJECTIVES

被引:31
作者
Calatroni, Luca [1 ]
Chambolle, Antonin [1 ]
机构
[1] Ecole Polytech, Ctr Math Appl CMAP, CNRS, F-91128 Palaiseau, France
关键词
composite optimization; forward-backward splitting; acceleration; backtracking; image denoising; elastic net; ALGORITHM; MINIMIZATION; OPTIMIZATION; CONVERGENCE; INEXACT;
D O I
10.1137/17M1149390
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
We present and analyze a backtracking strategy for a general fast iterative shrinkage/thresholding algorithm proposed by Chambolle and Pock [Acta Numer., 25 (2016), pp. 161-319] for strongly convex composite objective functions. Unlike classical Armijo-type line searching, our backtracking rule allows for local increasing and decreasing of the descent step size (i.e., proximal parameter) along the iterations. We prove accelerated convergence rates and show numerical results for some exemplar problems.
引用
收藏
页码:1772 / 1798
页数:27
相关论文
共 34 条
[1]  
[Anonymous], 2015, Journal of Computational Mathematics
[2]  
[Anonymous], WORKING PAPER
[3]  
[Anonymous], PREPRINT
[4]  
[Anonymous], 2004, INTRO LECT CONVEX OP
[5]  
[Anonymous], 2017, PREPRINT
[7]   STABILITY OF OVER-RELAXATIONS FOR THE FORWARD-BACKWARD ALGORITHM, APPLICATION TO FISTA [J].
Aujol, J. -F. ;
Dossal, Ch .
SIAM JOURNAL ON OPTIMIZATION, 2015, 25 (04) :2408-2433
[8]   Fast Gradient-Based Algorithms for Constrained Total Variation Image Denoising and Deblurring Problems [J].
Beck, Amir ;
Teboulle, Marc .
IEEE TRANSACTIONS ON IMAGE PROCESSING, 2009, 18 (11) :2419-2434
[9]   A Fast Iterative Shrinkage-Thresholding Algorithm for Linear Inverse Problems [J].
Beck, Amir ;
Teboulle, Marc .
SIAM JOURNAL ON IMAGING SCIENCES, 2009, 2 (01) :183-202
[10]   A VARIABLE METRIC FORWARD-BACKWARD METHOD WITH EXTRAPOLATION [J].
Bonettini, S. ;
Porta, F. ;
Ruggiero, V. .
SIAM JOURNAL ON SCIENTIFIC COMPUTING, 2016, 38 (04) :A2558-A2584