Optimally linearizing the alternating direction method of multipliers for convex programming

被引:47
|
作者
He, Bingsheng [1 ,2 ]
Ma, Feng [3 ]
Yuan, Xiaoming [4 ]
机构
[1] Southern Univ Sci & Technol China, Dept Math, Shenzhen, Peoples R China
[2] Nanjing Univ, Dept Math, Nanjing, Peoples R China
[3] Hightech Inst Xian, Xian 710025, Shaanxi, Peoples R China
[4] Univ Hong Kong, Dept Math, Hong Kong, Peoples R China
基金
中国国家自然科学基金;
关键词
Convex programming; Alternating direction method of multipliers; Linearized; Optimal step size; Proximal regularization; Convergence rate; RACHFORD SPLITTING METHOD; CONVERGENCE; ALGORITHM;
D O I
10.1007/s10589-019-00152-3
中图分类号
C93 [管理学]; O22 [运筹学];
学科分类号
070105 ; 12 ; 1201 ; 1202 ; 120202 ;
摘要
The alternating direction method of multipliers (ADMM) is being widely used in a variety of areas; its different variants tailored for different application scenarios have also been deeply researched in the literature. Among them, the linearized ADMM has received particularly wide attention in many areas because of its efficiency and easy implementation. To theoretically guarantee convergence of the linearized ADMM, the step size for the linearized subproblems, or the reciprocal of the linearization parameter, should be sufficiently small. On the other hand, small step sizes decelerate the convergence numerically. Hence, it is interesting to probe the optimal (largest) value of the step size that guarantees convergence of the linearized ADMM. This analysis is lacked in the literature. In this paper, we provide a rigorous mathematical analysis for finding this optimal step size of the linearized ADMM and accordingly set up the optimal version of the linearized ADMM in the convex programming context. The global convergence and worst-case convergence rate measured by the iteration complexity of the optimal version of linearized ADMM are proved as well.
引用
收藏
页码:361 / 388
页数:28
相关论文
共 50 条
  • [21] A Note on the Alternating Direction Method of Multipliers
    Han, Deren
    Yuan, Xiaoming
    JOURNAL OF OPTIMIZATION THEORY AND APPLICATIONS, 2012, 155 (01) : 227 - 238
  • [22] Accelerated Alternating Direction Method of Multipliers
    Kadkhodaie, Mojtaba
    Christakopoulou, Konstantina
    Sanjabi, Maziar
    Banerjee, Arindam
    KDD'15: PROCEEDINGS OF THE 21ST ACM SIGKDD INTERNATIONAL CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING, 2015, : 497 - 506
  • [23] A Linearized Alternating Direction Method of Multipliers with Substitution Procedure
    Chao, Miantao
    Cheng, Caozong
    Zhang, Haibin
    ASIA-PACIFIC JOURNAL OF OPERATIONAL RESEARCH, 2015, 32 (03)
  • [24] ALTERNATING DIRECTION METHOD OF MULTIPLIERS FOR LINEAR INVERSE PROBLEMS
    Jiao, Yuling
    Jin, Qinian
    Lu, Xiliang
    Wang, Weijie
    SIAM JOURNAL ON NUMERICAL ANALYSIS, 2016, 54 (04) : 2114 - 2137
  • [25] A Penalty Alternating Direction Method of Multipliers for Convex Composite Optimization Over Decentralized Networks
    Zhang, Jiaojiao
    Liu, Huikang
    Sow, Anthony Man-Cho
    Ling, Qing
    IEEE TRANSACTIONS ON SIGNAL PROCESSING, 2021, 69 : 4282 - 4295
  • [26] Alternating Direction Method of Multipliers for Separable Convex Optimization of Real Functions in Complex Variables
    Li, Lu
    Wang, Xingyu
    Wang, Guoqiang
    MATHEMATICAL PROBLEMS IN ENGINEERING, 2015, 2015
  • [27] ALTERNATING DIRECTION METHOD OF MULTIPLIERS WITH VARIABLE METRIC INDEFINITE PROXIMAL TERMS FOR CONVEX OPTIMIZATION
    Gu, Yan
    Yamashita, Nobuo
    NUMERICAL ALGEBRA CONTROL AND OPTIMIZATION, 2020, 10 (04): : 487 - 510
  • [28] Inexact alternating direction methods of multipliers for separable convex optimization
    Hager, William W.
    Zhang, Hongchao
    COMPUTATIONAL OPTIMIZATION AND APPLICATIONS, 2019, 73 (01) : 201 - 235
  • [29] DUAL DESCENT AUGMENTED LAGRANGIAN METHOD AND ALTERNATING DIRECTION METHOD OF MULTIPLIERS
    Sun, Kaizhao
    Sun, Xu Andy
    SIAM JOURNAL ON OPTIMIZATION, 2024, 34 (02) : 1679 - 1707
  • [30] A proximal point algorithm revisit on the alternating direction method of multipliers
    Cai XingJu
    Gu GuoYong
    He BingSheng
    Yuan XiaoMing
    SCIENCE CHINA-MATHEMATICS, 2013, 56 (10) : 2179 - 2186