Convergence of Augmented Lagrangian Methods for Composite Optimization Problems

被引:0
|
作者
Hang, Nguyen Thi Van [1 ,2 ]
Sarabi, Ebrahim [3 ]
机构
[1] Nanyang Technol Univ, Sch Phys & Math Sci, Singapore 639798, Singapore
[2] Vietnam Acad Sci & Technol, Inst Biotechnol, Hanoi 100000, Vietnam
[3] Miami Univ, Dept Math, Oxford, OH 45065 USA
基金
美国国家科学基金会;
关键词
augmented Lagrangian; nonunique Lagrange multiplier; second-order sufficient condition; C2-decomposable function; Q-linear convergence; LOCAL CONVERGENCE; REGULARITY;
D O I
10.1287/moor.2023.0324
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
Local convergence analysis of the augmented Lagrangian method (ALM) is established for a large class of composite optimization problems with nonunique Lagrange multipliers under a second-order sufficient condition. We present a new second-order variational property called the semistability of second subderivatives and demonstrate that it is widely satisfied for numerous classes of functions, which is important for applications in constrained and composite optimization problems. Using the latter condition and a certain second-order sufficient condition, we are able to establish Q-linear convergence of the primal-dual sequence for an inexact version of the ALM for composite programs.
引用
收藏
页数:31
相关论文
共 50 条
  • [1] Local convergence of augmented Lagrangian methods for composite optimization
    Kan, C.
    Bi, R.
    Song, W.
    OPTIMIZATION, 2024,
  • [2] LOCAL CONVERGENCE ANALYSIS OF AUGMENTED LAGRANGIAN METHODS FOR PIECEWISE LINEAR-QUADRATIC COMPOSITE OPTIMIZATION PROBLEMS
    Hang, Nguyen T., V
    Sarabi, M. Ebrahim
    SIAM JOURNAL ON OPTIMIZATION, 2021, 31 (04) : 2665 - 2694
  • [3] Augmented Lagrangian Duality for Composite Optimization Problems
    Chao Kan
    Wen Song
    Journal of Optimization Theory and Applications, 2015, 165 : 763 - 784
  • [4] Augmented Lagrangian Duality for Composite Optimization Problems
    Kan, Chao
    Song, Wen
    JOURNAL OF OPTIMIZATION THEORY AND APPLICATIONS, 2015, 165 (03) : 763 - 784
  • [5] Constrained composite optimization and augmented Lagrangian methods
    Alberto De Marchi
    Xiaoxi Jia
    Christian Kanzow
    Patrick Mehlitz
    Mathematical Programming, 2023, 201 : 863 - 896
  • [6] Constrained composite optimization and augmented Lagrangian methods
    De Marchi, Alberto
    Jia, Xiaoxi
    Kanzow, Christian
    Mehlitz, Patrick
    MATHEMATICAL PROGRAMMING, 2023, 201 (1-2) : 863 - 896
  • [7] ON THE CONVERGENCE OF AUGMENTED LAGRANGIAN METHODS FOR CONSTRAINED GLOBAL OPTIMIZATION
    Luo, H. Z.
    Sun, X. L.
    Li, D.
    SIAM JOURNAL ON OPTIMIZATION, 2008, 18 (04) : 1209 - 1230
  • [8] On the convergence of inexact augmented Lagrangian methods for problems with convex constraints
    Galvan, Giulio
    Lapucci, Matteo
    OPERATIONS RESEARCH LETTERS, 2019, 47 (03) : 185 - 189
  • [9] Convergence properties of augmented Lagrangian methods for constrained global optimization
    Luo, Hezhi
    Sun, Xiaoling
    Wu, Huixian
    OPTIMIZATION METHODS & SOFTWARE, 2008, 23 (05): : 763 - 778
  • [10] Augmented Lagrangian Methods for Convex Matrix Optimization Problems
    Ying Cui
    Chao Ding
    Xu-Dong Li
    Xin-Yuan Zhao
    Journal of the Operations Research Society of China, 2022, 10 : 305 - 342