Inexact primal-dual gradient projection methods for nonlinear optimization on convex set

被引:8
作者
Zhang, Fan [1 ,2 ,3 ]
Wang, Hao [1 ]
Wang, Jiashan [4 ]
Yang, Kai [5 ]
机构
[1] ShanghaiTech Univ, Sch Informat Sci & Technol, Shanghai, Peoples R China
[2] Chinese Acad Sci, Shanghai Inst Microsyst & Informat Technol, Shanghai, Peoples R China
[3] Univ Chinese Acad Sci, Beijing, Peoples R China
[4] Univ Washington, Dept Math, Washington, DC USA
[5] Tongji Univ, Dept Comp Sci, Shanghai, Peoples R China
基金
中国国家自然科学基金;
关键词
Inexact optimization; gradient projection methods; l(1)-ball projection; first-order methods; proximal methods; ALGORITHMS; SPARSITY; SIMPLEX; POINT;
D O I
10.1080/02331934.2019.1696338
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
In this paper, we propose a novel primal-dual inexact gradient projection method for nonlinear optimization problems with convex-set constraint. This method only needs inexact computation of the projections onto the convex set for each iteration, consequently reducing the computational cost for projections per iteration. This feature is attractive especially for solving problems where the projections are computationally not easy to calculate. Global convergence guarantee and ergodic convergence rate of the optimality residual are provided under loose assumptions. We apply our proposed strategy to -ball constrained problems. Numerical results exhibit that our inexact gradient projection methods for solving -ball constrained problems are more efficient than the exact methods.
引用
收藏
页码:2339 / 2365
页数:27
相关论文
共 50 条
  • [31] A prediction-correction-based primal-dual hybrid gradient method for linearly constrained convex minimization
    Ma, Feng
    Bi, Yiming
    Gao, Bin
    NUMERICAL ALGORITHMS, 2019, 82 (02) : 641 - 662
  • [32] A Primal-Dual Splitting Method for Convex Optimization Involving Lipschitzian, Proximable and Linear Composite Terms
    Condat, Laurent
    JOURNAL OF OPTIMIZATION THEORY AND APPLICATIONS, 2013, 158 (02) : 460 - 479
  • [33] Kernel-function-based primal-dual interior-point methods for convex quadratic optimization over symmetric cone
    Cai, Xinzhong
    Wu, Lin
    Yue, Yujing
    Li, Minmin
    Wang, Guoqiang
    JOURNAL OF INEQUALITIES AND APPLICATIONS, 2014,
  • [34] Extension of primal-dual interior point methods to diff-convex problems on symmetric cones
    Valkonen, Tuomo
    OPTIMIZATION, 2013, 62 (03) : 345 - 377
  • [35] Fast primal-dual algorithm via dynamical system for a linearly constrained convex optimization problem
    He, Xin
    Hu, Rong
    Fang, Ya-Ping
    AUTOMATICA, 2022, 146
  • [36] Asynchronous parallel primal-dual block coordinate update methods for affinely constrained convex programs
    Xu, Yangyang
    COMPUTATIONAL OPTIMIZATION AND APPLICATIONS, 2019, 72 (01) : 87 - 113
  • [37] Complexity analysis and numerical implementation of primal-dual interior-point methods for convex quadratic optimization based on a finite barrier
    Cai, Xinzhong
    Wang, Guoqiang
    Zhang, Zihou
    NUMERICAL ALGORITHMS, 2013, 62 (02) : 289 - 306
  • [38] ON CONVERGENCE ANALYSIS OF GRADIENT BASED PRIMAL-DUAL METHOD OF MULTIPLIERS
    Zhang, Guoqiang
    O'Connor, Matthew
    Li, Le
    2018 IEEE STATISTICAL SIGNAL PROCESSING WORKSHOP (SSP), 2018, : 11 - 15
  • [39] Primal-dual and forward gradient implementation for quantitative susceptibility mapping
    Kee, Youngwook
    Deh, Kofi
    Dimov, Alexey
    Spincemaille, Pascal
    Wang, Yi
    MAGNETIC RESONANCE IN MEDICINE, 2017, 78 (06) : 2416 - 2427
  • [40] A FORWARD-BACKWARD VIEW OF SOME PRIMAL-DUAL OPTIMIZATION METHODS IN IMAGE RECOVERY
    Combettes, P. L.
    Condat, L.
    Pesquet, J-C
    Vu, B. C.
    2014 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING (ICIP), 2014, : 4141 - 4145