Computation of second-order directional stationary points for group sparse optimization

被引:10
作者
Peng, Dingtao [1 ]
Chen, Xiaojun [2 ]
机构
[1] Guizhou Univ, Sch Math & Stat, Guiyang, Guizhou, Peoples R China
[2] Hong Kong Polytech Univ, Dept Appl Math, Hong Kong, Peoples R China
关键词
Group sparse optimization; nonconvex and nonsmooth optimization; composite folded concave penalty; directional stationary point; smoothing method; OPTIMALITY CONDITIONS; VARIABLE SELECTION; REGRESSION; NONSMOOTH; ALGORITHMS; LASSO;
D O I
10.1080/10556788.2019.1684492
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
We consider a nonconvex and nonsmooth group sparse optimization problem where the penalty function is the sum of compositions of a folded concave function and the vector norm for each group variable. We show that under some mild conditions a first-order directional stationary point is a strict local minimizer that fulfils the first-order growth condition, and a second-order directional stationary point is a strong local minimizer that fulfils the second-order growth condition. In order to compute second-order directional stationary points, we construct a twice continuously differentiable smoothing problem and show that any accumulation point of the sequence of second-order stationary points of the smoothing problem is a second-order directional stationary point of the original problem. We give numerical examples to illustrate how to compute a second-order directional stationary point by the smoothing method.
引用
收藏
页码:348 / 376
页数:29
相关论文
共 37 条
[1]   DIFFERENCE-OF-CONVEX LEARNING: DIRECTIONAL STATIONARITY, OPTIMALITY, AND SPARSITY [J].
Ahn, Miju ;
Pang, Jong-Shi ;
Xin, Jack .
SIAM JOURNAL ON OPTIMIZATION, 2017, 27 (03) :1637-1665
[2]   Error bounds for compressed sensing algorithms with group sparsity: A unified approach [J].
Ahsen, M. Eren ;
Vidyasagar, M. .
APPLIED AND COMPUTATIONAL HARMONIC ANALYSIS, 2017, 43 (02) :212-232
[3]  
[Anonymous], 2006, J ROYAL STAT SOC B
[4]   NECESSARY AND SUFFICIENT OPTIMALITY CONDITIONS FOR A CLASS OF NON-SMOOTH MINIMIZATION PROBLEMS [J].
BENTAL, A ;
ZOWE, J .
MATHEMATICAL PROGRAMMING, 1982, 24 (01) :70-91
[5]   Optimality and Complexity for Constrained Optimization Problems with Nonconvex Regularization [J].
Bian, Wei ;
Chen, Xiaojun .
MATHEMATICS OF OPERATIONS RESEARCH, 2017, 42 (04) :1063-1084
[6]   Complexity analysis of interior point algorithms for non-Lipschitz and nonconvex minimization [J].
Bian, Wei ;
Chen, Xiaojun ;
Ye, Yinyu .
MATHEMATICAL PROGRAMMING, 2015, 149 (1-2) :301-327
[7]   Group descent algorithms for nonconvex penalized linear and logistic regression models with grouped predictors [J].
Breheny, Patrick ;
Huang, Jian .
STATISTICS AND COMPUTING, 2015, 25 (02) :173-187
[8]   From Sparse Solutions of Systems of Equations to Sparse Modeling of Signals and Images [J].
Bruckstein, Alfred M. ;
Donoho, David L. ;
Elad, Michael .
SIAM REVIEW, 2009, 51 (01) :34-81
[9]  
Chang T.H., 2017, LOCAL MINIMIZERS 2 O
[10]   OPTIMALITY CONDITIONS AND A SMOOTHING TRUST REGION NEWTON METHOD FOR NONLIPSCHITZ OPTIMIZATION [J].
Chen, Xiaojun ;
Niu, Lingfeng ;
Yuan, Yaxiang .
SIAM JOURNAL ON OPTIMIZATION, 2013, 23 (03) :1528-1552