A QCQP-based splitting SQP algorithm for two-block nonconvex constrained optimization problems with application

被引:16
作者
Jian, Jinbao [1 ]
Liu, Pengjie [2 ]
Yin, Jianghua [3 ]
Zhang, Chen [4 ]
Chao, Miantao [2 ]
机构
[1] Guangxi Univ Nationalities, Guangxi Key Lab Hybrid Computat & IC Design Anal, Coll Math & Phys, Ctr Appl Math & Artificial Intelligence, Nanning 530006, Peoples R China
[2] Guangxi Univ, Coll Math & Informat Sci, Nanning 530004, Peoples R China
[3] Guangxi Sci & Technol Normal Univ, Coll Math & Comp Sci, Laibin 546199, Peoples R China
[4] Univ Shanghai Sci & Technol, Sch Mech Engn, Shanghai 200093, Peoples R China
基金
中国国家自然科学基金;
关键词
Two-block nonconvex constrained optimization; Quadratically constrained quadratic programming; Sequential quadratic programming; Splitting method; Convergence analysis; QUADRATIC-PROGRAMMING ALGORITHM; ALTERNATING DIRECTION METHOD; GLOBAL CONVERGENCE; SUPERLINEAR CONVERGENCE; DESCENT METHODS; PENALTY; FILTER; MULTIPLIERS;
D O I
10.1016/j.cam.2020.113368
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
This paper discusses a class of two-block nonconvex smooth optimization problems with nonlinear constraints. Based on a quadratically constrained quadratic programming (QCQP) approximation, an augmented Lagrangian function (ALF), and a Lagrangian splitting technique into small-scale subproblems, we propose a novel sequential quadratic programming (SQP) algorithm. First, inspired by the augmented Lagrangian method (ALM), we penalize the quadratic equality constraints associated with the QCQP approximation subproblem in its objective by means of the ALF, and then split the resulting subproblem into two small-scale ones, but both of them are not quadratic programming (QP) due to the square of the quadratic equality constraints in the objective. Second, by ignoring the three-order infinitesimal arising from the squared term, the two small-scale subproblems are reduced to two standard QP subproblems, which can yield an improved search direction. Third, taking the ALF of the discussed problem as a merit function, the next iterate point is generated by the Armijo line search. As a result, a new SQP method, called QCQP-based splitting SQP method, is proposed. Under suitable conditions, the global convergence, strong convergence, iteration complexity and convergence rate of the proposed method are analyzed and obtained. Finally, preliminary numerical experiments and applications were carried out, and these show that the proposed method is promising. (C) 2020 Elsevier B.V. All rights reserved.
引用
收藏
页数:23
相关论文
共 58 条
  • [11] Global convergence of a trust-region SQP-filter algorithm for general nonlinear programming
    Fletcher, R
    Gould, NIM
    Leyffer, S
    Toint, PL
    Wächter, A
    [J]. SIAM JOURNAL ON OPTIMIZATION, 2003, 13 (03) : 635 - 659
  • [12] On the global convergence of a filter SQP algorithm
    Fletcher, R
    Leyffer, S
    Toint, PL
    [J]. SIAM JOURNAL ON OPTIMIZATION, 2002, 13 (01) : 44 - 59
  • [13] Nonlinear programming without a penalty function
    Fletcher, R
    Leyffer, S
    [J]. MATHEMATICAL PROGRAMMING, 2002, 91 (02) : 239 - 269
  • [14] A sequential quadratically constrained quadratic programming method for differentiable convex minimization
    Fukushima, M
    Luo, ZQ
    Tseng, P
    [J]. SIAM JOURNAL ON OPTIMIZATION, 2003, 13 (04) : 1098 - 1119
  • [16] A GLOBALLY CONVERGENT STABILIZED SQP METHOD
    Gill, Philip E.
    Robinson, Daniel P.
    [J]. SIAM JOURNAL ON OPTIMIZATION, 2013, 23 (04) : 1983 - 2010
  • [17] Fast alternating linearization methods for minimizing the sum of two convex functions
    Goldfarb, Donald
    Ma, Shiqian
    Scheinberg, Katya
    [J]. MATHEMATICAL PROGRAMMING, 2013, 141 (1-2) : 349 - 382
  • [18] Fast Alternating Direction Optimization Methods
    Goldstein, Tom
    O'Donoghue, Brendan
    Setzer, Simon
    Baraniuk, Richard
    [J]. SIAM JOURNAL ON IMAGING SCIENCES, 2014, 7 (03): : 1588 - 1623
  • [19] Stabilized sequential quadratic programming
    Hager, WW
    [J]. COMPUTATIONAL OPTIMIZATION AND APPLICATIONS, 1999, 12 (1-3) : 253 - 273
  • [20] Convergence Study on the Symmetric Version of ADMM with Larger Step Sizes
    He, Bingsheng
    Ma, Feng
    Yuan, Xiaoming
    [J]. SIAM JOURNAL ON IMAGING SCIENCES, 2016, 9 (03): : 1467 - 1501