Stabilized sequential quadratic programming for optimization and a stabilized Newton-type method for variational problems

被引:46
|
作者
Fernandez, Damian [1 ]
Solodov, Mikhail [1 ]
机构
[1] IMPA, BR-22460320 Rio De Janeiro, Brazil
关键词
Stabilized sequential quadratic programming; Karush-Kuhn-Tucker system; Variational inequality; Newton methods; Superlinear convergence; Error bound; ERROR-BOUNDS; DEGENERATE; ALGORITHM; SQP;
D O I
10.1007/s10107-008-0255-4
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
The stabilized version of the sequential quadratic programming algorithm (sSQP) had been developed in order to achieve fast convergence despite possible degeneracy of constraints of optimization problems, when the Lagrange multipliers associated to a solution are not unique. Superlinear convergence of sSQP had been previously established under the strong second-order sufficient condition for optimality (without any constraint qualification assumptions). We prove a stronger superlinear convergence result than the above, assuming the usual second-order sufficient condition only. In addition, our analysis is carried out in the more general setting of variational problems, for which we introduce a natural extension of sSQP techniques. In the process, we also obtain a new error bound for Karush-Kuhn-Tucker systems for variational problems that holds under an appropriate second-order condition.
引用
收藏
页码:47 / 73
页数:27
相关论文
共 50 条
  • [21] A globally convergent proximal Newton-type method in nonsmooth convex optimization
    Mordukhovich, Boris S.
    Yuan, Xiaoming
    Zeng, Shangzhi
    Zhang, Jin
    MATHEMATICAL PROGRAMMING, 2023, 198 (01) : 899 - 936
  • [22] Examples of dual behaviour of Newton-type methods on optimization problems with degenerate constraints
    A. F. Izmailov
    M. V. Solodov
    Computational Optimization and Applications, 2009, 42 : 231 - 264
  • [23] INEXACT NEWTON-TYPE OPTIMIZATION WITH ITERATED SENSITIVITIES
    Quirynen, Rien
    Gros, Sebastien
    Diehl, Moritz
    SIAM JOURNAL ON OPTIMIZATION, 2018, 28 (01) : 74 - 95
  • [24] Examples of dual behaviour of Newton-type methods on optimization problems with degenerate constraints
    Izmailov, A. F.
    Solodov, M. V.
    COMPUTATIONAL OPTIMIZATION AND APPLICATIONS, 2009, 42 (02) : 231 - 264
  • [25] The Sequential Quadratic Programming Method
    Fletcher, Roger
    NONLINEAR OPTIMIZATION, 2010, 1989 : 165 - 214
  • [26] A truly globally convergent feasible Newton-type method for mixed complementarity problems
    Han, D
    JOURNAL OF COMPUTATIONAL MATHEMATICS, 2004, 22 (03) : 347 - 360
  • [27] New Proximal Newton-Type Methods for Convex Optimization
    Adler, Ilan
    Hu, Zhiyue T.
    Lin, Tianyi
    2020 59TH IEEE CONFERENCE ON DECISION AND CONTROL (CDC), 2020, : 4828 - 4835
  • [28] A modified sequential quadratic programming method for sparse signal recovery problems
    Alamdari, Mohammad Saeid
    Fatemi, Masoud
    Ghaffari, Aboozar
    SIGNAL PROCESSING, 2023, 207
  • [29] Newton-Type Alternating Minimization Algorithm for Convex Optimization
    Stella, Lorenzo
    Themelis, Andreas
    Patrinos, Panagiotis
    IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 2019, 64 (02) : 697 - 711
  • [30] A new local and global optimization method for mixed integer quadratic programming problems
    Li, G. Q.
    Wu, Z. Y.
    Quan, J.
    APPLIED MATHEMATICS AND COMPUTATION, 2010, 217 (06) : 2501 - 2512