pith. sign in

arxiv: 2606.08424 · v1 · pith:R6L6ORNKnew · submitted 2026-06-07 · 🧮 math.OC

Augmented Lagrangian methods for nonlinear semidefinite programming with complementarity constraints

Daiana O. Santos , Carina M. Costa , Evelin H. M. Krulikovski e Marina Geremia This is my paper

Pith reviewed 2026-06-27 18:17 UTC · model grok-4.3

classification 🧮 math.OC
keywords semidefinite programmingcomplementarity constraintsaugmented LagrangianW-stationarityC-stationarityRobinson conditionspectral decompositionnonlinear programming
0
0 comments X

The pith

An augmented Lagrangian method on a spectral reformulation of SDCMPCC produces W-stationary accumulation points under an extended Robinson condition.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper addresses nonlinear semidefinite programs with complementarity constraints, known as SDCMPCC, which are highly degenerate so that classical optimality conditions fail. It introduces a reformulation of the complementarity structure based on spectral decomposition that preserves local solutions and permits standard analysis. An augmented Lagrangian method is applied to the reformulated problem. Under a suitable extension of Robinson's condition, every accumulation point of the generated sequence is shown to be W-stationary; a stricter condition on how the subproblems are solved upgrades this to C-stationarity.

Core claim

The central claim is that after reformulating SDCMPCC via spectral decomposition of the complementarity structure, the augmented Lagrangian method generates a sequence whose accumulation points are W-stationary whenever an extended Robinson condition holds, and C-stationary when the subproblems are solved under a stricter condition.

What carries the argument

Spectral decomposition reformulation of the complementarity structure, which preserves local solutions and enables tractable stationarity analysis of the subsequent augmented Lagrangian iterates.

If this is right

  • Every accumulation point of the sequence satisfies W-stationarity under the extended Robinson condition.
  • Stricter accuracy in solving the subproblems upgrades the limit to C-stationarity.
  • The reformulation leaves the set of local solutions unchanged.
  • Standard augmented Lagrangian convergence theory applies once the problem is rewritten in the decomposed form.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same decomposition technique may allow convergence proofs for other penalty or sequential methods on SDCMPCC.
  • Numerical tests on concrete SDCMPCC instances could reveal how often the extended Robinson condition holds in practice.

Load-bearing premise

The spectral decomposition reformulation of the complementarity structure preserves local solutions of the original problem.

What would settle it

A concrete SDCMPCC instance satisfying the extended Robinson condition in which the augmented Lagrangian sequence has an accumulation point that is not W-stationary would falsify the convergence claim.

read the original abstract

We consider nonlinear semidefinite programming problems with complementarity constraints (SDCMPCC), a class of highly degenerate problems where classical optimality conditions typically fail. In this context, weak stationarity conditions have been developed to address their degeneracy. While these notions are well understood, their algorithmic implications remain largely unexplored in semidefinite complementarity settings. We introduce a reformulation based on the spectral decomposition of the complementarity structure, which preserves local solutions and enables a tractable analysis. Within this framework, we analyze an augmented Lagrangian method for SDCMPCC and prove that under a suitable extension of Robinson's condition, every accumulation point of the generated sequence is W-stationary, or C-stationary under a stricter condition for solving the subproblems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 3 minor

Summary. The paper considers nonlinear semidefinite programming problems with complementarity constraints (SDCMPCC). It introduces a reformulation of the complementarity structure based on spectral decomposition that preserves local solutions, then analyzes an augmented Lagrangian method applied to the reformulated problem. Under a suitable extension of Robinson's condition, every accumulation point of the generated sequence is shown to be W-stationary; under a stricter condition on the subproblem solver, the limit points are C-stationary.

Significance. The result is significant for providing the first convergence analysis of an augmented Lagrangian method to weak stationarity notions in the SDCMPCC setting, where standard KKT conditions fail due to degeneracy. The spectral reformulation is a key technical device that renders the analysis tractable while preserving local optimality. The paper delivers a conditional convergence theorem under explicitly stated constraint qualifications rather than parameter-tuned or self-referential assumptions.

minor comments (3)
  1. [Section 3] The definition and precise statement of the extended Robinson condition (used in the main convergence theorem) should be stated explicitly in the main text rather than deferred entirely to an appendix, to improve readability for readers focused on the algorithmic result.
  2. [Theorem 4.3] In the proof of the main accumulation-point result, the passage from the spectral reformulation back to the original complementarity variables could be expanded with one additional sentence clarifying why the W-stationarity notion transfers without additional assumptions.
  3. [Section 4] Notation for the spectral decomposition (eigenvalues and projectors) is introduced in Section 2 but reused in Section 4 without a brief reminder; a one-line cross-reference would prevent minor confusion.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive summary, significance assessment, and recommendation for minor revision. No specific major comments were raised in the report.

Circularity Check

0 steps flagged

No significant circularity; derivation relies on external conditions and standard analysis

full rationale

The paper introduces a spectral reformulation of SDCMPCC that is claimed to preserve local solutions, then proves convergence of an augmented Lagrangian method to W-stationarity (or C-stationarity) under an explicitly extended Robinson condition. Robinson's condition is a standard external concept from nonlinear programming literature, not derived within the paper. No equations reduce a claimed result to a fitted parameter, self-definition, or self-citation chain; the proof steps are presented as independent analysis under stated assumptions. The central claim does not collapse to its inputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

Based on abstract only; the central claim rests on the validity of the spectral reformulation preserving solutions and on an extended Robinson condition whose precise statement is not given here.

axioms (2)
  • domain assumption Spectral decomposition reformulation preserves local solutions
    Invoked to justify the reformulation step that enables the subsequent analysis.
  • domain assumption Extended Robinson condition
    The convergence result is conditioned on this extension holding for the problem.

pith-pipeline@v0.9.1-grok · 5661 in / 1139 out tokens · 18685 ms · 2026-06-27T18:17:27.226112+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

28 extracted references · 23 canonical work pages

  1. [1]

    Andreani, G

    R. Andreani, G. Haeser, and J. M. Mart´ ınez. On sequential optimality conditions for smooth constrained optimization.Optimization, 60(5):627–641, 2011. doi:10.1080/02331930903578700

    work page doi:10.1080/02331930903578700 2011

  2. [2]

    Andreani, G

    R. Andreani, G. Haeser, L. D. Secchin, and P. J. S. Silva. New sequential optimality conditions for mathematical problems with complementarity constraints and algorithmic consequences.SIAM Journal on Optimization, 29(4):3201–3230, 2019. doi:10.1137/18M121040X

    work page doi:10.1137/18m121040x 2019

  3. [3]

    Andreani, G

    R. Andreani, G. Haeser, and D. S. Viana. Optimality conditions and global convergence for nonlinear semidefinite programming.Mathematical Programming, 180(1):203–235, 2020. doi:10.1007/s10107-018- 1354-5

    work page doi:10.1007/s10107-018- 2020

  4. [4]

    Andreani, G

    R. Andreani, G. Haeser, L. M. Mito, H. Ram´ ırez, and T. P. Silveira. First- and second-order optimal- ity conditions for second-order cone and semidefinite programming under a constant rank condition. Mathematical Programming, 202:473–514, 2023. doi:10.1007/s10107-022-01863-2

    work page doi:10.1007/s10107-022-01863-2 2023

  5. [5]

    Andreani, G

    R. Andreani, G. Haeser, L.M. Mito, and H. Ram´ ırez. Sequential constant rank constraint qualifications for nonlinear semidefinite programming with applications.Set-Valued and Variational Analysis, 31(3), 2023

    2023

  6. [6]

    Andreani, G

    R. Andreani, G. Haeser, M. da Rosa, and D. O. Santos. On constraint qualifications for lower-level sets and an augmented lagrangian method. 2025. Preprint

    2025

  7. [7]

    E. G. Birgin, N. Krejic, and J. M. Mart´ ınez. On the minimization of possibly discontinuous functions by means of pointwise approximations.Optimization Letters, 11(8):1623–1637, 2017. doi:10.1007/s11590- 016-1068-7

    work page doi:10.1007/s11590- 2017

  8. [8]

    L. F. Bueno, G. Haeser, and F. N. Rojas. Optimality conditions and constraint qualifications for generalized nash equilibrium problems and their practical implications.SIAM Journal on Optimization, 29(1):31–54, 2019. doi:10.1137/17M1162524

    work page doi:10.1137/17m1162524 2019

  9. [9]

    L. F. Bueno, G. Haeser, F. Lara, and F. N. Rojas. An augmented lagrangian method for quasi-equilibrium problems.Computational Optimization and Applications, 76(3):737–766, 2020. doi:10.1007/s10589-020-00180-4

    work page doi:10.1007/s10589-020-00180-4 2020

  10. [10]

    S. Dempe. Annotated bibliography on bilevel programming and mathematical programs with equilib- rium constraints.Optimization, 52(3):333–359, 2003. doi:10.1080/0233193031000149894

    work page doi:10.1080/0233193031000149894 2003

  11. [11]

    S. Dempe. Bilevel optimization: theory, algorithms, applications and a bibliography.Bilevel optimiza- tion: advances and next challenges, pages 581–672, 2020

    2020

  12. [12]

    Dempe, F

    S. Dempe, F. M. Kue, and P. Mehlitz. Optimality conditions for special semidefinite bilevel optimization problems.SIAM Journal on Optimization, 28(2):1564–1587, 2018. doi:10.1137/16M1099303

    work page doi:10.1137/16m1099303 2018

  13. [13]

    C. Ding, D. Sun, and J. J. Ye. First order optimality conditions for mathematical programs with semidefinite cone complementarity constraints.Mathematical Programming, 147:539–579, 2014. doi:10.1007/s10107-013-0735-z

    work page doi:10.1007/s10107-013-0735-z 2014

  14. [14]

    Feng and S

    M. Feng and S. Li. An approximate strong kkt condition for multiobjective optimization.TOP, 26(3): 489–509, 2018. doi:10.1007/s11750-018-0491-6. 19

    work page doi:10.1007/s11750-018-0491-6 2018

  15. [15]

    M. C. Ferris and J. S. Pang. Engineering and economic applications of complementarity problems. SIAM Review, 39(4):669–713, 1997. doi:10.1137/S0036144595285963

    work page doi:10.1137/s0036144595285963 1997

  16. [16]

    M. L. Flegel and C. Kanzow. On the guignard constraint qualification for mathematical programs with equilibrium constraints.Optimization, 54(6):517–534, 2005. doi:10.1080/02331930500342591

    work page doi:10.1080/02331930500342591 2005

  17. [17]

    Giorgi, B

    G. Giorgi, B. Jimenez, and V. Novo. Approximate karush-kuhn-tucker condition in multiobjective op- timization.Journal of Optimization Theory and Applications, 171(1):70–89, 2016. doi:10.1007/s10957- 016-0986-y

    work page doi:10.1007/s10957- 2016

  18. [18]

    Guo and N

    S. Guo and N. Deng. A new augmented lagrangian method for mpccs: theoretical and numerical com- parison with existing augmented lagrangian methods.Computational Optimization and Applications, 82(3):779–808, 2022. doi:10.1007/s10589-022-00382-0

    work page doi:10.1007/s10589-022-00382-0 2022

  19. [19]

    Haeser and M

    G. Haeser and M. L. Schuverdt. On approximate kkt condition and its extension to continuous variational inequalities.Journal of Optimization Theory and Applications, 149(3):528–539, 2011. doi:10.1007/s10957-011-9802-x

    work page doi:10.1007/s10957-011-9802-x 2011

  20. [20]

    A. F. Izmailov, M. V. Solodov, and E. I. Uskov. Global convergence of augmented lagrangian methods applied to optimization problems with degenerate constraints, including problems with complementarity constraints.SIAM Journal on Optimization, 22(4):1579–1606, 2012. doi:10.1137/120868359

    work page doi:10.1137/120868359 2012

  21. [21]

    X. Jia, C. Kanzow, P. Mehlitz, and G. Wachsmuth. An augmented lagrangian method for optimization problems with structured geometric constraints.Mathematical Programming, 199:1365–1415, 2023. doi:10.1007/s10107-022-01870-z

    work page doi:10.1007/s10107-022-01870-z 2023

  22. [22]

    Kanzow, D

    C. Kanzow, D. Steck, and D. Wachsmuth. An augmented lagrangian method for optimization problems in banach spaces.SIAM Journal on Control and Optimization, 56(1):272–291, 2018. doi:10.1137/16M1107103

    work page doi:10.1137/16m1107103 2018

  23. [23]

    Kanzow and D

    D. Kanzow and D. Steck. Augmented lagrangian and exact penalty methods for quasi-variational inequalities.Computational Optimization and Applications, 69(3):801–824, 2018. doi:10.1007/s10589- 017-9963-0

    work page doi:10.1007/s10589- 2018

  24. [24]

    Z. Luo, J. Pang, and D. Ralph.Mathematical programs with equilibrium constraints. Cambridge University Press, 1996

    1996

  25. [25]

    J. I. Madariaga and H. Ram´ ırez. Facial approach for constructing stationary points for mathematical programs with cone complementarity constraints.Journal of Optimization Theory and Applications, 204:15, 2025. doi:10.1007/s10957-024-02562-8

    work page doi:10.1007/s10957-024-02562-8 2025

  26. [26]

    A. Ramos. Mathematical programs with equilibrium constraints: a sequential optimality condition, new constraint qualifications and algorithmic consequences.Optimization Methods and Software, 36: 45–81, 2019. doi:10.1080/10556788.2019.1702661

    work page doi:10.1080/10556788.2019.1702661 2019

  27. [27]

    R. T. Rockafellar and R. Wets.Variational Analysis, volume 317 ofGrundlehren der Mathematischen Wissenschaften. Springer–Verlag, Berlin, 1998

    1998

  28. [28]

    Yan and M

    T. Yan and M. Fukushima. Smoothing method for mathematical programs with symmetric cone com- plementarity constraints.Optimization, 60:113–128, 2011. doi:10.1080/02331934.2010.541458. 20

    work page doi:10.1080/02331934.2010.541458 2011