Abstract
In this paper, we propose an augmented Lagrangian method with Backtracking Line Search for solving nonconvex composite optimization problems including both nonlinear equality and inequality constraints. In case the variable spaces are homogeneous, our setting yields a generic nonlinear mathematical programming model. When some variables belong to the real Hilbert space and others to the integer space, one obtains a nonconvex mixed-integer/-binary nonlinear programming model for which the nonconvexity is not limited to the integrality constraints. Together with the formal proof of its iteration complexity, the proposed algorithm is then numerically evaluated to solve a multi-constrained network design problem. Extensive numerical executions on a set of instances extracted from the SNDlib repository are then performed to study its behavior and performance as well as identify potential improvement of this method. Finally, analysis of the results and their comparison against those obtained when solving its convex relaxation using mixed-integer programming solvers are reported.
Similar content being viewed by others
Data Availability
The network topology datasets generated during and/or analysed during the current study are available in the SNDlib repository (Orlowski et al. 2010), http://sndlib.zib.de/problems.overview.action. The demand datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.
Notes
Remember that, following the right scalar multiplication rule with \(\gamma > 0\), the proximal operator associated with the function \(\gamma f(\gamma ^{-1} x)\) is \(\gamma {\text {prox}}_{\gamma ^{-1} f}(\gamma ^{-1} x)\).
\({\mathcal {N}}_C\) denotes the limiting (or Mordukhovich) normal cone; this cone is generally nonconvex; hence, it cannot be polar to any tangential approximation of C.
The set of KKT multipliers (\(\lambda _1,\lambda _2\)) at \(u^{\dagger }\) is nonempty.
If c(u) is a vector function, \(c(u)_+ {:}{=}\max [c(u), 0]\), where the maximum is taken componentwise.
References
Andreani R, Birgin EG, Martínez JM, Schuverdt ML (2008) Augmented Lagrangian methods under the constant positive linear dependence constraint qualification. Math Program Ser B 111:5–32
Attouch H, Bolte J, Redont P, Soubeyran A (2010) Proximal alternating minimization and projection methods for nonconvex problems. An approach based on the Kurdyka–Lojasiewicz inequality. Math Oper Res 35(2):438–457
Attouch H, Bolte J, Svaiter BF (2013) Convergence of descent methods for semi-algebraic and tame problems: proximal algorithms, forward-backward splitting, and regularized Gauss–Seidel methods. Math Program Ser A 137:91–129
Bauschke HH, Combettes PL (2017) Convex analysis and monotone operator theory in Hilbert spaces, 2nd edn. Springer, New York
Beck A, Teboulle M (2012) Smoothing and first order methods: a unified framework. SIAM J Optim 22(2):557–580
Becker S, Bobin J, Candès EJ (2011) NESTA: a fast and accurate first-order method for sparse recovery. SIAM J Imaging Sci 4(1):1–39
Ben-Ameur W, Ouorou A (2006) Mathematical models of the delay constrained routing problem. Algorithmic Oper Res 1(2):94–103
Bolte J, Daniilidis A, Lewis AS (2007) The Lojasiewicz inequality for nonsmooth subanalytic functions with applications to subgradient dynamical systems. SIAM J Optim 17(4):1205–1223
Bolte J, Sabach S, Teboulle M (2018) Nonconvex Lagrangian-based optimization: monitoring schemes and global convergence. Math Oper Res 43(4):1210–1232
Burke JV, Engle A (2018) Line search and trust-region methods for convex-composite optimization. arXiv:1806.05218
Fischer A (1992) A special Newton-type optimization method. Optimization 24(3–4):269–284
Gabay D (1983) Applications of the method of multipliers to variational inequalities. In: Fortinand M, Glowinski R (eds) Augmented Lagrangian methods: applications to the solution of boundary-value problems, North-Holland, Amsterdam
Garrigos PG, Garrigos G, Peypouquet J (2015) Splitting methods with variable metric for Kurdyka–Łojasiewicz functions and general convergence rates. J Optim Theory Appl 165:874–900
Giorgi G (2018) A guided tour in constraint qualifications for nonlinear programming under differentiability assumptions. In: DEM working paper series. ISSN: 2281-1346
Grapiglia GN, Yuan Y (2021) On the complexity of an augmented Lagrangian method for nonconvex optimization. IMA J Numer Anal 41(2):1546–1568
Guo L, Ye JJ (2018) Necessary optimality conditions and exact penalization for non-Lipschitz nonlinear programs. Math Program Ser B 168:571–598
Hestenes MR (1969) Multiplier and gradient methods. J Optim Theory Appl 4(5):303–320
Hijazi HL, Bonami P, Ouorou A (2013) Robust delay-constrained routing in telecommunications. Ann Oper Res 206:163–181
Khintchine AY (1932) Mathematical theory of a stationary queue. Mat Sb 39(4):73–84
Li Z et al (2021) Rate-improved inexact augmented Lagrangian method for constrained nonconvex optimization. In: Proceedings of the 24th international conference on artificial intelligence and statistics (AISTATS) 2021, San Diego, California, USA. PMLR: Volume 130
Luenberger DG, Yase Y et al (2007) Linear and nonlinear programming, vol 2, 3rd edn. Springer, Berlin
McCormick GP (1976) Computability of global solutions to factorable nonconvex programs: part I—convex underestimating problems. Math Program 10:147–175
Murray W, Ng K (2010) An algorithm for nonlinear optimization problems with binary variables. Comput Optim Appl 47(2):257–288
Nesterov Y (2005) Smooth minimization of non-smooth functions. Math Program Ser A 103:127–152
Orlowski, S Pioro M, Tomaszewski A, Wessaly R (2007) SNDlib 1.0—survivable network design library. In: Proceedings of the 3rd international network optimization conference (INOC 2007), Spa, Belgium. http://sndlib.zib.de. Extended version accepted in Networks, 2009
Orlowski S, Wessäly R, Pióro M, Tomaszewski A (2010) SNDlib 1.0–survivable network design library. Networks 55:276–286
Papadimitriou D, Fortz B, Gorgone E (2015) Lagrangian relaxation for the time-dependent combined network design and routing problem. In: Proceedings of 2015 IEEE international conference on communications (ICC), pp 6030–6036
Pollaczek F (1930) Uber eine Aufgabe der Wahrscheinlichkeitstheorie. Math Z 32:64–100
Poss M, Raack C (2013) Affine recourse for the robust network design problem: between static and dynamic routing. Networks 61(2):180–198
Powell MJD (1969) A method for non-linear constraints in minimization problems. In: Fletcher R (ed) Optimization. Academic Press, New York, pp 283–298
Qi L, Wei Z (2000) On the constant positive linear dependence condition and its application to SQP methods. SIAM J Optim 10:963–981
Sahin FM, Eftekhari A, Alacaoglu A, Latorre F, Cevher V (2019) An inexact augmented Lagrangian framework for nonconvex optimization with nonlinear constraints. In: Proceedings of advances in neural information processing systems 32 (NeurIPS 2019), pp 13943–13955
Scholtes S (2001) Convergence properties of a regularization scheme for mathematical programs with complementarity constraints. SIAM J Optim 11(4):918–936
Stein O (2004) On constraint qualifications in nonsmooth optimization. J Optim Theory Appl 121(3):647–671
Valkonen T (2014) A primal-dual hybrid gradient method for nonlinear operators with applications to MRI. Inverse Prob 30(5):055012
Xie Y, Wright SJ (2021) Complexity of proximal augmented Lagrangian for Nonconvex optimization with nonlinear equality constraints. J Sci Comput 86(38):1–30
Yuan G, Ghanem B (2017) An exact penalty method for binary optimization based on MPEC formulation. In: Proceedings of the thirty-first AAAI conference on artificial intelligence (AAAI-17), pp 2867–2875
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Papadimitriou, D., Vũ, B.C. An augmented Lagrangian method for nonconvex composite optimization problems with nonlinear constraints. Optim Eng (2023). https://doi.org/10.1007/s11081-023-09867-z
Received:
Revised:
Accepted:
Published:
DOI: https://doi.org/10.1007/s11081-023-09867-z