Please use this identifier to cite or link to this item: http://hdl.handle.net/10397/106713
PIRA download icon_1.1View/Download Full Text
Title: A Newton-CG based augmented Lagrangian method for finding a second-order stationary point of nonconvex equality constrained optimization with complexity guarantees
Authors: He, C
Lu, Z
Pong, TK 
Issue Date: 2023
Source: SIAM journal on optimization, 2023, v. 33, no. 3, p. 1734-1766
Abstract: In this paper we consider finding a second-order stationary point (SOSP) of noncon-vex equality constrained optimization when a nearly feasible point is known. In particular, we firstpropose a new Newton-conjugate gradient (Newton-CG) method for finding an approximate SOSPof unconstrained optimization and show that it enjoys a substantially better complexity than theNewton-CG method in [C. W. Royer, M. O'Neill, and S. J. Wright, Math. Program., 180 (2020),pp. 451--488]. We then propose a Newton-CG based augmented Lagrangian (AL) method for find-ing an approximate SOSP of nonconvex equality constrained optimization, in which the proposedNewton-CG method is used as a subproblem solver. We show that under a generalized linear indepen-dence constraint qualification (GLICQ), our AL method enjoys a total inner iteration complexity of\widetilde \scrO (\epsilon 7/2) and an operation complexity of \widetilde \scrO (\epsilon 7/2 min\{ n, \epsilon 3/4\} ) for finding an (\epsilon , \surd \epsilon )-SOSP of non-convex equality constrained optimization with high probability, which are significantly better thanthe ones achieved by the proximal AL method in [Y. Xie and S. J. Wright, J. Sci. Comput., 86 (2021),pp. 1--30]. In addition, we show that it has a total inner iteration complexity of \widetilde \scrO (\epsilon 11/2) and anoperation complexity of \widetilde \scrO (\epsilon 11/2 min\{ n, \epsilon 5/4\} ) when the GLICQ does not hold. To the best of ourknowledge, all the complexity results obtained in this paper are new for finding an approximate SOSPof nonconvex equality constrained optimization with high probability. Preliminary numerical resultsalso demonstrate the superiority of our proposed methods over the other competing algorithms.
Keywords: Augmented Lagrangian method
Iteration complexity
Newton conjugate gradient method
Nonconvex equality constrained optimization
Operation complexity
Second-order stationary point
Publisher: Society for Industrial and Applied Mathematics Publications
Journal: SIAM journal on optimization 
ISSN: 1052-6234
EISSN: 1095-7189
DOI: 10.1137/22M1489824
Rights: Copyright © by SIAM. Unauthorized reproduction of this article is prohibited.
The following publication He, C., Lu, Z., & Pong, T. K. (2023). A Newton-CG Based Augmented Lagrangian Method for Finding a Second-Order Stationary Point of Nonconvex Equality Constrained Optimization with Complexity Guarantees. SIAM Journal on Optimization, 33(3), 1734-1766 is available at https://doi.org/10.1137/22m1489824.
Appears in Collections:Journal/Magazine Article

Files in This Item:
File Description SizeFormat 
22m1489824.pdf2.18 MBAdobe PDFView/Open
Open Access Information
Status open access
File Version Version of Record
Access
View full-text via PolyU eLinks SFX Query
Show full item record

Page views

6
Citations as of Jun 30, 2024

Downloads

4
Citations as of Jun 30, 2024

SCOPUSTM   
Citations

2
Citations as of Jun 21, 2024

WEB OF SCIENCETM
Citations

1
Citations as of Jun 27, 2024

Google ScholarTM

Check

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.