A Semismooth Newton based Augmented Lagrangian Method for Nonsmooth Optimization on Matrix Manifolds
arXiv:2103.02855
Abstract
This paper is devoted to studying an augmented Lagrangian method for solving a class of manifold optimization problems, which have nonsmooth objective functions and nonlinear constraints. Under the constant positive linear dependence condition on manifolds, we show that the proposed method converges to a stationary point of the nonsmooth manifold optimization problem. Moreover, we propose a globalized semismooth Newton method to solve the augmented Lagrangian subproblem on manifolds efficiently. The local superlinear convergence of the manifold semismooth Newton method is also established under some suitable conditions. We also prove that the semismoothness on submanifolds can be inherited from that in the ambient manifold. Finally, numerical experiments on compressed modes and (constrained) sparse principal component analysis illustrate the advantages of the proposed method.
arXiv commentWe moved the technical proofs of lemmas into Appendix
References in corpus (5)
- Compressed Modes for Variational Problems in Mathematics and Physics
- A Second Order Non-Smooth Variational Model for Restoring Manifold-Valued Images
- Riemannian SVRG: Fast Stochastic Optimization on Riemannian Manifolds
- An inexact augmented Lagrangian method for nonsmooth optimization on Riemannian manifold
- A Global Version of the Newton Method for Finding a Singularity of the Nonsmooth Vector Fields on Riemannian Manifolds