Convergence properties of augmented Lagrangian methods for constrained global optimization
Optimization Methods & Software - THE JOINT EUROPT-OMS CONFERENCE ON OPTIMIZATION, 4-7 JULY, 2007, PRAGUE, CZECH REPUBLIC, PART I
Bilinear factorization via augmented lagrange multipliers
ECCV'10 Proceedings of the 11th European conference on Computer vision: Part IV
An augmented Lagrangian fish swarm based method for global optimization
Journal of Computational and Applied Mathematics
A parallel, asynchronous method for derivative-free nonlinear programs
ICMS'06 Proceedings of the Second international conference on Mathematical Software
Augmented Lagrangian method with nonmonotone penalty parameters for constrained optimization
Computational Optimization and Applications
Saddle points of general augmented Lagrangians for constrained nonconvex optimization
Journal of Global Optimization
An inexact restoration strategy for the globalization of the sSQP method
Computational Optimization and Applications
Hi-index | 0.00 |
We consider the global and local convergence properties of a class of augmented Lagrangian methods for solving nonlinear programming problems. In these methods, linear and more general constraints are handled in different ways. The general constraints are combined with the objective function in an augmented Lagrangian. The iteration consists of solving a sequence of subproblems; in each subproblem the augmented Lagrangian is approximately minimized in the region defined by the linear constraints. A subproblem is terminated as soon as a stopping condition is satisfied. The stopping rules that we consider here encompass practical tests used in several existing packages for linearly constrained optimization. Our algorithm also allows different penalty parameters to be associated with disjoint subsets of the general constraints. In this paper, we analyze the convergence of the sequence of iterates generated by such an algorithm and prove global and fast linear convergence as well as show that potentially troublesome penalty parameters remain bounded away from zero.