Imperial College London
Portrait Picture

ProfessorEricKerrigan

Faculty of Engineering, Department of Electrical and Electronic Engineering

Professor of Control and Optimization
 
 
 
//

Contact

 

+44 (0)20 7594 6343e.kerrigan Website

 
 
//

Assistant

 

Mrs Raluca Reynolds +44 (0)20 7594 6281

 
//

Location

 

1114Electrical EngineeringSouth Kensington Campus

//

Summary

 

Publications

Citation

BibTex format

@unpublished{Neuenhofen:2020,
author = {Neuenhofen, M and Kerrigan, E},
publisher = {arXiv},
title = {A modified augmented lagrangian method for problems with inconsistentconstraints},
url = {http://arxiv.org/abs/2012.10673v1},
year = {2020}
}
Download

RIS format (EndNote, RefMan)

TY  - UNPB
AB  - We present a numerical method for the minimization of objectives that areaugmented with linear inequality constraints and large quadratic penalties ofover-determined inconsistent equality constraints. Such objectives arise fromquadratic integral penalty methods for the direct transcription of optimalcontrol problems.  The Augmented Lagrangian Method (ALM) has a number of advantages over theQuadratic Penalty Method (QPM) for solving this class of problems. However, ifthe equality constraints are inconsistent, then ALM might not converge to apoint that minimizes the %unconstrained bias of the objective and penalty term.Therefore, in this paper we show a modification of ALM that fits our purpose.  We prove convergence of the modified method and prove under local uniquenessassumptions that the local rate of convergence of the modified method ingeneral exceeds the one of the unmodified method.  Numerical experiments demonstrate that the modified ALM can minimize certainquadratic penalty-augmented functions faster than QPM, whereas the unmodifiedALM converges to a minimizer of a significantly different problem.
AU  - Neuenhofen,M
AU  - Kerrigan,E
PB  - arXiv
PY  - 2020///
TI  - A modified augmented lagrangian method for problems with inconsistentconstraints
UR  - http://arxiv.org/abs/2012.10673v1
UR  - http://hdl.handle.net/10044/1/90958
ER  -
Download