Publication: A trust region SQP algorithm for equality constrained parameter estimation with simple parameter bounds
Introduction
Applications
Tools
Research Groups
Workshops
Publications
   List Publications
   Advanced Search
   Info
   Add Publications
My Account
About

A trust region SQP algorithm for equality constrained parameter estimation with simple parameter bounds

- Article in a journal -
 

Area
Chemistry

Author(s)
Nikhil Arora , Lorenz T. Biegler

Published in
Computational Optimization and Applications

Year
2004

Publisher
Springer

Abstract
We describe a new algorithm for a class of parameter estimation problems, which are either unconstrained or have only equality constraints and bounds on parameters. Due to the presence of unobservable variables, parameter estimation problems may have non-unique solutions for these variables. These can also lead to singular or ill-conditioned Hessians and this may be responsible for slow or non-convergence of nonlinear programming (NLP) algorithms used to solve these problems. For this reason, we need an algorithm that leads to strong descent and converges to a stationary point. Our algorithm is based on Successive Quadratic Programming (SQP) and constrains the SQP steps in a trust region for global convergence. We consider the second-order information in three ways: quasi-Newton updates, Gauss-Newton approximation, and exact second derivatives, and we compare their performance. Finally, we provide results of tests of our algorithm on various problems from the CUTE and COPS sets

AD Tools
ADOL-C

AD Theory and Techniques
Hessian

Related Applications
- A Trust Region SQP Algorithm

BibTeX
@ARTICLE{
         Arora2004Atr,
       title = "A trust region {SQP} algorithm for equality constrained parameter estimation with
         simple parameter bounds",
       author = "Nikhil Arora and Lorenz T. Biegler",
       publisher = "Springer",
       year = "2004",
       journal = "Computational Optimization and Applications",
       volume = "28",
       pages = "51--86",
       doi = "10.1023/B:COAP.0000018879.40214.11",
       abstract = "We describe a new algorithm for a class of parameter estimation problems, which are
         either unconstrained or have only equality constraints and bounds on parameters. Due to the presence
         of unobservable variables, parameter estimation problems may have non-unique solutions for these
         variables. These can also lead to singular or ill-conditioned Hessians and this may be responsible
         for slow or non-convergence of nonlinear programming (NLP) algorithms used to solve these problems.
         For this reason, we need an algorithm that leads to strong descent and converges to a stationary
         point. Our algorithm is based on Successive Quadratic Programming (SQP) and constrains the SQP steps
         in a trust region for global convergence. We consider the second-order information in three ways:
         quasi-Newton updates, Gauss-Newton approximation, and exact second derivatives, and we compare their
         performance. Finally, we provide results of tests of our algorithm on various problems from the CUTE
         and COPS sets",
       ad_area = "Chemistry",
       ad_tools = "ADOL-C",
       ad_theotech = "Hessian"
}


back
  

Contact:
autodiff.org
Username:
Password:
(lost password)