A new trust-region method for solving systems of equalities and inequalities

In this paper, we propose a new trust-region method for solving nonlinear systems of equalities and inequalities. The algorithm combines both standard and adaptive trust-region frameworks to construct the steps of the algorithm. The trust-region subproblem is solved in the first iteration using a gi...

Full description

Saved in:
Bibliographic Details
Published in:Computational & applied mathematics 2017-03, Vol.36 (1), p.769-790
Main Authors: Saeidian, Z., Reza Peyghami, M., Habibi, M., Ghasemi, S.
Format: Article
Language:English
Subjects:
Adaptive algorithms
Adaptive systems
Applications of Mathematics
Applied physics
Computational mathematics
Computational Mathematics and Numerical Analysis
Construction standards
Inequalities
Iterative methods
Mathematical Applications in Computer Science
Mathematical Applications in the Physical Sciences
Mathematics
Mathematics and Statistics
Nonlinear systems
Optimization algorithms
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, we propose a new trust-region method for solving nonlinear systems of equalities and inequalities. The algorithm combines both standard and adaptive trust-region frameworks to construct the steps of the algorithm. The trust-region subproblem is solved in the first iteration using a given initial radius. Then, in each iteration, the standard trust-region method is followed whenever the current trial step is accepted, otherwise, the subproblem is resolved using an adaptive scheme. The convergence results for the new proposed algorithm are established under some mild and standard assumptions. Numerical results on some least-squares test problems show the efficiency and effectiveness of the proposed algorithm in practice too.
ISSN:0101-8205
2238-3603
1807-0302
DOI:10.1007/s40314-015-0257-9