Ефективна реалізація прискореного методу розв’язання варіаційних нерівностей
A nonlocally converging algorithm for solving variational inequalities with strongly monotone operator and convex constraints-inequalities has been constructed. The algorithm has a high rate of convergence. The method is based on a combination of the global first-order algorithm that uses an iterati...
Saved in:
| Date: | 2014 |
|---|---|
| Main Authors: | , |
| Format: | Article |
| Language: | Russian |
| Published: |
The National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute"
2014
|
| Online Access: | http://journal.iasa.kpi.ua/article/view/33326 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Journal Title: | System research and information technologies |
Institution
System research and information technologies| Summary: | A nonlocally converging algorithm for solving variational inequalities with strongly monotone operator and convex constraints-inequalities has been constructed. The algorithm has a high rate of convergence. The method is based on a combination of the global first-order algorithm that uses an iterative sequence in the space of direct variables with Newton's method of solving the Kuhn-Tucker conditions of variational inequalities in the neighborhood of the solution. The effective implementation of the proposed algorithm has been performed. The computational aspects associated with the two time-consuming sub-tasks of a presented algorithm — the quadratic programming problem and solving a system of nonlinear equations have been considered. The implementation of the method has been tested by solving the variational inequalities with a nonpotential operator. A comparative analysis of the accelerated algorithm and the first order algorithm has been performed. The high convergence of the proposed algorithm has been confirmed by the results of computational experiments. |
|---|