Title : ( A novel recurrent neural network based of NCP function for solving convex quadratic programming problems )
Authors: Sohrab Effati ,
Full TextAbstract
In this paper we propose a novel neural network model to solving linear and (convex) quadratic programming problems. The neural network model is derived based of an NCP function. In theoretical aspect, global convergence of the new model is proved. As an application, we show that the proposed model can be used directly to solve linear complementary problems with positive semidefinite matrices. The validity and transient behavior of the neural network model are demonstrated by using four numerical examples.
Keywords
, NCP function, Dynamical system, Linear Programming, Quadratic Programming, Stability, Global convergence.
@inproceedings{paperid:1009463,
author = {Effati, Sohrab},
title = {A novel recurrent neural network based of NCP function for solving convex quadratic programming problems},
booktitle = {The 2nd international Conference on Control and Optimization with Industrial Applications},
year = {2008},
location = {باکو},
keywords = {NCP function; Dynamical system; Linear Programming; Quadratic Programming;
Stability; Global convergence.},
}
%0 Conference Proceedings
%T A novel recurrent neural network based of NCP function for solving convex quadratic programming problems
%A Effati, Sohrab
%J The 2nd international Conference on Control and Optimization with Industrial Applications
%D 2008
