Métodos de punto interior para optimización cuadrática convexa con matrices no definidas positivas

In this article a modification of the recursive algorithm of Cholesky is obtainedthat allows the factorization of Semi Definite Positive Matrices, even though theseare not positive defined, without increasing the computational cost. Thanks to thisfactorization Convex Quadratic Programming Problems a...

Descripción completa

Autores Principales: Palencia F., Gonzalo, Hing C., Rosina, Rojas C., Mariledy, Medina S., Denysde
Formato: Artículo
Idioma: Español
Publicado: 2015
Acceso en línea: http://revistas.ucr.ac.cr/index.php/matematica/article/view/284
http://hdl.handle.net/10669/12934
Sumario: In this article a modification of the recursive algorithm of Cholesky is obtainedthat allows the factorization of Semi Definite Positive Matrices, even though theseare not positive defined, without increasing the computational cost. Thanks to thisfactorization Convex Quadratic Programming Problems are transformed into SecondOrder Conical Problems, which are solved with the aid of the generalization of thePredictor-Corrector algorithm of Mehrotra for these problems. There are carried outnumeric experiments for validating the results.Keywords: convex quadratic programming, second-order cones, interior point methods.