Numerical experiments with nonconvex MINLP problems

• Autor(es): Costa, M. Fernanda P. ; Fernandes, Florbela P. ; Fernandes, Edite Manuela da G. P.
• Data: 2010
• Identificador Persistente: http://hdl.handle.net/1822/16506
• Origem: RepositóriUM - Universidade do Minho
• Assunto(s): Mixed-integer programming; Branch-and-bound; Stochastic method
• Tipo de Documento: Documento de conferência
• Idioma: Inglês

Descrição

We present a methodology to solve nonconvex Mixed-Integer Nonlinear Programming problems, that combines the Branch-and-Bound and simulated annealing type methods, which was implemented in MATLAB. A set of benchmark functions with simple bounds and different dimensions was used to analyze its practical behaviour. We exhibit computational results showing the good performance of the method.