DESEMPE ̃NO COMPUTACIONAL DE ESTRATEGIAS H ́IBRIDAS PARA LA SOLUCI ́ON DE PROBLEMAS CUADR ́ATICOS NO CONVEXOS CON RESTRICCIONES DE CAJA

Authors

  • Ridelio Miranda Pérez Universidad de la Habana
  • Sira Allende Alonso Universidad de la Habana
  • Boris Pérez Cañedo Universidad de Cienfuegos
  • Gemayqzel Bouza Allende Universidad de la Habana

Keywords:

Non-convex quadratic programming, Parametric optimization, Branch-and-Bound, Double non negative relaxation, Semidefinite programming

Abstract

In this paper the minimization of a non-convex quadratic function under box constraints is con- sidered. For solving this problem, we propose a hybrid strategy combining a cut and continuation method with a Branch and Bound procedure, based on double non-negative relaxation. We compare the results of the new approach with the Branch and Bound method and with the preceding hybrid algorithm.The influence of characteristics of the problem in the algorithm’s behavior is explored. Computational experiments include problems of higher dimension than the ones reported in the re- ferred literature and show that the new strategy improves the results concern to number of explored nodes and computation time.

Downloads

Download data is not yet available.

Published

2023-04-12

How to Cite

Miranda Pérez, R., Allende Alonso, S., Pérez Cañedo, B., & Bouza Allende, G. (2023). DESEMPE ̃NO COMPUTACIONAL DE ESTRATEGIAS H ́IBRIDAS PARA LA SOLUCI ́ON DE PROBLEMAS CUADR ́ATICOS NO CONVEXOS CON RESTRICCIONES DE CAJA. Investigación Operacional, 39(1). Retrieved from https://revistas.uh.cu/invoperacional/article/view/4065

Most read articles by the same author(s)

1 2 3 4 5 6 > >>