An iterative refinement algorithm for the minimum branch vertices problem

  • Authors:

  • Diego M. Silva;Ricardo M. A. Silva;Geraldo R. Mateus;José F. Gonçalves;Mauricio G. C. Resende;Paola Festa

  • Affiliations:

  • Dept. of Computer Science, Federal University of Lavras, Lavras, MG, Brazil and Dept. of Computer Science, Federal University of Minas Gerais, Belo Horizonte, MG, Brazil;Dept. of Computer Science, Federal University of Lavras, Lavras, MG, Brazil and Center of Informatics, Federal University of Pernambuco, Recife, PE, Brazil;Dept. of Computer Science, Federal University of Minas Gerais, Belo Horizonte, MG, Brazil;LIAAD, Faculdade de Economia do Porto, Porto, Portugal;Internet and Network Systems Research, AT&T Labs Research, Florham Park, NJ;Dept. of Mathematics and Applications "R. Caccioppoli", University of Napoli FEDERICO II, Napoli, Italy

  • Venue:
  • SEA'11 Proceedings of the 10th international conference on Experimental algorithms
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper presents a new approach to solve the NP-complete minimumbranch vertices problem (MBV) introduced by Gargano et. al[1]. In spite of being a recently proposed problem in the network optimization literature, there are some heuristics to solve it [3]. The main contribution of this paper consists in a new heuristic based on the iterative refinement approach proposed by Deo and Kumar [2]. The experimental results suggest that this approach is capable of finding solutions that are better than the best known in the literature. In thiswork, for instance, the proposed heuristic found better solutions for 78% of the instances tested. The heuristic looks very promising for the solution of problems related with constrained spanning trees.