Implementation and Comparison of Heuristics for the Vertex Cover Problem on Huge Graphs

Eric Angel (1) , Romain Campigotto (2) , Christian Laforest (3)

1 IBISC - Informatique, Biologie Intégrative et Systèmes Complexes
2 LAMSADE - Laboratoire d'analyse et modélisation de systèmes pour l'aide à la décision
3 LIMOS - Laboratoire d'Informatique, de Modélisation et d'optimisation des Systèmes

We present in this paper an experimental study of six heuristics for a well-studied NP-complete graph problem: the vertex cover. These algorithms are adapted to process huge graphs. Indeed, executed on a current laptop computer, they offer reasonable CPU running times (between twenty seconds and eight hours) on graphs for which sizes are between 200 *106 and 100 *109 vertices and edges. We have run algorithms on specific graph families (we propose generators) and also on random power law graphs. Some of these heuristics can produce good solutions. We give here a comparison and an analysis of results obtained on several instances, in terms of quality of solutions and complexity, including running times.

Mots clés: implementation of algorithms; experimental analysis; huge graphs; low memory; vertex cover

Domaines: Recherche opérationnelle [math.OC]

Eric Angel, Romain Campigotto, Christian Laforest. Implementation and Comparison of Heuristics for the Vertex Cover Problem on Huge Graphs. 11th International Symposium on Experimental Algorithm (SEA 2012), Jun 2012, Bordeaux, France. pp.39--50, ⟨10.1007/978-3-642-30850-5_5⟩. ⟨hal-00741605⟩