Issue |
RAIRO-Oper. Res.
Volume 35, Number 2, April June 2001
ROADEF'99
|
|
---|---|---|
Page(s) | 211 - 228 | |
DOI | https://doi.org/10.1051/ro:2001112 | |
Published online | 15 August 2002 |
Semi-Definite positive Programming Relaxations for Graph Kn-Coloring in Frequency Assignment
1
Projet NUMOPT,
INRIA Rhône-Alpes, 655 avenue de l'Europe,
38330 Montbonnot-Saint-Martin, France; Philippe.Meurdesoif@inrialpes.fr.
2
Direction des Technologies Nouvelles,
Bouygues, 1 avenue Eugène Freyssinet, 78061
Saint-Quentin-en-Yvelines, France; brottemb@challenger.bouygues.fr.
Received:
May
1999
In this paper we will describe a new class of coloring problems, arising from military frequency assignment, where we want to minimize the number of distinct n-uples of colors used to color a given set of n-complete-subgraphs of a graph. We will propose two relaxations based on Semi-Definite Programming models for graph and hypergraph coloring, to approximate those (generally) NP-hard problems, as well as a generalization of the works of Karger et al. for hypergraph coloring, to find good feasible solutions with a probabilistic approach.
Résumé
Dans cet article, nous décrivons une nouvelle classe de problèmes de coloration rencontrés en Allocation de Fréquences militaire : nous voulons minimiser le nombre de n-uplets distincts utilisés pour colorier un ensemble doné de n-cliques d'un graphe. Pour approcher ces problèmes généralement NP-difficiles, nous proposons deux relaxations basées sur les modélisations semi-définies de la coloration de graphes et d'hypergraphes, ainsi qu'une généralisation des travaux de Karger et al. à la coloration d'hypergraphes, pour trouver de bonnes solutions faisables par une approche probabiliste.
Key words: Discrete optimization / semidefinite programming frequency assignment / graph coloring / hypergraph coloring.
© EDP Sciences, 2001
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.