• Summary
• Related Documents
  • Active Bibliography
  • Co-citation
• Version History

Abstract:

Graph Drawing addresses the problem of nding a layout of a graph that satises a given aesthetic objective. This problem is very important for programs that visualize diagrams such as those used in Software Engineering methodologies, Data Mining functional dependencies, Semantic and Bayesian Networks, etc. This paper is concerned with developing a general approach for Graph Drawing based on Evolutionary Algorithms. We use (1+1)-Evolution Strategies which is justied by previous work. Experimental results on 146 graphs demonstrate the algorithms capacity for solving the most important problem in the graph drawing eld: edge-crossings minimization. In addition, a preliminary study has shown how to acquire and satisfy subjective user criteria. The algorithm has shown to be fast enough what allows its use in real-world programs that visualize diagrams. Un etude du trace evolutif de graphes Resume Le trace automatique de graphes est un probleme tres important pour toutes les applications ayant besoin de visualiser des diagrammes tels ceux utilises en Genie Logiciel, en extraction de connaisances a partir de donnees, en reseaux bayesiens et semantiques, etc. L'objet de la recherche presentee est de decrire une approche generale pour le trace de graphes utilisant des algorithmes evolutifs. Nous nous sommes bases sur les travaux precedents pour choisir l'algorithme

Citations