ARTIFICIAL INTELLIGENCE ALGORITHMS IN THE PROBLEM OF GRAPH VISUALIZATION

Graphs are used for simulation of various objects and relations between them in different areas of science and technology. At the present stage of the information technology development the necessity of applying of graphs for the complex data analysis increases. The graphs allow to represent information in a clear and easy for understanding view. That is why the issue of development of algorithms for automatic graphs placement on the plane is actual today.  The article analyzes the performance criteria of graphs visualization for various subject areas, as well as the algorithms for graphs visualization in accordance with the predetermined criteria. The authors distinguished the following criteria: the number of ribs intersections, the ribs length unification, the placement area, the number of folds, symmetry. The analysis shows that the existing algorithms of automatic graphs location work well only for certain graphs classes. When visualizing trees and acyclic graphs the standard placement algorithms give rather good picture. But rather long ribs, the extra intersections, and the superposition of ribs on the state points may appear in the pictures produced for arbitrary graphs. The article presents some results of study of the efficiency of simulated annealing algorithm applying to the issue of the graphs visualization of genetic algorithms developed by the authors. The advantage of these algorithms is that they are the versatile approaches: using them it is possible to develop graph representation in accordance with the predetermined quality criterion (or even several criteria), they can be used for any graph class.