Efficient Visual Navigation- A Study by the Example of Hierarchically Structured Graphs

Visual navigation of hierarchically structured graphs is a technique forinteractively exploring large graphs with an additional hierarchical structuresuch as call graphs of telephone networks. In order to reduce the complexityof the graph only the interesting subgraphs are shown in detail while theothers are collapsed. As the areas of interest change over time clusters areexpanded or contracted for which an efficient data structure is needed. Inthis work a new variant of this problem is introduced which supportsinsertion and deletion of graph nodes. So far efficient data structures wereknown only for more static variants. The proposed data structure builds on anovel technique of superimposing a search tree over an ordered list maintenancestructure. For visualizing the expanding and contracting a localupdate scheme for the well-known algorithm of Sugiyama and Misue ispresented which has applications ranging from UML diagrams to biochemicalpathways. Experimental results show that this update scheme ismore efficient than redrawing yet the performance gain is not at the expenseof quality as regards the area of the drawing and the number of crossings.