Improving the Quality of Clustered Graph Drawing through a Dummy Element Approach

Abstract

Clustered graphs have been widely consisted in Graph Drawing to overcome the problem of drawing large (or huge) graphs with thousands, or perhaps millions of nodes. Force-directed algorithm is one of the available approaches to draw such graphs with good layouts. However, the force directed methods are often produce very different layout outcomes based on different clustering algorithms that greatly affect the quality of the final result. In this paper, we propose a 'dummy element' approach for drawing clustered graphs by using the traditional force-directed algorithm. Our approach attempts to maximize the satisfaction of aesthetics in graph drawing. The experimental results shown that our new method can reduce the side effects on choosing different clustering methods and improve the quality of final layouts.