New method to optimize Force-directed layouts of large graphs

Abstract

This paper describes a novel method to optimize the force-directed placement algorithm for 3D drawing of large graphs. The main idea behind our approach consists in optimizing the layout by equitably distributing vertices in space. We consider the largest sphere inscribed in the 3D space and the vertices are then assigned random initial positions that are improved by force-directed placement. In order to ensure the effectiveness of the algorithm, we propose a new energy function minimization which uses the conjugated gradient of Fletcher-Reeves. Our algorithm is not only addressed to general undirected graphs but it also produces good layouts for large trees. This work is motivated by our need to offer 3D visualization tools for large computing networks but this first phase will be focused on the graph representation.