A Speedup Spatial Rearrangement Algorithm for Dynamic Network Visualization

In this paper, we propose a speedup spatial rearrangement algorithm for visualizing evolutionary process while preserving mental map of previous layout. Hierarchical structure of network is extracted in data preprocessing stage, which can help to find the minimum subnets with topological changes. Multi-constraint force directed algorithm is applied to changed subnets. Evolution parameter is used to restrain the movement of existing nodes to be consistent with their previous positions. Energy parameter, on the other hand, is used to get rid of the constrain of previous position and achieve aesthetic layout. After rearrange all changed subnets, A time-varying stress model is introduced to adjust the position of all other subnets from bottom to top, which can avoid recalculating new positions for all nodes and thus improve efficiency.