A NETWORK CONTRACTION METHOD BASED ON GOMORY-HU TREE WITH NODE AGGREGATION TO COMMUNITY CENTROID

Abstract

In recent years, the data available for road network analysis has become more sophisticated and detailed, backed by improved computer performance and observation technology. In Japan, road network data with high resolution is also available, and the computational cost for analysis is an issue. In this study, a simple method for network reduction is developed for large-scale networks. Specifically, we assume the existence of a community core node, such as a centroid, which is set as a representative point of demand generation concentration in each community, and aim to shrink the entire network by aggregating the nodes in the community. The Gomory-Hu tree is used to identify communities, giving a partition such that the boundaries between communities have the smallest capacity. The proposed method is applied to the North Sea road network to check its behaviour.