Identifying critical traffic jam areas with node centralities interference and robustness
Carlo Laudanna, Giovanni Scardoni · Networks and Heterogeneous Media · 2012
We introduce the notions of centrality interference and centrality robustness, as measures of variation ofcentrality values when the structure of a network is modified by removing oradding individual nodes from/to anetwork. Centrality analysis allows categorizing nodes according to theirtopological relevance in a network. Thus, centrality interference analysisallows understanding which parts of a network are mostly influenced by anode and, conversely, centrality robustness allowsquantifying the functional dependency of a node from other nodes in thenetwork. We examine the theoretical significance of these measures and apply them toclassify nodes in a road network to predict the effects on the traffic jamdue to variations in the structure of the network.In these case the interference analysis allows to predict which are the distinct regions of thenetwork affected by the function of different nodes.Such notions, when applied to a variety of different contexts, opens newperspectives in network analysis since they allow predicting the effectsof local network modifications on single node as well as global networkfunctionality.