Weighted Betweenness for Multipath Networks

International audienceTypical betweenness centrality metrics neglect thepotential contribution of nodes that are near but not exactlyon shortest paths. The idea of this paper is to give morevalue to these nodes. We propose a weighted betweennesscentrality, a novel metric that assigns weights to nodes basedon the stretch of the paths they intermediate against theshortest paths. We compare the proposed metric with thetraditional and the distance-scaled betweenness metrics usingfour different network datasets. Results show that the weightedbetweenness centrality pinpoints and promotes nodes that areunderestimated by typical metrics, which can help to avoidnetwork disconnections and better exploit multipath protocols