A faster algorithm to calculate centrality based on Shortest Path Layer

Network centrality is an important index to measure the centrality of a node in a network. Calculating node centrality parameters is difficult, as networks become Large-Scale, it is essential to find out new algorithms to calculate the centrality of the more complex and bigger networks. Motivated by this trend, a new data structure and faster algorithm for nodes' centrality is provided in this paper. Our algorithm is based on the foundation of the Shortest Path Layer, we used this algorithm to calculate degree centrality, betweenness centrality and closeness centrality for unweighted networks simultaneously. To test its calculation speed, we did 540 simulation experiments using Ulrik Brandes' algorithm, using SPL without evolution and SPL with evolution, respectively. Results show that our structure and algorithm can improve calculation speed by at least 170.43% compared with Brandes' algorithm, and margin of betweenness result error is less than 10¹⁰.