A Tabu Search Optimization Algorithm with 2.5K Null Model

The 2.5K null model has the same joint degree distribution and clustering spectrum with the original network. As a frame of reference, the design of 2.5K null model is extremely significant tool to study the clustering characteristics of the original network. The existing 2.5K null model algorithms are mainly generated based on random scrambling methods, which lead to redundant exchange problems during the exchange of edges and usually can not effectively reach the 2.5K targets for high-clustering networks. To solve these problems, we propose a tabu search optimization algorithm with 2.5K null model, which is constructed by setting weight of edges in initial network-2K null model, selecting edges to be disconnected according to the probability and limiting the newborn edges depended on tabu search. The simulation results show that the null model generated by the proposed algorithm is more accurate and effective than the existing ones.