An Adaptive Brain Storm Optimization Based on Hierarchical Learning for Community Detection.

Community structure is an important feature of complex networks and it is essential for analyzing complex networks. In recent years, community detection based on heuristic algorithms has received much attention in various fields. To improve the effectiveness and accuracy of the algorithm in big data era, an adaptive brain storm optimization based on hierarchical learning (ABSO-HL) is proposed. Instead of the fixed probability, the adaptive probability is adopted in mutation and crossover operations of the proposed ABSO-HL. The proposed updated strategy selects one or two solutions for mutation and crossover to generate a new one. The proposed hierarchical learning strategy is used to accelerate the process by searching in the neighborhood of the new solution, and obtain the optimal partition in an efficient way. The usefulness and effectiveness of the proposed algorithm were demonstrated through a lot of experiments on both real-world and synthetic networks.