Research On Complex Network Community Detection Algorithm Based On Evolutionary Multi-objective Optimization

In real life,many complex systems can be expressed as networks.Common ones are: social networks,citation networks,expert collaboration networks,and protein networks.And complex networks have many characteristics,of which community structure is one of the most important network topology attributes,occupying an increasingly important position.Deep mining and detecting the community structure of the network is of great significance.In recent years,many algorithms for community detection have been proposed,but there are defects such as low accuracy,high time complexity,slow population convergence,and easy to fall into local optimum.According to the above problems,this paper proposes three community detection algorithms,the main contents are as follows:For the traditional non-directional and non-overlapping small and medium-sized networks,based on the idea of multi-objective optimization,this paper using the Memetic framework in the evolutionary algorithm,the crossover and mutation methods of the population are improved accordingly,and a multi-objective based on multi-objective is proposed.Complex network community detection method based on adaptive Memetic algorithm.By dynamically adjusting the crossover and mutation probabilities,while maintaining the diversity of the population,the search space is reduced and the efficiency of the algorithm is improved.How to realize the detection of large-scale overlapping communities becomes a new research direction for real-life large-scale networks.Based on the idea of multi-objective optimization and the NSGA-II algorithm framework,a large-scale overlapping community detection algorithm based on node priority is proposed.The algorithm evaluates the tightness between adjacent nodes by defining a priority function,exploring potential community structures and reducing the size of the network.The algorithm improved the accuracy of large-scale overlapping community testing,and had better optimization results.Aiming at the high time complexity problem of large-scale overlapping community detection,the research direction is transferred to the research of large-scale overlapping community detection algorithm with low time complexity,and the algorithm uses the nearly linear time complexity advantage of label propagation to perform algorithm.The improvement proposes a label propagation overlapping community detection algorithm based on node intimacy.Based on the local information of the network and the module degree increment,the nodes in the network are coarsely clustered to realize the preliminary division of the nodes.Then,the algorithm defines the node affinity function for label update and selection.The algorithm effectively improved the accuracy and stability of large-scale overlapping community testing.