| As network science continues to develop,complex network systems have been widely applied in various fields of people’s lives,including power network systems,transportation systems,small-scale social networks and so on.In complex network systems,the community structure is one of the important features.Detecting the community structure in complex networks can help reveal the principles and characteristics of complex networks.Existing multi-objective community detection algorithms often suffer from high complexity,large search space,low accuracy,and difficulty in controlling mutation directions.To address these issues,this thesis proposes a multi-objective community detection algorithm based on hybrid encoding and a multi-objective community detection algorithm based on heuristic mutation.The main contents are as follows:(1)To address the issues of high time complexity and large search space caused by initialization encoding,this thesis proposes a multi-objective community detection algorithm based on hybrid encoding using the NSGA-II framework.The advantages of neighborhood representation and label representation are complementary,and bidirectional crossover is used to prevent loss of node information.Extensive experiments have shown that the proposed algorithm achieves more accurate results on both real and synthetic LFR networks.(2)To address the problem of blind mutation strategy and difficult control over mutation direction,this thesis proposes a heuristic mutation-based multi-objective community detection algorithm by redesigning the mutation operator based on existing effective information of individuals to guide the mutation process and obtain better mutated individuals.The research results show that this algorithm can accelerate the convergence speed of the population and effectively improve the accuracy of community partition results.This thesis proposes two multi-objective community detection algorithms to address the problems of existing algorithms.The proposed algorithms greatly improve the accuracy of community partitioning in complex networks and have significant practical implications for community detection in complex networks. |
