基于免疫离散差分进化算法的复杂网络社区发现

针对复杂网络社区发现问题,在标准差分进化算法的框架下,提出一种新型免疫离散差分进化算法(Immune discrete differential evolution, IDDE).该算法通过标签传播策略生成初始种群,采用离散差分进化策略来保证种群在问题空间的全局搜索能力,同时对种群中的优秀个体执行针对性的高频克隆变异操作,以提高算法的局部开发能力,改善算法的收敛性能.在计算机生成网络与真实世界网络中的仿真实验结果表明:IDDE算法具有较强的寻优性能与鲁棒性,能够有效探测复杂网络中存在的社区结构.

Community Detection in Complex Networks Using Immune Discrete Differential Evolution Algorithm

Aimed at the existing problem of community detection in complex networks, a novel immune discrete differential evolution(IDDE) is proposed in the framework of standard differential evolution. In the proposed method, the initial population is generated through label propagation, and the discrete differential evolution strategy is utilized to ensure the global searching ability of the IDDE; meanwhile, the high-frequency clonal selection mutation operation is applied to excellent individuals of the population to improve the local exploitation ability and the convergence performance of the IDDE. Artificial networks and several real networks are employed to test the performance of the IDDE, and the testing results show that the IDDE achieves better searching ability and stronger robustness, and that it can detect the community structure in complex networks effectively.