Community detection in multilayer networks plays a key role in revealing the multiple aspects of information spreading and in comprehending the relationships and interactions within and between each layer. However, most existing algorithms are prone to local optimality, and they are also difficult to extend to high-dimensional networks. To address these challenges, we propose here a multi-objective algorithm for community detection that is based on the genetic algorithm. In particular, the modularity is introduced to optimize each network layer iteratively, and the local search is combined with genetic operations to overcome local optimality. Comparative benchmarks with other algorithms on artificial and real-world networks show that the proposed algorithm performs better, especially on high-dimensional networks.

多层网络中的社区检测在揭示信息传播的多个方面以及理解每一层内部和之间的关系和相互作用方面起着关键作用。然而，现有的大多数算法都容易出现局部最优性，也很难扩展到高维网络。为了应对这些挑战，我们在这里提出了一种基于遗传算法的社区检测多目标算法。特别地，引入模块化以迭代优化每个网络层，并且将局部搜索与遗传操作相结合以克服局部最优性。在人工和现实世界网络上与其他算法的比较基准表明，所提出的算法性能更好，尤其是在高维网络上。