Evolutionary Large‐Scale Multi‐Objective Optimization for Community Detection in Complex Networks

This chapter presents two ideas of enhancing the scalability of multi‐objective evolutionary algorithms (MOEAs) in solving the community detection problems for large‐scale complex networks, and two MOEAs, namely, network reduction for multi‐objective evolutionary algorithms (RMOEA) and the parallel framework for multi‐objective evolutionary algorithms (PMOEAs). The RMOEA utilizes the local communities detected during the evolution to reduce the size of complex networks, so that the scalability of MOEA to large‐scale complex networks can be enhanced. The PMOEA is based on the idea of parallelization to tackle large‐scale complex networks, where the community detection of the entire network is decomposed into detecting the communities associated with a number of key nodes performed in parallel. Although the reduction and parallelization are widely used for large‐scale optimization, there are many other technologies to deal with large‐scale optimization, such as decomposition and search enhancement.