Topology Optimization of Complex Network based on NSGA-II

Abstract

This paper proposes a complex network topology optimization method based on NSGA-II with degrees and edges number unchanged. The method selects the elite parent networks by applying non-dominated layer sorting and congestion degree calculation through the network crossover and mutation operation to achieve maximum network efficiency and network invulnerability. Simulation analysis shows that the method can significantly improve the invulnerability and network efficiency of the initial networks at the same time. Optimizing the network efficiency makes the network connection show a negative correlation. The network presents a "multi-core decentralized" structure; optimizing the network invulnerability makes the connection show a positive correlation, the network presents a "close core and scattered periphery" structure; the cooperative optimization makes the network present a "tighter core and large periphery" structure, which is proved with good theoretical generality and practical application value.