Multi-objective disintegration of multilayer networks

Abstract

The multiple relationships between nodes in complex systems or the dependencies between multiple subsystems can be effectively represented by multilayer networks. The network disintegration problem seeks to identify a set of nodes whose removal can minimize network performance. Existing research on the disintegration of multilayer networks often focuses on overall network connectivity. This study examines the multi-objective disintegration problem of multilayer networks to find groups of nodes that maximize damage to different layers. The multi-objective optimization model is established, and the nondominated sorting genetic algorithm is improved to solve it. The search efficiency for approximating the Pareto front is enhanced by innovating initial population generation and crossover operation coding methods. Additionally, we employ the technique for order preference by similarity to ideal solution to assess the effectiveness of various multi-objective disintegration strategies. Experiments in the model and real multilayer networks show that the relative optimal strategy in the Pareto solution set can effectively balance the disintegration effect of different layers. This research offers valuable insight into safeguarding infrastructure systems and controlling disease transmission.