A Probabilistic Theory Based Method for Robustness Assessment of Bipartite Networks

Abstract

Complex networks play a significant role in our daily lives. Due to the fact that complex networks, in reality, could suffer from manifold perturbations and network components could break down, it is, therefore, pivotal to investigate the robustness of complex networks in the face of perturbations. In the literature, enormous endeavors have been made to study network robustness. However, existing studies mainly deal with the interdependent networks, while little attention is paid to study the robustness of bipartite networks. Many networks, such as economical networks and biological networks, can be modeled as bipartite networks. It is therefore important to study the robustness of bipartite networks. Because the structures of bipartite networks differ from that of networks in common sense, existing robustness analysis methods cannot be applied to bipartite networks. With regard to this, this paper proposes a generic method based on the probabilistic theory to assess the robustness of bipartite networks under random node failures. The proposed method mathematically indicates that bipartite networks are robust to random perturbations. The experiments on random bipartite Erdos–Renyi and scale-free networks demonstrate that the theoretical results yielded by the proposed method coincide quite well with the simulation results. This paper also compares the proposed method against the well-studied method. The experiments indicate that the proposed method outperforms the comparison method.