Cascading Dynamics of Invulnerable Complex Networks Based on Markov Decision

This paper proposes a new cascading failure model that introduces a transition probability matrix in Markov Decision to characterize the dynamic process of load flow. And following the inherent logic of betweenness centrality, the degree of nodes and the distance between nodes are used to define the initial load distribution of edges in the network, which satisfies the actual load migration preference and load conservation law. The model is tested on nine actual and artificial networks, the effectiveness and generalization of which have been verified. In addition, we found the intrinsic relationship between cascading failures and network capacity, and the existence of a cascading failures threshold, from which some enlightenments are obtained for the invulnerability of realistic complex networks.