Exact analysis of generalized degree-based percolation without memory

Abstract

We study the problem of generalized degree-based percolation without memory, where the probability of node removal depends on a non-negative function of node degree within the remaining network. We derive a set of nonlinear ordinary differential equations describing the evolution of the degree distribution during the percolation process. Employing generating function methods, we calculate various quantities of interest such as the probability of a randomly selected node belonging to the giant component, the mean cluster size, and the critical point at which a network transitions from connected to fragmented. Validation of our approach is achieved through extensive Monte Carlo simulations and numerical solutions. Our results indicate that removing highly connected nodes with small preference probabilities significantly lowers the critical point compared to random removal. Moreover, our analytical framework is applicable to random networks generated by the configuration model with any bounded degree distribution. It provides a comprehensive and systematic approach to analyze network robustness against various degree-based attack strategies, offering valuable insights into network design and protection.