An edge deletion model for complex networks

We propose a new random graph model–edge popularity–for the web graph and other complex networks, in which edges are deleted over time and an edge is chosen to be deleted with probability inversely proportional to the in-degree of the destination. We show that, with probability tending to one as time tends to infinity, the model generates graphs whose degree distribution follows a power law. Depending on the parameters of the model, the exponent of the power law can be any number in (2,∞).