On a General class of stochastic co-evolutionary dynamics

This paper presents a unified framework to study the co-evolution of networks and play, using the language of evolutionary game theory. We show by examples that the set-up is rich enough to encompass many recent models discussed by the literature. We completely characterize the invariant distribution of such processes and show how to calculate stochastically stable states by means of a treecharacterization algorithm. Moreover, specializing the process a bit further allows us to completely characterize the generated random graph ensemble. This new result demonstrates a new and rather general relation between random graph theory and evolutionary models with endogenous interaction structures.