| Summary: | Describing a structured population by means of a graph has been proven an invaluable tool for obtaining complex features of evolutionary processes. Notably, progress has been made in using network models to provide a theoretical basis for the emergence of cooperative behavior. As a price paid for the higher complexity, the resulting models are typically not amenable to a complete analytic treatment and require non-rigorous approximation schemes to be described. Another
conceptual limitation is the rigidity of using a fixed network structure in describing processes that are dynamic in nature. In this thesis, we develop a general framework to produce evolutionary graph processes where the network structure is randomized by the introduction of a reshuffling step conditioned to the value of chosen graph parameters, which can encompass clustering features of the community. The resulting models are dynamic but considerably less complex, allowing for better
rigorous analytical descriptions and lower numerical simulation times. We apply this framework in the study of the evolution of cooperation, demonstrating the fundamental role that clustering has in allowing cooperative behavior to survive.--Author's abstract
|
|---|
