Evolution of Cooperative Networks and the Emergence of Leadership

A generic property of biological, social and economical networks is their ability to evolve in time, creating or supressing links. We model this situation with an adaptive network of agents playing a Prisoner's Dilemma game. Each agent plays with its local neighbors, collects an aggregate payoff and imitates the strategy of its best neighbor. Furthermore we allow the agents adapt their local neighborhood according to their satisfaction level and the strategy played. Therefore each agent will have diverse environments that induces an interesting dynamics in the cooperation fraction of the whole network. In the absence of noise, a steady state is always reached, where the strategies and the neighborhoods remain stationary, and where for a wide range of parameter values, an almost full cooperative outcome is obtained. The topology of the network in these states reveals that cooperators with a large number of connections emerges. These "leaders" are shown to be very important in understanding the global stability of the final steady state. If the "leaders" are perturbated, then global cascades arise and the system oscillates between the nearly full defection network and the fully cooperative outcome, before settling again in a nearly fully cooperative outcome.