Emerging cooperation in the prisoner's dilemma on dynamic networks

We consider the classical prisoner's dilemma being played repeatedly on a dynamic network, where agents may choose their actions as well as their co-players. Agents act profit-maximizing, fully rationally and base their decision only on local information. Individual decisions are made such that links to defecting agents are resolved and such that cooperating agents build up links, as we assume that new interrelations are established via a process of recommendation. Unlike in networks with fixed topology, the network dynamics allows for a higher profit when choosing cooperation rather than defection, thus resolving the dilemma. In the case of perfect synchronization of strategy updates, the system exhibits oscillatory dynamics of overall cooperation and linkage. The cyclical behavior is lost and the system is stabilized at a predominant level of overall cooperation when players are updating strategies in a heterogeneous way; stable cooperative structures then form. Between the highly synchronized and the low-frequency regime, we find a region where network structure resembles 'complex' hierarchical topology and the level of total cooperation in the system reaches a maximum