Prisoner's Dilemma game as a model for cooperation dynamics on scale-free networks with small-world properties

Abstract

The paper presents a game-theoretic model for cooperation dynamics among individuals on a highly clustered scale-free small world-network. It outlines details on generating a network which reflects the salient properties of real social networks. To model cooperation dynamics, the iterated Prisoners' dilemma game is simulated on the network. It is verified that the optimal outcome is for all individuals to cooperate and that when the temptation to defect exceeds a certain critical value, cooperation ceases. However it is possible to observe cooperative behavior even when the temptation to defect is beyond the critical value by allowing for some irrational moves by the players. This unexpected rise in cooperative behavior can be attributed to an avalanche effect when neighboring individuals imitate a random cooperate strategy. It is also shown that the addition of shortcut in the network further enhances this effect.