A set of agents located in a lattice play the Prisoners Dilemma with their neighbors during a given number of iterations. At each iteration the agents choose to cooperate or to defect depending on the strategy chosen by the most successful player of their environment.
The model is a bidimensional cellular automata with a transition rule that depends on the accumulated payoffs of the agents when they play a Prisoners Dilemma with their neighbors.
The dynamics of the game generates in many cases complex chaotic patterns.