Properties of Prisoner´s Dilemma Game based on genetic algorithm

Evolutionary games are widely studied in the fields of biology, sociology, economics and informatics. People´s interests focus on how cooperation emerged in a system that individuals are selfish. We studied the iterated Prisoner´s Dilemma Game evolved on BA network with genetic algorithm. Nodes in network could remember the game historical strategies and coded into gene. We show that memory length has gentle effect on cooperation level, and that some characteristic genes and frequently used genes emerge after some generations. We also find that cooperate strategy distributed on different degree-nodes equably. These findings give new perspectives to the illustration of cooperation behavior in biological group and society.