Achieving Cooperation Through Multi agent Reinforcement Learning In Iterated Prisoner's Dilemma

Nowadays, the prisoner’s dilemma is one of the primary and important issues in game theory. In this dilemma, there is a Nash Equilibrium, and if the agents behave rationally, they play at point; For this purpose, the agents choose defection between the two actions of cooperation and defection to achieve greater profit. However there is a better point for the agents than the Nash Equilibrium, it is that both agents choose the cooperation. However there is a better point for the agents than the Nash Equilibrium, it is that both agents choose the cooperation. Therefore, in order to increase the rate of cooperation of the agents, the prisoner's dilemma has been considered as iterated prisoner's dilemma with a reinforcement learning approach. The results of the article show that the desired approach let has increased the rate of cooperation of the agents, and if one agent choose the cooperation, the other agent also chooses cooperation and vice versa.