Nash Equilibrium: Better Strategy for Agents Coordination

Intelligent agents operate in an environment that can be mostly not dynamic. This requires learning by trial and error in order to know better the environment and reach their goals. They also need cooperation among themselves. We use reinforcement learning and game theory techniques to discuss how intelligent agents learn and cooperate. As known, one of the properties of agents is that they are social. They must therefore be cooperative in the social environment where they are. This is possible by learning the environment facts by doing, sharing instantaneous information and learned knowledge. The cooperative agents will perform better than independent agents. What is the advantage of such cooperation? As an example, in this paper we shall especially show how cooperative trading agents can maximize their profits due to a good coordination by playing Nash equilibrium to ensure that each agent chooses the best strategy which gives a good payoff.