GLAD: a simple adaptive strategy that yields cooperation in dilemma games

We introduce a simple adaptive rule where agents choose a cooperative effort on a grid. Agents can adjust this effort step by step and G ains and L osses A djust D irections. We show that this process converges to the cooperative outcome in a two-person Prisoners' Dilemma game, and we provide simulations showing that the results also holds with a larger number of agents.