Satisficing paths in multi-agent reinforcement learning allow for exploration and convergence to equilibrium strategies.