Evolving Equilibrium Policies for a Multiagent Reinforcement Learning Problem with State Attractors