ResPubID20806, component, agent communication, reinforcement learning, multi-agent systems, complex system, learning rates, neural network