Adaptive cybersecurity response using Q-learning reinforcement learning to learn optimal defensive actions against simulated cyber attacks.