Model-Free Reinforcement Learning Control for Resilient Cyber-Physical Systems
A new research paper published on arXiv explores the effectiveness of model-free reinforcement learning (RL) controllers in enhancing the resilience of cyber-physical systems against cyberattacks. The study analyzes four RL reward types, finding that the Lyapunov reward offers superior resilience with minimal tracking error. Proximal Policy Optimization (PPO) demonstrated better performance than Deep Deterministic Policy Gradient (DDPG), significantly reducing key performance indicator variance. AI
IMPACT This research could lead to more robust AI-driven security for critical infrastructure and industrial control systems.