Toward Universal Controller: Performance-Aware Self-Optimizing Reinforcement Learning for Discrete-Time Systems With Uncontrollable Factors