Environment Parameter Gradient Theorem for Co-Design in Reinforcement Learning

Amber Srivastava

Reinforcement learning (RL) is traditionally concerned with learning a control policy for a fixed environment. In many engineering systems, however, the environment itself is alter…