Sufficient Conditions

Exploiting Linear Models for Model-Free Nonlinear Control: A Provably Convergent Policy Gradient Approach