model predictive control

Formulating and implementing receding-horizon optimal-control strategies (including stochastic and decentralized variants) to generate online feedback policies or trajectories that can be integrated with manipulation tasks, RL buffers, or closed-loop selection processes.