Dynamic programming and model predictive control

Presents a criterion for assessing the stability of nonlinear model predictive control (MPC). The criterion is based on the monotonicity of the MPC cost function as a function of horizon length, and is derived using dynamic programming theory in a deterministic setting.