A hierarchical distributed economic NMPC architecture based on neighboring-extremal updates

A hierarchical distributed algorithm for nonlinear optimal control is presented. The algorithm enables the change of the operating point for continuous processes in real-time by further reducing computational time especially for large-scale systems with several coupled subsystems. The algorithm is an extension of a two-layer nonlinear control architecture which provides a rigorous solution to the optimal control problem by a slow controller and a sensitivity-based update for the optimal trajectories under disturbances by a fast neighboring-extremal (NE) controller. The direct incorporation of disturbances within the fast NE controller results in a parametric QP which is split up in a set of distributed parametric QP to be solved in parallel.