Implementable distributed model predictive control with guaranteed performance properties

This article describes an implementable distributed MPC framework with guaranteed nominal stability and performance properties. The proposed distributed MPC framework consists of three main components (i) distributed estimator (ii) centralized/distributed target calculation (iii) distributed regulator. State estimation for distributed MPC is addressed using the well established Kalman filtering framework. Disturbance models are employed to eliminate steady-state offset due to modeling errors/unmeasured disturbances. Algorithms with well defined properties are advanced for distributed target calculation and distributed regulation. Incorporation of the proposed distributed MPC framework provides a means to achieve optimal systemwide control performance employing subsystem-based MPCs