Stability and optimality of distributed, linear model predictive control∗ Part I: state feedback

A new framework is presented for distributed, linear model predictive control (MPC) with guaranteed nominal stability and performance properties. We first show that modeling the interactions between subsystems and exchanging trajectory information among MPCs (communication) is insufficient to provide even closed-loop stability. We next propose a cooperative distributedMPC framework, in which the objective functions of the local MPCs are modified to achieve systemwide control objectives. This approach allows practitioners to tackle large, interacting systems by building on the local MPC systems already in place. The iterations generated by the proposed distributed MPC algorithm are systemwide feasible, and the controller based on any intermediate termination of the algorithm provides nominal closed-loop stability and zero steady-state offset. If iterated to convergence, the distributed MPC algorithm achieves optimal, centralized MPC control. Three examples are presented to illustrate the benefits of the proposed distributed MPC framework.