Parallel MPC for Linear Systems With State and Input Constraints

This letter proposes a parallelizable algorithm for linear-quadratic model predictive control (MPC) problems with state and input constraints. The itself is based on parallel MPC scheme that has originally been designed systems In this context, one contribution of the construction time-varying yet separable constraint margins ensuring recursive feasibility asymptotic stability sub-optimal in general setting, which also includes Moreover, it shown how to tradeoff online run-time guarantees versus conservatism introduced by tightened corresponding performance proposed method as well cost analyzed context controlling large-scale mechatronic system. illustrated numerical experiments system more than 100 states 60 inputs leading run-times millisecond range.