摘要：AbstractThis paper proposes a real-time model predictive control (MPC) algorithm for problems with convex quadratic objectives, linear dynamic systems, as well as polytopic state-and control constraints. The method features, on the one hand, explicit model predictive control algorithms and parametric optimization, but, on the other hand, also uses ideas from iterative distributed optimization. This leads to an explicit real-time algorithm with constant and verifiable run-time bounds that scale linearly with the prediction horizon length of the MPC problem. Moreover, unlike standard explicit MPC algorithms, the memory requirements of the proposed algorithm do not dependent on the prediction horizon length. The practical advantages of the method are illustrated with a numerical example.
