Quadratic programming algorithms for large-scale model predictive control

摘要: Abstract Quadratic programming (QP) methods are an important element in the application of model predictive control (MPC). As larger and more challenging MPC applications considered, attention needs to be focused on construction tailoring efficient QP algorithms. In this study, we tailor apply a new method, called QPSchur, large applications, such as cross directional problems paper machines. Written C++, QPSchur is object oriented implementation novel dual space, Schur complement algorithm. We compare approach three widely applied algorithms show that significantly (up two orders magnitude) than other addition, detailed simulations considered demonstrate importance flexible, along with additional features for constraint handling, warm starts partial solution.