A Unifying Complexity Certification Framework for Active-Set Methods for Convex Quadratic Programming

In model-predictive control (MPC), an optimization problem has to be solved at each time step, which in real-time applications makes it important solve these efficiently and have good upper bounds on worst-case solution time. Often for linear MPC problems, the question is a quadratic program (QP) that depends parameters such as system states reference signals. A popular class of methods solving QPs active-set methods, where sequence systems equations solved. We propose algorithm computing subproblems will solve, every parameter interest. These sequences can used set how many iterations, floating-point operations, and, ultimately, maximum requires converge. The usefulness proposed method illustrated originating from by exact number iterations primal dual algorithms require reach optimality.