OPTEC - K.U.Leuven Center of Excellence: Optimization in Engineering

When linear model predictive control (MPC) shall be applied in areas like e.g. the automotive industry, solution times in the order of milli- or even microseconds are required. One particular popular approach, explicit MPC, requires a precomputation of all possible solutions of a parametric quadratic programming problem. This, however, becomes prohibitive for nontrivial problem dimensions. Thus the development of fast online quadratic programming (QP) solvers becomes necessary and a challenging task.

In our opinion, it is not sufficient just to apply fast offline QP solvers to problems arising in MPC applications. Instead, fast online QP solvers must take into account the special structure of the problems. Normally, the problems within the MPC context do not differ much from one QP to the next making it essential to incorporate this knowledge into efficient solvers.

The Online QP Benchmark Collection aims at providing benchmark QPs from several MPC applications that comprise a whole sequence of neighbouring QPs (instead of stand-alone QP test problems). We focus on the mathematical task to solve such online QP efficiently leaving the modelling process away. Thus the QPs may seem to appear out of the box but we affirm that they all stem from a proper optimisation problem formulation.

Please send remarks and questions to This e-mail address is being protected from spambots. You need JavaScript enabled to view it .
Last update: 26/06/2008