2. A Stabilizing Real-Time Implementation of Nonlinear Model Predictive Control

Moritz Diehl, Rolf Findeisen, Frank AllgöwerSociety 2007 · DOI · OpenAlex

Abstract

2.1 Introduction Nonlinear model predictive control (NMPC) is a feedback control technique based on the real-time solution of optimal control problems. It has attracted significant interest over the past decades [2, 25, 34, 36], not only from theoreticians but also from practitioners. This interest is mainly motivated by the fact that NMPC allows for a wide flexibility in formulating the objective and the process model, and facilitates directly taking equality and inequality constraints on the states and inputs into account. One important precondition for the successful application of NMPC is the availability of reliable and efficient numerical dynamic optimization algorithms, since at every sampling time a nonlinear dynamic optimization problem must be solved in real time. Solving such an optimization problem efficiently and quickly is a nontrivial task; see, e.g., [3, 4, 7, 12, 15, 17, 29, 31, 35, 38, 39, 40]. Often classical offline dynamic optimization algorithms are used to solve the optimization problems arising in NMPC. This is done as fast as possible, and the solution is applied to the system once it is available. For small systems using fast computers the feedback delay due to the required computation time is often negligible if compared to the time scale of the system. In this case the computational delay can be considered as a small disturbance, which NMPC, under certain conditions, is able to reject [21, 24, 30, 37].
BibTeX
@inproceedings{diehl2007stabilizing,
  author = {Moritz Diehl and Rolf Findeisen and Frank Allgöwer},
  title = {{2. A Stabilizing Real-Time Implementation of Nonlinear Model Predictive Control}},
  booktitle = {Society},
  year = {2007},
  doi = {10.1137/1.9780898718935.ch2},
}

Loading related content...

Figure