Commented University Calendar

Neural and Genetic Computing for Control of Dynamic Systems

Instructor:

Herbert Werner

Course Format:

2 hours Lecture

Period:

Summer Semester

Language:

English

Recommended Previous Knowledge:

Control Systems Theory and Design

Contents:

  • Introduction to multilayer perceptron networks
  • Nonlinear system identification using neural networks
  • Neural network based predictive control
  • Introduction to Genetic Algorithms (GA)
  • Tuning PID controllers using GA
  • Design of controllers with fixed structure using a hybrid Riccati-GA approach
  • Introduction to relevant Matlab toolboxes (Neural Network Based System Identification , Neural Network Based Control System Design , Genetic Algorithm )
  • Case studies and exercises in Matlab/Simulink

Learning Outcomes:

  • knowledge: nonlinear system identification, predictive control, fixed-structure controller synthesis
  • competence of methods: application of neural networks and evolutionary search in control engineering
  • competence of systems: nonconvex optimization in control engineering
  • social competence: communication in English

Reading Resources:

Werner, H., Lecture Notes „Neural Networks for Control of Dynamic Systems“, “Genetic Algorithms for Control”
L. Ljung "System Identification - Theory for the User" Prentice Hall, 1999
M. Norgaard, O. Ravn, N.K. Poulsen and L.K. Hansen "Neural Networks for Modelling and Control of Dynamic
Systems", Springer Verlag, London, 2003
M.T. Hagan, H.B. Demuth and M.H. Beale "Neural Network Design", Brooks Cole, 1995
Z. Michalewicz and D.B. Fogel, "How to Solve It: Modern Heuristics" (2nd Edition), Springer Verlag, Berlin

Performance Record:

oral exam

Workload:

90 hours total

Further Information:

www.tuhh.de/rts

Contact:

regelungstechnik(at)tuhh(dot)de

Credit points of this module can be found in the course plan for the corresponding course of study.

Last change: 24 Sep 2014