15th Triennial World Congress of the International Federation of Automatic Control
  Barcelona, 21–26 July 2002 
SYNTHESIZING MULTI-OBJECTIVE H2/H∞ DYNAMIC CONTROLLER USING EVOLUTIONARY ALGORITHMS
Gerulf K. M. Pedersen* Anders S. Langballe** Rafal Wiśniewski*
* Aalborg University, Department of Control Engineering,
Fredrik Bajers Vej 7, DK-9220 Aalborg Ø, Denmark
** TDC Tele Danmark, Digtervejen 1, DK-9200 Aalborg SV, Denmark
This paper covers the design of an Evolutionary Algorithm (EA), which should be able to synthesize a multi-objective H2/H∞ controller. It will be shown how a system can be expressed as Matrix Inequalities (MI) and these will then be used in the design of the EA. The main objective is to use EA to examine whether a multi-objective H2/H∞ controller is feasible, and if so, how the optimal multi-objective controller might be found.
Keywords: Multiobjective optimization, Optimal control, Robust control, Genetic algorithm
E-mail: goofy@control.auc.dk
Session slot T-Th-M16: Evolutionary Algorithms in Control Design/Area code 2a : Control Design