Multi-objective control systems design with criteria reduction

Abstract

Control systems design may be based on many criteria. These optimization problems are nonconvex, therefore evolutionary multi-objective optimization algorithms (EMOA) are methods of choice. In engineering design problems it is desirable to find the one solution only as in single criterion optimisation. We describe a new method based on reduction of objectives while keeping relevant Pareto sets changes bounded. In the illustrative control design six objectives from optimal control, mixed norm robust optimization and standard control methods are reduced to three, which enables visualisation of the Pareto front. © 2010 Springer-Verlag.