Discrete objective-based control for self-optimizing systems

In this contribution we present a new, objective-based controller which works on the basis of a Pareto set, that is the solution of a multiobjective optimization problem. The controller can be used for self-optimizing systems to adapt the system behavior in terms of the desired objectives in spite of unknown disturbances over time. New system configurations are computed, that fit best to the current situation. The objective-based control is realized by means of a discrete controller which operates in a large sample time due to objective evaluation. Parametric reduced models are used to shorten the simulation time during controller design. An active suspension system of a railbound vehicle serves as an application example. The controller is tested with different linear models of the suspension system and is finally applied to a Hardware-in-the-Loop test rig.