Continuous objective-based control for self-optimizing systems with changing operation modes

Self-optimization enables technical systems to adapt their behavior to varying environmental conditions and changing operation modes. Objective functions serve as evaluation criteria for the system behavior. In this paper we present a novel approach of an objective-based controller. This controller is based on Pareto sets, which are solutions of multiobjective optimization problems and computed offline. The controller can be used to continuously adjust the system behavior by computing optimal configurations out of the Pareto sets without solving an optimization problem at runtime. The goal of the controller is to drive the objective values toward a desired relative weighting. Furthermore, the controller is able to cope with varying operation modes, which arise in case of component faults, for example. For each operation mode an entire Pareto set is computed separately. For designing the objective-based controller we use a simplified model of the closed-loop system. The objective values are approximated at runtime by means of a lowpass filter. An active suspension system affected by unknown track excitations serves as an application example. We use a particular control reconfiguration approach in case of an actuator fault as a change of the operation mode. Finally, simulations with a nonlinear model of the system validate our approach.