Cognitive self-optimization for quality control loops - potentials and future challenges in research

On all company levels, control loops are a proper way to control and optimize processes. As energy and resource efficiency come into focus and the increasing wish of the customer for individualized products adds even more complexity to the control task, standard control systems are unable to meet these increased demands. In the field of Self-Optimization, researchers try to enhance these control loops with more degrees of freedom by an adaption of the inner rules and objectives of the controller and the application of cognitive systems. This paper presents some promising developments that have been made in the field of self-optimization and shows which challenges research will have to face, like the missing standardization, the complexity of multi-parameter optimization, the problem of complex and conflicting goal systems and the technical risk management for non-deterministic system behavior.