Hierarchical Hybrid Modeling and Control of Cognitive Manufacturing Systems

In current manufacturing systems, short times for product development and ramp-up as well as high production rates and customized products are of increasing importance. To push the efficiency (and thus the economic gain) of production further, recent investigations combine aspects of flexible manufacturing with technical cognition – the latter refers to a system’s capability to perceive a momentary situation and to pro-actively derive a production scheme that maximizes profit. Within this context, the contribution of this paper is twofold: (1) a hierarchical modeling scheme is presented which is suitable to represent dynamics depending on a perceived situation, (2) an approach to adapt production online to changes in product specifications or production constraints by combining high-level optimal plan generation with low-level learningand model-based control.