Machine Learning Surrogate Models are used to optimize the automated assembly, as well as the layout of components of a product in early stages of the product design process. An electrolyzer is used to validate the developed concepts.

In practice, the integration of heterogeneous IoT components is often a complex and confusing process. The different ways of representing information about individual assets create the desire for a standardized representation. The asset administration shell creates a digital exchange format for this purpose with the potential to significantly increase the degree of automation in the integration of IoT components.