In an attempt to handle some of the challenges of modern production, intel-
ligent automation systems offer solutions that are flexible, adaptive, and col-
laborative. Contrary to traditional solutions, intelligent automation systems
emerged just recently and thus lack the supporting tools and infrastructure
that traditional systems nowadays take for granted. To support efficient devel-
opment, commissioning, and control of such systems, this thesis summarizes
various lessons learned during years of implementation.
Based on what was learned, this thesis investigates key features of an in-
frastructure for modern and flexible intelligent automation systems, as well as
number of important design solutions. For example, an important question is
raised whether to decentralize the global state or to give complete access to
the main controller.
Moreover, in order to develop such systems, a framework for virtual prepa-
ration and commissioning is presented, with the main goal to offer support for
engineers. As traditional virtual commissioning solutions are not intended for
preparing highly flexible, collaborative, and dynamic systems, this framework
aims to provide some of the groundwork and point to a direction for fast and
integrated preparation and virtual commissioning of such systems.
Finally, this thesis summarizes some of the investigations made on planning
as satisfiability, in order to evaluate how different methods improve planning
performance. Throughout the thesis, an industrial material kitting use-case
exemplifies presented perspectives, lessons learned, and frameworks.