Keywords:
Astract: Research in areas of Internet of Things (IoT), Industry 4.0 (I4.0), Cyber Physical Systems (CPS), consider system-of-systems composed of hardware and software systems that seamless work together. Interoperability ranges from compatibility (systems do not disturb other systems, but no interaction between systems) to integration (systems that share a common model/worldview). Between these extremes a continuum of interoperability exists. In any case, interoperability is a model-driven approach - in contrast to data-driven or machine-learning approaches, for example. In general, models contain modules, connections, and behavior descriptions of different types. Integration and interoperability may not only be seen as (goal) states but also as processes. Changes in one system might trigger adaptation in other systems. However, in loose integration and interoperability settings there is more support for the evolution of models.

Keywords:
Astract: Nowadays, manufacturers are increasingly able to collect and analyze data from sensors on manufacturing equipment, and also from other types of machinery, such as smart meters, pipelines, delivery trucks, etc. Nevertheless, many manufacturers are not yet ready to use analytics beyond a tool to track historical performance. However, just knowing what happened and why it happened does not use the full potential of the data and is not sufficient anymore. Manufacturers need to know what happens next and what actions to take in order to get optimal results. It is a challenge to develop advanced analytics techniques including machine learning and predictive algorithms to transform data into relevant information for gaining useful insights to take appropriate action. In the proposed research we target new analytic methods and tools that make insights not only more understandable but also actionable by decision makers. The latter requires that the results of data analytics have an immediate effect on the processes of the manufacturer. Thereby, data analytics has the potential to improve industrial processes by reducing maintenance costs, avoiding equipment failures and improving business operations. Accordingly, the overall objective of this project is to develop a set of tools - including algorithms, analytic machinery, planning approaches and visualizations - for industrial process improvements based on data analytics.

Keywords:
Astract: OPC Unified Architecture (UA) is a powerful technology for modeling and instantiating domain-specific information in a standardized manner. Its initial application scenario is in the domain of automated production systems, that increasingly have to deal with variability information, (i) regarding the products being manufactured and (ii) regarding the production systems themselves. In this work we propose a non-intrusive OPC UA information model for the modeling and querying of variability information using feature models, which are a well-known paradigm in the management of software product lines. Our information model can be applied "aside" existing domain information without interfering with their internal structure.

Keywords:

Keywords: Modell, Versionierung, SYSML, Modellierung, Modellversionierung
Astract: With the increasing popularity and usage of domain specific modeling languages, model versioning is be getting more and more important in software development, especially when model-based development methods come into play. This is also an important aspect for model-based systems engineering. Models and the distributed, collaborative development of models are constantly gaining significance. SysML is a domain-specific modeling language for systems engineering. The open source development platform Eclipse and its outstanding Eclipse Modeling Framework are an excellent base for model-driven software development. The initiative „Collaborative Modeling with Eclipse“ provides a modeling platform, founded on the named technologies and the modeling tool Papyrus, that tries to enable the comfortable working with models in teams. This work tries to improve the results, retrieved from the modeling platform when comparing SysML models. From an analysis of carried out model comparisons, several weak spots in differencing and visualization of those differences were found. Through the development of appropriate extensions, the identified weak spots could be addressed.

Keywords:

Keywords:
Astract: At the faculty for Informatics, TU Wien, there is a great heterogenity among students regarding their educational requirements, especially concerning prior knowledge of programming. Nevertheless, programming is an essential part of the first year curricula of CS and Business Informatics bachelor studies and it constitutes a base, a tool or content of further courses throughout the further study progress. To support prospective students in programming already prior to their start of studies, a MOOC (massive open online course) on introductory programming has been developed at the Faculty of Informatics at TU Wien. During summer 2017 it has been successfully implemented with a limited participants group of 579 prospective students and a small group of other internal and external test persons. Ongoing evaluations have already shown that the course and its implementation as a MOOC has the significant potential to serve as an additional bridging-course offer to prospective MINT students.

Keywords:

Keywords: UML, fUML, code generation, code verification, execution semantics, model-driven engineering
Astract: The concept of abstraction is the foundation of the evolution of computer science. Complex concepts and techniques are composed to a level of abstraction on which new techniques can be built. The evolution of programming languages is no exception. The assembly language was an abstraction of binary machine code instructions, making them interpretable by humans. Up to the nineties, many programming languages required the user to manually allocate and deallocate memory - a task that is carried out by all modern languages without the knowledge of the user. Model-driven engineering (MDE) is a software development methodology that introduces models as the central element, raising the level of abstraction beyond programming languages. Software systems described by the means of modeling languages have certain advantages: they are easier to understand, better to maintain as the described system changes over time, and they can be translated to be used on different target platforms. The latter is carried out by model transformations and code generators. UML is a standardized modeling language that is widely used to express the structure and behavior of systems. The transformation of systems described by means of UML concepts into executable code has been a tedious task for many years. This is because of the lack of formal semantics, which specify how models that conform to the abstract syntax of the modeling language have to be interpreted, which required the developers of code generators to rely on individual interpretations of the meaning of the modeling concepts. This circumstance was addressed with the fUML standard published in 2011. It covers the formal specification of the semantics of a subset of UML consisting of class modeling, activity modeling and action language concepts. Along with the standard, a conforming virtual machine was introduced that allows the user to execute fUML compliant models. However, up until today, no code generator compliant to fUML has been develped impeding the automated generation of implementation artifacts from fUML models for different target platforms. The first goal of this thesis is to elaborate and implement a code generation approach that generates executable code from fUML compliant models. The main requirement of the generator is to produce code that, when executed, behaves equivalent to the fUML model from which it was generated. The second goal of this thesis is to develop a component that verifies the correctness of the code generated from an input model by comparing its execution to the execution of the model carried out by the fUML virtual machine. Both goals aim to provide support for one of the key features of MDE for fUML: the automated transformation from a higher level of abstraction to a lower level. Thereby, an increase in productivity and efficiency for users of fUML shall be achieved.

Keywords:
Astract: Resource-Event-Agent (REA) has been applied to various engineering and business domains, with a focus on transfer activities rather than transformation activities. In the context of smart manufacturing, vertical integration of IT systems (e.g., business applications and production control systems) is a key factor. In this work, we shed light on the integration of REA concepts into production environments by investigating properties of REA transformations and aligning them with concepts from an international standard for enterprise-control system integration (ISA-95).