Keywords: Variant Management, Feature Models, Conceptual Modeling, Customer Requirements
Astract: The degree of variability in modern products is steadily increasing. Markets have changed in terms of customers with individual requirements and expectations on products. This brings new challenges to the process of product development. The number of possible variants increases more and more, as well as the complexity to handle variability. Therefore, there is a need for supporting information systems, when it comes to selecting the correct variants in terms of, e.g. optimal costs or fulfillment of individual requirements. Customer requirements different than those of average users create a greater difficulty in the management of product variants. This is especially a struggle when developing or evolving products in the health care domain. Persons with disabilities often have completely different requirements on products than people without do. This thesis proposes a methodology to analyze and evaluate variable products, in consideration of specific customer requirements. Techniques from Quality Function Deployment - in particular the House of Quality - and modeling techniques from variability modeling - namely Feature Modeling - are combined. Feature Models are a common variability modeling technique. The House of Quality is a method to evaluate the level of fulfillment of customer requirements by correlating them with technical characteristics. By combining concepts from both techniques, a user centric evaluation of variable products is provided. In order to prove the feasibility and applicability of this approach, a case study within the health care domain is done as proof-of-concept. Personas with disabilities are defined as well as real product configurations. The introduced approach is able to identify the product fitting best for each of the defined personas. The final step evaluates the methodology by means of a case study. This case study also revealed areas for improvement in future work. The approach was presented to and discussed with the variability modeling community at a workshop and further investigations were done by interviews with experts on variability modeling.

Keywords: Model Engineering, Standardization, UML, fUML, Enterprise Architect
Astract: The rise of Model Driven Development (MDD) has renewed the interest in the execution of models. The predominant modeling language applied in MDD is the Unified Modeling Language (UML). Unfortunately, UML lacks clear execution semantics. This has lead to a plethora of different interpretations both in academia and industry, hindering the interoperability of UML tools supporting the execution of models. A possible solution to this problem is the so-called Foundational Subset For Executable UML Models (fUML) standard published by the Object Management Group (OMG). fUML is an extension of UML, defining a standardized execution semantics for a subset of UML. fUML has, however, not yet been widely adopted in commercial tooling. This thesis investigates the integration of fUML with existing commercial UML modeling tools to contribute to the adoption of fUML and identify challenges arising in the integration. To this end, this thesis aims to answer the following research questions: 1. Can fUML be integrated into a proprietary UML model execution environment? Which challenges arise in the integration? 2. Does the standardized fUML model execution environment provide the same functionality as a proprietary UML model execution environment? 3. Is the performance of the standardized fUML model execution environment comparable to the performance of a proprietary UML model execution environment? To explore these questions, a prototypical integration of fUML into the commercial UML modeling tool Enterprise Architect (EA) has been implemented in this thesis. The goal of the prototype is to allow execution and debugging of UML state machines and sequence diagrams via EA-s execution environment, in conjunction with the execution of UML activity diagrams via the fUML execution environment. The evaluation of the prototype has shown that the integration of the execution environments has been completed successfully. The prototype is capable of executing and debugging state machines, sequence diagrams and fUML compliant activity diagrams in conjunction, while still providing the same functionality as the proprietary execution environment. The performance analysis has shown that the prototype is slower than the proprietary execution environment provided by EA. This is mostly due to the necessity of running two different execution environments in parallel.

Keywords: Comparison of Curricula, Business Informatics
Astract: Due to the interdisciplinarity of business informatics between business sciences and computer sciences there are a lot of topics for research and teaching enabling a multifarious scope for designing curricula. So a comparison of course contents in terms of benchmarking is obvious, not least because of the increasing university autonomy. This thesis deals with the question which course contents of the bachelor programme Business Informatics at the Vienna University of Technology diverge to which extent compared to contents of similar programmes at other selected universities. The structure and the teaching contents of the undergraduate studies were analyzed by pre-defined categories and a grading system for the similarity of content. Domain experts of the Vienna University of Technology reviewed the results per field of examination (Business Informatics, Business Sciences, Information Technology and Foundational Knowledge) and their feedback was worked into. Main findings of this master thesis are: There is no balance between the four domains of education with respect to the amount of work in ECTS: Information Technology and Business Informatics require an above-average amount of work, whereas the amount of work for Business Sciences and Foundational Knowledge are below-average. The individual selection of specialization areas is quite limited, their size is very low compared to the total amount of work in ECTS. Probably strengths of the curricula bachelor programme Business Informatics [TUWien11b] in terms of a great variety and depth of topics were found in the following modules: -Engineering of Web Applications- and -Modeling- (Business Informatics), -Algorithm and data structures- and -Programming- (Information Technology), -Statistics and probability theory- (Foundational Knowledge). On the other hand there are also identified weaknesses, as follows: Missing courses or lectures to -Accounting-, especially double-entry bookkeeping, -Introduction to Business Informatics-, -Information Management-, -Project Management-, -Computer Networks- -Network Protocols and Network Management-. These findings should be considered in further discussions about the design of the bachelor programme Business Informatics.

Keywords: Model Checking, Model Driven Software Development
Astract: In today's software, no matter how security and safety critical it may be, defects and failures are common. With the rising complexity of software and our growing dependency on its correct functioning as it permeates our every day life the software development process requires new approaches to integrate formal verification techniques. This thesis presents approaches on efficiently verifying software systems described by model-driven software development artifacts. These artifacts comprise the implementation of the system and consist of both structural and behavioral models. We present two model checking approaches, MocOCL and Gryphon, to verify the temporal specification of a system against its model-based implementation. Central to our approach is the twofold use of graphs to describe the system under verification. First, we describe the admissible static structure of an instance of the system by means of attributed type graphs with inheritance and containment relations, often referred to as metamodel. Second, we represent a state of the system as an object graph that enumerates a system's active objects, the references among them, and their attribute values. A change in the system, e.g., the deletion of an object or the modification of an attribute value, triggers a state change. The behavior of the system is thus described by actions that modify the state of the system. In this thesis we employ graph transformations to model such state-changing actions as they provide suitable means to formally describe modifications on graphs. The specification of the system, on the other hand, is written in our temporal extension of the Object Constraint Language (OCL) that is based on Computation Tree Logic (CTL). A specification written in our CTL extension for OCL, called cOCL, can be verified against a model-based implementation of the system with our explicit-state model checker MocOCL. Gryphon aims to increase the efficiency and scalability of the verification process and implements a symbolic model checking approach, that focuses the verification on safety specifications. The work presented in this thesis also encompasses a survey and a feature-based classification of verification approaches that can be used to verify artifacts of model-driven software development. Methodologically, it provides the motivation for our work on MocOCL and Gryphon. Both our approaches are novel in their own respect; MocOCL for its capability to verify CTL-extended OCL expressions and Gryphon for its use of relational logic to build a symbolic representation of the system that can be verified with any model checker participating in the Hardware Model Checking Competition. Finally, MocOCL and Gryphon are evaluated performance-wise on a set of three representative benchmarks that demonstrate the model checkers' preferred fields of application and its limitations.

Keywords:

Keywords:

Keywords:
Astract: The capability of UML profiles to serve as annotation mechanism has been recognized in both industry and research. With JUMP, we have presented a fully automatic approach to generate profiles from annotation-based Java libraries. We have demonstrated the practical value of JUMP by contributing profiles that facilitate reverse-engineering and forward-engineering scenarios for the Java platform. Its evaluation shows that automatically generated profiles are equal or even improved in quality compared to profiles currently used in practice.

Keywords:
Astract: REAlist is a prototypical web-based ERP system built on top of a generic REA core system. It is implemented as a multi-tenant aware software-as-a-service. We intend to give a live demo of REAlist and show how various REA concepts are used to model the required ERP artefacts.

Keywords:
Astract: The German working committee for "Industrie 4.0" identified the horizontal integration throughout value networks and the vertical integration of networked manufacturing systems as key issues in the context of smart factories. For this purpose we aim for a universal model-driven industrial engineering framework spanning over production chains and value networks. Thereby, we build up on the Resource Event Agent (REA) business ontology (ISO/IEC 15944-4) to describe external activities requiring horizontal integration with business partners and internal activities serving for vertical integration within a manufacturing enterprise. We plan to apply the ISA-95 industry standard (ANSI/ISA-95; DIN EN 62264) to describe the vertical integration within an enterprise and its decentralized, networked production plants. As a first step, presented in this paper, we extend the REA ontology by useful concepts known from ISA-95 towards an integrating modeling framework.

Keywords:
Astract: Enterprise resource planning (ERP) systems are often cumbersome to customize to a client´s needs. Model-driven approaches promise to simplify these attempts. In this work we present an ERP prototype based on the Resource-Event-Agent (REA) business ontology that follows a "model-at-runtime" approach: one may customize the ERP system during runtime by changing the underlying REA models. We are using Retail-H as the reference framework for building a retail information system (RIS). Our main contribution is the prototypical implementation of a domain agnostic REA engine that can be loaded at runtime with domain specific business models-these models can further be manipulated at runtime. On that basis we have exemplarily modeled main concepts of Retail-H in REA. Validation of the implemented components is realized by applying real business activities and requirements received from our partner, a business software solution provider.