Model Based System Development

Research Vision

The central topics driving the research carried out in the section on model-based system development (MBSD) are models and "modeling". Models provide abstractions of systems, artificial or natural, that allow reasoning about properties of these systems, ignoring extraneous details while focusing on relevant ones. Explicit models have always played a key role in science and engineering. As computer scientists we have the obligation to precisely characterize the terms we use and to develop general mathematical theories that allow us to prove particular properties of interest of a given artifact. The significance of models in computing and information science has increased over the years. Prominent, for instance, is the use of the Unified Modeling Language (UML) and the Archimate modelling language for enterprise architecture. Other examples include the use of automata in system simulation and model checking, and logical and probabilistic models in model-based diagnosis within artificial intelligence. There is a clear trend towards the systematic use of models as the primary artifacts throughout the engineering lifecycle of computer based systems. Requirements, behavior, functionality, construction and testing strategies are all described in terms of models. To facilitate construction, communication and exploitation, models are often given a graphical representation. Models are not only used to reason about a system, but also to allow all the stakeholders to participate in the development process and to communicate with each other, to simulate or test the behavior of a system, to generate implementations and to facilitate reuse. However, much research is still needed to turn the vision of Model Based Sysem Development into reality.

Research Mission

To carry out fundamental research on the use of (mathematical) models in the development of computer systems, in particular on

• model based reasoning, verification and analysis,
• model transformation,
• model construction and validation, and
• model-based application generation,

in relationship to application areas, on the one hand, and basic technologies, on the other. The effectiveness of our theories, methods and tools are empirically validated by the development of challenging applications from business, industry, government and health care, thus aiming at bridging the gap between theory and practice.

Research Strategy and Policy

History and Context

The MBSD group was created at the beginning of 2008, as part of an effort to restructure the research within ICIS in somewhat larger units. The main motivation for creating MBSD was the insight that models - and the vision of model based system development - played a key role in the research of four groups within ICIS. All these groups have a strong international reputation concerning specific aspects of model based system development:

• the Informatics for Technical Applications (ITA) group, headed by prof. Frits Vaandrager, applies computer-aided verification and analysis techniques,
• the Software Technology (ST) group, headed by prof. Rinus Plasmeijer, applies advanced functional programming and code generation techniques for model driven software development,
• the Model Based Reasoning (MBR) group, headed by dr. Peter Lucas, applies probabilistic graphical models, and
• the Theories for Enterprise Engineering (TEE) group headed by prof. Erik Proper, applies conceptual modeling to study the modeling process and the use of models as a means for communication between stakeholders.

All four groups are committed to apply their basic techniques (theories, methods and tools) to challenging practical applications:

• ITA is well-known for its works on embedded and distributed systems,
• ST and TEE are well-known for their work on workflow management systems and enterprise engineering,
• MBR is well-known for its work on medical decision support.

The MBSD research themes

Companies and organizations typically are not interested in just one specific application of models. Once they have gone through the (costly) effort of constructing models for, say, verification, they also want to use them for code generation, testing, fault-diagnosos, etc. By bringing the four groups together within an encompassing MBSD group, we have created a unique combination of research expertise with significant added value. By the integrated use of our basic techniques during the whole model based system development life cycle, we can handle practical applications more effectively.The displayed diagram shows the seven research themes represented within the MBSD group, and the three activities of model based system development that link them together.

Characteristics

Some characteristics for the research within our group are:

• The inspiration comes from practice, in the form of challenging problems.
• This inspiration subsequently forms the basis for fundamental research at an international level on basic techniques. Thus, the research approach is both society-driven and knowledge-driven.
• Software tools play an essential role within our research: formal methods are considered incomplete without tools such as compilers, theorem provers, test tools, and model checkers.
• The results of this scientific research contribute to the solution of the original industrial problems that formed its inspiration. Indeed, it is the group's ambition to have an internationally acknowledged leading position both in developing new scientific theories and in applying them in (industry, government, health care) practice, following the industry-as-laboratory approach. Two of our professors work four days per week in industry, one of our associate professors and one of our professors work four days per week at the Embedded Systems Institute, and several of our PhD students and postdocs are supported by industry or have an industrial background.
• Being able to cover a wide spectrum ranging from theory to practice is highly regarded. Much of our research is carried out in a multidisciplinary context.
• External funding is seen as an important criterion for the quality and societal relevance of the research, and as an essential structuring factor.
• Pure curiosity-driven or fundamental research (in this context in areas such as methodology, automata theory, logic) is seen as an essential part of an academic attitude.
• Influence in and outside academia is sought not only via publications in the scientific literature and joint projects, but also via contributions in non-scientific periodicals and in the press, contact with high school students, via participation in scientific steering committees (within funding bodies and program committees of conferences), and professional organizations such as the Netherlands Architecture Forum (NAF). Thus, the audience ranges from colleagues in the scientific community (including PhD students) to practitioners and managers.
• The ITA and ST subgroups of MBSD participate in the Dutch graduate school Institute for Programming research and Algorithmics (IPA). The MBR and TEE subgroups of MBSD participate in the Dutch graduate school Information and Knowledge Systems.

Basic Technologies

We will now elaborate on our research strategy concerning the four basic technologies that are being used and further developed within our group:

Application Areas

We apply our theories, methods and tools on applications selected from three different areas:

MBSD Collaborations

Since the inception of MBSD, the following joint projects between subgroups have been set up:

• Peter Lucas and Frits Vaandrager are jointly involved in the ESI project Octopus. The Octopus project adresses the issue of system adaptability in the setting of digital document printers from Océ; i.e. the ability of a system to adapt itself to fluctuations in the environment, the use of the product, etc. Within this project, the expertise of Peter Lucas on model-based diagnosis and probabilistic graphical models is complemented by the expertise of Vaandrager on model checking and real-time systems.
• Application of model checking for analysis medical guidelines
• Stijn Hoppenbrouwers and Peter Lucas have started exploring the links between enterprise modeling (EE) and knowledge acquisition (AI). In particular, they are investigating the possibility to create game-like procedures for knowledge elicitation from domain experts without skills in formal modeling. The idea is to collect structured but naturally available information (rationale patterns) and automatically derive specific formal models from these. This has led to a publication at an AI convention. Ilona Wilmont recently started a PhD project in this area, at the TEE group.
• Collaboration Rinus/Peter
• Collaboration Rinus/Erik
• Erik Proper and Frits Vaandrager are jointly involved in a FES project proposal on dynamic cooperative traffic and travel management. The aim of this research is to select and improve architectures and communication standards for mobility, and to assess the quality of their specifications. Here the expertise of Proper on architectures is complemented by the expertise of Vaandrager on embedded systems.
• Links between testing work Jan/Pieter