Model-Driven Policy Refinement

Start date: 01.01.2008
Funded by: Universität Augsburg
Local head of project: Dr. Raphael Romeikat
Local scientists: Prof. Dr. Bernhard Bauer


Development of complex systems requires experts from different domains to collaborate with each other. Development processes usually start with specifying high-level system requirements which must be refined and implemented at lower levels. Coordination of business experts, domain experts, and technicians to align high-level system behavior and low-level implementation is a time-consuming task due to different background, vocabulary, and domain expertise. This project improves the collaboration of experts in the development process. For this purpose, an approach to system development is developed that is based on policy-based and model-driven techniques. System behavior is first modeled using abstract policies on the business layer, which are then refined step-by-step across various abstraction layers down to a technical layer, where they are finally represented in a machine-executable way. During the refinement process, information modeled at a certain layer is reused at lower layers using model transformations. The application of model-driven techniques brings benefits of MDSD to the policy domain. It simplifies the policy refinement process and at the same time increases the degree of automation.