Abstract EN:

In the last years, strategies to improve or assure the quality in information systems have been addressed by several approaches. In practice, however, each application domain developed its own quality management procedure providing a specific vision of quality as well as a suite of tools to solve quality problems. Still, the solutions provided by such approaches are not sufficient to deal with broader user's requirements. Moreover, they have their own quality models, terminology and access patterns, which makes interoperability between ther a technical challenge. Aiming at filling this gap, this thesis proposes an approach whose main objective is to facilitate the definition of appropriate quality metrics and measurement methods tailored to specific quality organization. The quality in information systems also includes the exploitation of the quality information in specific tasks, e. G. , the retieval of objects satisfying a set of criteria (some of them related to quality attributes). In this context, the problem of process retrieval gained special attention due to the investment of organizations on process management practices and due to the consequent growth of the repositories of business processes. Current approaches for process retrieval propose strategies for handling processes with reasonable size and define ranking metrics that improve the relevance of the answers. However, these works address only the structural representation of the process and do not properly deal with the non-functional aspects at the different granularity levels. Motivated by these problems, this thesis also proposes an approach for evaluating quality preferences in process model matching.