The first papers reporting the model-driven tools for (semi)automatic data model design were published in the mid-1990s. Although the idea of model-driven design of the data model is almost 30 years old, the fully automatic model-driven synthesis of the data model (MDSDM) is still the subject of intensive research. In the existing literature there are only a small number of papers presenting the implemented automatic model- driven generator of the target data model with the corresponding evaluation results, while the great majority of papers only present modest achievements in (semi)automated, or even manual, data model synthesis. The main objective of this project is automatic business-process-model-driven data model synthesis, i.e. automatic synthesis of the target data model (conceptual database model - CDM) based on a collection of models representing business processes of an enterprise.
The first ideas, and the corresponding prototype implementation called ADBdesign, were presented in 2010. The first implementation was based on business process models (BPMs) represented by UML activity diagrams, while the target CDM was represented by the UML class diagram. The initial rules for automatic CDM synthesis were covering the automatic synthesis of classes and their associations for di erent types of participants and business objects used in a single business process. The set of formal rules enabling automatic CDM synthesis has been additionally amended and applied to collaborative BPMs represented by BPMN. The most recent achievements are related to the transition from direct to indirect two-phase MDSDM, through the introduction of a domain specific language as an intermediate between different concrete source business process modelling notations and the target data modelling notation, which minimizes dependency of the data model generation process on the source notations.
A more detailed presentation of our BPM-driven approach to CDM synthesis, and the corresponding experimental results are available here, while the presentation about our online BPM-driven generator of the CDM can be downloaded from here.
Contact us for any additional information.
Patre 5, 78000 Banja Luka, Republic of Srpska, Bosnia and Herzegovina
+387 51 221 820
m-lab (at) etf.unibl.org