OptML framework and its application to model optimization

Abstract Model-based software development, in particular Model-Driven Engineering (MDE), has received a lot of attention during the last decade. More and more companies adopt MDE methods in software development. As a consequence, software engineers may need to deal with a large model base which becomes the main company asset. Moreover, there may be many configurations of the same model resulting in a large set of alternatives for the software engineers to take into account. This chapter presents the OptML Framework, which accepts various models defined in the Ecore MDE environment and computes the optimal models which satisfy user-defined constraints. Commonly used models in the MDE context such as class model, feature model, platform model, and process model are utilized in this chapter for illustration purposes. The examples of user-defined criteria are timeliness, energy consumption, and precision. To the best of our knowledge, this is the first generic MDE framework that is suitable for model optimization. The term generic here refers to a tool workbench architecture that is deliberately designed to extend the workbench with new features if necessary. The framework is fully implemented and tested for a selected set of models defined in the image processing domain.