Model Repair with Quality-Based Reinforcement Learning

Domain modeling is a core activity in Model-Driven Engineering, and these models must be correct. A large number of artifacts may be constructed on top of these domain models, such as instance models, transformations, and editors. Similar to any other software artifact, domain models are subject to the introduction of errors during the modeling process. There are a number of existing tools that reduce the burden of manually dealing with correctness issues in models. Although various approaches have been proposed to support the quality assessment of modeling artifacts in the past decade, the quality of the automatically repaired models has not been the focus of repairing processes. In this paper, we propose the integration of an automatic evaluation of domain models based on a quality model with a framework for personalized and automatic model repair. The framework uses reinforcement learning to find the best sequence of actions for repairing a broken model.