An Optimization Approach for Matching Textual Domain Models with Existing Code

We address the task of mapping a given textual domain model with the source code of an application which is in the same domain but was developed independently of the domain model. The key novelty of our approach is to use mathematical optimization to find a mapping between the elements in the two sides that maximizes the instances of clusters of related elements on each side being mapped to clusters of similarly related elements on the other side. We describe experiments wherein we apply our approach to the task of matching two real, open-source applications to corresponding industry-standard domain models. In comparison with previous approaches that leverage relationships, but are formulated as heuristics rather than as a principled optimization problem, our approach gives up to 40% higher precision given a desired level of recall.