CARE - A Constraint-Based Approach for Re-Establishing Conformance-Relationships

In Model-Driven Engineering, models have to conform to their associated linguistic and ontological metamodels. While linguistic metamodels are usually not subject to frequent changes, ontological metamodels are. Thus, ontological conformance may be easily corrupted by changes on the metamodel or by the incautious modification of models. Current approaches for re-establishing conformance relationships are often (i) deeply woven into specific tools to record changes and to derive resolutions, or (ii) require extensive user effort to guide the resolution process, and (iii) the output of these approaches usually is one single solution, whereas alternative solutions remain unexplored. To allow for exploring a broader solution space independent of specific tools and to avoid extensive user involvement by utilizing predefined repair actions, we propose a logic programming approach called CARE, for accomplishing multiple solutions. In particular, CARE is based on a formalization of the ontological conformance relationship as constraints, accompanying repair actions for counteracting constraint violations, as well as quality criteria for ranking the solutions. This paper reports on the realization of CARE based on Answer Set Programming and summarizes lessons learned from applying the approach in several experiments.