Data Independent Type Reduction for Zinc

Zinc is a high-level constraint modelling language with support for decision variables of complex types. Current constraint solvers only support decision variables of a subset of the types Zinc offers. Therefore, these variables and the constraints that involve them need to be reduced to a form that is supported by the target solvers. In this paper, we show how to reduce the type of decision variables in a systematic way, independent of a model’s instance data. We use the Cadmium model transformation language for a concise and modular implementation of these transformations.