Constraints satisfaction problems in data modeling

Abstract

We present a new approach to study the global coherence of constraints in data modeling. Unlike most of other approaches, we take a best part of these constraints into account during modeling processing. Thus, it is useless to generate the database - which may be partially or totally empty - if the constraints are in conflict. To this end, we formalize the constraints in mathematical inequalities form combined with expressions in first-order logic. We propose two approaches to detect and localize conflicts. In the first one, an algorithm, based on Fourier-Motzkin elimination, is suggested. In the second approach, the conceptual schema is formalized using a knowledge base.