Fuzzy Prolog: a new approach using soft constraints propagation

Sergio Guadarrama, Susana Muñoz-Hernández, Claudio Vaucheret

Abstract: We present a definition of a Fuzzy Prolog Language that models B([0,1])<math><mtext>B</mtext><mtext>([0,1])</mtext></math>-valued Fuzzy Logic, and subsumes former approaches because it uses a truth value representation based on a union of sub-intervals on [0,1] and is defined using general operators that can model different logics. This extension to Prolog is implemented by interpreting fuzzy reasoning as a set of constraints that are propagated through the rules by means of aggregation operators. Declarative and procedural semantics for Fuzzy Logic programs are given and their equivalence is proven. In addition, we present the implementation of an interpreter for this conceived language using constraint logic programming over real numbers.