RT Journal Article
T1 Proximity-based Unification: an Efficient Implementation Method
A1 Julián-Iranzo, Pascual
A1 Sáenz Pérez, Fernando
AB Unification is a central concept in logic systems based on the resolution principle. As well, in knowledge representation, proximity relations (i.e., reflexive, symmetric, fuzzy binary relations) are useful for introducing semantics into a syntactic level by modelling the semantic closeness of different syntactic objects and managing vague or imprecise information. Proximity relations, in combination with the unification algorithm, make possible expressing certain forms of approximate reasoning in a logic programming framework. We use proximity relations in the context of a (fuzzy) logic programming system, called Bousi_Prolog, as a way of solving the limitations introduced by similarity relations (i.e., transitive proximity relations) to correctly represent fuzzy information. Recently, we introduced an accurate definition of proximity between expressions (terms or atomic formulas) and a new unification algorithm able to manage proximity relations properly. However, the so-called weak unification algorithm, which is an extension of Martelli and Montanari’s unification algorithm supported by the new notion of proximity, does not have an efficient implementation. In this paper, we present a method that facilitates such an efficient implementation, including an adaptation of the weak SLD resolution rule based on the new unification algorithm, and its integration and implementation into the fuzzy logic programming system Bousi_Prolog. A performance analysis to show its efficiency is also presented.
PB IEEE
SN 1063-6706
YR 2021
FD 2021-05
LK https://hdl.handle.net/20.500.14352/114021
UL https://hdl.handle.net/20.500.14352/114021
LA eng
NO P. Julián-Iranzo and F. Sáenz-Pérez, "Proximity-Based Unification: An Efficient Implementation Method," in IEEE Transactions on Fuzzy Systems, vol. 29, no. 5, pp. 1238-1251, May 2021, doi: 10.1109/TFUZZ.2020.2973129.
DS Docta Complutense
RD 7 abr 2025