A resolution-like strategy based on a lattice-valued logic
Jun Liu, Da Ruan, Yang Xu, Zhenming Song · IEEE Transactions on Fuzzy Systems · 2003
As the use of nonclassical logics becomes increasingly important in computer science, artificial intelligence and logic programming, the development of efficient automated theorem proving based on nonclassical logic is currently an active area of research. This paper aims at the resolution principle for the Pavelka type fuzzy logic (1979). Pavelka showed that the only natural way of formalizing fuzzy logic for truth-values in the unit interval [0, 1] is by using the Lukasiewicz's implication operator a/spl rarr/b=min{1,1-a+b} or some isomorphic forms of it. Hence, we first focus on the resolution principle for the Lukasiewicz logic L/sub /spl aleph// with [0, 1] as the truth-valued set. Some limitations of classical resolution and resolution procedures for fuzzy logic with Kleene implication are analyzed. Then some preliminary ideals about combining resolution procedure with the implication connectives in L/sub /spl aleph// are given. Moreover, a resolution-like principle in L/sub /spl aleph// is proposed and the soundness theorem of this resolution procedure is also proved. Second, we use this resolution-like principle to Horn clauses with truth-values in an enriched residuated lattice and consider the L-type fuzzy Prolog.