Parameterized Simplification Logic: Reasoning With Implications in an Automated Way

Abstract

In this sequel to our previous article (Cordero et al., 2020) on general inference systems for reasoning with if–then dependencies, we study transformations of if–then rules to semantically equivalent collections of if–then rules suitable to solve several problems related to reasoning with data dependencies.Wework in a framework of general lattice-based if–then rules whose semantics is parameterized by systems of isotone Galois connections. This framework allows us to obtain theoretical insight as well as algorithms on a general level and observe their special cases by choosing types of parameterizations. This way, we study methods for automated reasoning with different types of if–then rules in a single framework that covers existing as well as novel types of rules. Our approach supports a large family of if–then rules, including fuzzy if–then rules with various types of semantics. Themain results in this article include new observations on the syntactic inference of if–then rules, complete collections of rules, reduced normal forms of collections of rules, and automated reasoning methods. We demonstrate the generality of the framework and the results by examples of their particular cases focusing on fuzzy if–then rules.