arXiv subjects

Dusan Guller

Publications and source records attributed to Dusan Guller.

3 recordsLinked to original sources

arXiv2023

Hyperresolution for Multi-step Fuzzy Inference in Goedel Logic

This paper is a continuation of our work concerning the logical and computational foundations of multi-step fuzzy inference. We bring further results on the implementation of the Mamdani-Assilian type of fuzzy rules and inference in Goedel logic with truth constants. In our previous work, we have provided translation of Mamdani-Assilian fuzzy rules to formulae of Goedel logic, and subsequently, to suitable clausal form. Moreover, we have outlined a class of problems regarding general properties of fuzzy inference and shown its reduction to a class of deduction/unsatisfiability problems. We now focus on solving such problems using an adapted hyperresolution calculus.

cs.LO↗

arXiv2023

On Multi-step Fuzzy Inference in Goedel Logic

This paper addresses the logical and computational foundations of multi-step fuzzy inference using the Mamdani-Assilian type of fuzzy rules by implementing such inference in Goedel logic with truth constants. We apply the results achieved in the development of a hyperresolution calculus for this logic. We pose three fundamental problems: reachability, stability, the existence of a k-cycle in multi-step fuzzy inference and reduce them to certain deduction and unsatisfiability problems. The corresponding unsatisfiability problems may be solved using hyperresolution.

cs.LO↗

arXiv2023

A DPLL Procedure with Dichotomous Branching for Propositional Product Logic

The propositional product logic is one of the basic fuzzy logics with continuous t-norms, exploiting the multiplication t-norm on the unit interval [0,1]. Our aim is to combine well-established automated deduction (theorem proving) with fuzzy inference. As a first step, we devise a modification of the procedure of Davis, Putnam, Logemann, and Loveland (DPLL) with dichotomous branching inferring in the product logic. We prove that the procedure is refutation sound and finitely complete. As a consequence, solutions to the deduction, satisfiability, and validity problems will be proposed for the finite case. The presented results are applicable to a design of intelligent systems, exploiting some kind of multi-step fuzzy inference.

cs.LO↗