Details

Categorical Unification

by Galán García, María Ángeles

Abstract (Summary)
This thesis deals with different aspects towards many-valued unification which have been studied in the scope of category theory. The main motivation of this investigation comes from the fact that in logic programming, classical unification has been identified as the provision of coequalizers in Kleisli categories of term monads. Continuing in that direction, we have used categorical instrumentations to generalise the classical concept of a term. It is expected that this approach will provide an appropriate formal framework for useful developments of generalised terms as a basis for many-valued logic programming involving an extended notion of terms.As a first step a concept for generalised terms has been studied. A generalised term is given by a composition of monads that again yields a monad, i.e. compositions of powerset monads with the term monad provide definitions for generalised terms. A composition of monads does, however, not always produce a monad. In this sense, techniques for monads composition provide a helpful tool for our concerns and therefore the study of these techniques has been a focus of this research.The composition of monads make use of a lot of equations. Proofs become complicated, not to mention the challenge of understanding different steps of the equations. In this respect, we have studied visual techniques and show how a graphical approach can provide the support we need.For the purpose of many-valued unification, similarity relations, generalised substitutions and unifiers have been defined for generalised terms.
Bibliographical Information:

Advisor:

School:Umeå universitet

School Location:Sweden

Source Type:Doctoral Dissertation

Keywords:TECHNOLOGY; Information technology; Computer science; Computer science; Datalogi; Monad compositions; generalised terms; many-valued logic; Datalogi; Computing Science; administrativ databehandling

ISBN:91-7305-642-1

Date of Publication:01/01/2004

© 2009 OpenThesis.org. All Rights Reserved.