Deakin University
Browse

File(s) under permanent embargo

Research developments in multiple inheritance with exceptions

journal contribution
posted on 1994-03-01, 00:00 authored by Peter EklundPeter Eklund
The inheritance problem can be simply stated: for any instantiation of an inheritance network, say a specific hierarchy T, find a conclusion set for T. In other words, find out what is logically entailed by F. This can be done in two ways: either by defining a deductive or proof theoretic definition to determine what paths are entailed by a network; or by translating the individual links in the network to a more general nonmonotonic logic and using its model and proof theory to generate entailments that correspond to what one would expect from “viewing” the inheritance hierarchy. Two approaches to a solution to the inheritance problem structure this paper. The first is widely known as the “path-based” or “proof theoretic”, and the second, the “model-based” or “model theoretic”. The two approaches result in both a different interpretation of default links as well as a variation in the entailment strategy for a solution to the inheritance problem. In either case, the entailments produced need some intuitive interpretation, which can be either credulous or skeptical. The semantics of both skeptical and credulous inheritance reasoners are examined.

History

Journal

Knowledge engineering review

Volume

9

Issue

1

Pagination

21 - 55

Publisher

Cambridge University Press

Location

Cambridge, Eng.

ISSN

0269-8889

eISSN

1469-8005

Language

eng

Publication classification

C1.1 Refereed article in a scholarly journal

Copyright notice

1994, Cambridge University Press