Granularity in Relational Formalisms-With Application to Time and Space Representation

Temporal and spatial phenomena can be seen at a more or less precise granularity, depending on the kind of perceivable details. As a consequence, the relationship between two objects may differ depending on the granularity considered. When merging representations of different granularity, this may r...

Full description

Saved in:
Bibliographic Details
Published in:Computational intelligence 2001-11, Vol.17 (4), p.703-738
Main Author: Euzenat, Jérôme
Format: Article
Language:English
Subjects:
Artificial Intelligence
Computer Science
granularity
interval algebra
product
relation algebra
space representation
time representation
weakening
Citations: Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Temporal and spatial phenomena can be seen at a more or less precise granularity, depending on the kind of perceivable details. As a consequence, the relationship between two objects may differ depending on the granularity considered. When merging representations of different granularity, this may raise problems. This paper presents general rules of granularity conversion in relation algebras. Granularity is considered independently of the specific relation algebra, by investigating operators for converting a representation from one granularity to another and presenting six constraints that they must satisfy. The constraints are shown to be independent and consistent and general results about the existence of such operators are provided. The constraints are used to generate the unique pairs of operators for converting qualitative temporal relationships (upward and downward) from one granularity to another. Then two fundamental constructors (product and weakening) are presented: they permit the generation of new qualitative systems (e.g. space algebra) from existing ones. They are shown to preserve most of the properties of granularity conversion operators.
ISSN:0824-7935
1467-8640
DOI:10.1111/0824-7935.00170