Spatial Information Revision: A Comparison between 3 Approaches

The present paper deals with spatial information revision in geographical information system (GIS). These systems use incomplete and uncertain information and inconsistency can result, therefore the definition of revision operations is required. Most of the proposed belief revision operations are characterized by a high complexity and since GIS use large amount of data, adjustments of existing strategies are necessary. Taking advantage of the specificity of spatial information allows to define heuristics which speed up the general algorithms. We illustrate some suitable adjustments on 3 approaches of revision: binary decision diagrams, preferred models and Reiter's algorithm for diagnostic. We formally compare them and we experiment them on a real application. In order to deal with huge amount of data we propose a divide and revise strategy in the case where inconsistencies are local.