Efficient algorithms for qualitative reasoning about imprecise space

Abstract

This paper addresses the problem of qualitative spatial reasoning and presents efficient algorithms to deal it. We assume a representation which views space as a totality of objects surrounded by a haze area and connected in terms of qualitative spatial relations. Statements relating objects in this represesentation are expressed in terms of haze-order constraints. Reasoning about haze-orders involves, first, determining the consistency of a set of haze-order constraints, and, second, deducing new relations from those that are already known. The developed reasoning algorithms make use of a data structure called haze-order graph which trades space for efficiency. Experimental results illustrate the efficiency of the algorithms.