Generating Fuzzy Regions from Conflicting Spatial Information

Applications such as geographic information retrieval need to deal with spatial information of a very diverse nature, typically involving a mixture of qualitative spatial relations and quantitative geographic constraints. Since existing approaches to spatial reasoning cannot easily be adapted to such a setting, we argue for a more direct approach, in which available information is viewed as constraints on possible representations of spatial regions, and we propose a specific implementation based on genetic algorithms. In the second part of this chapter, we turn our attention to the problem of inconsistencies. As no complete procedures for consistency checking are available, in general, instead of trying to repair the given knowledge base, we try to find representations of regions that satisfy available information to the best extent possible. To avoid arbitrary decisions, and to obtain solutions that emphasize, rather than ignore, the conflicting nature of different views, we advocate the use of fuzzy sets to represent regions. Finally, the results of a number of experiments are presented to demonstrate the effectiveness of the overall approach.