Search strategies for reasoning about spatial ontologies

The utilization of spatial ontologies to representing and reasoning about real-world problems exhibits some advantages when compared with traditional approaches. Namely reification, flexibility, efficiency and process understanding are considerably improved. However, the search mechanism, which underlies the reasoning process, always has exponential complexity, mainly because real-world problems produce models with high branching factors and solution depths. Therefore, the study and implementation of control strategies and cooperation among them are essential to meet the challenge of search complexity reduction. In particular, improvements on two fundamental steps of the process-successor generation and state completion can reduce drastically the search effort. In this paper, some search strategies are introduced, such as the utilization factor and the use of several lists of open nodes. Some classical approaches as sub-goaling and heuristic search are applied to the problem of controlling the process of reasoning about spatial ontologies. The other aim of this paper is to discussing some kinds of cooperation among the previous search strategies to reaching a search complexity reduction. The proposed search strategies were incorporated in a reasoner implemented in Prolog, whose performance has shown drastical improvements