CHOROS 2: Improving the Performance of Qualitative Spatial Reasoning in OWL

Abstract

We investigate on potential improvements to reasoning methods for topological and directional spatial information in OWL. Building upon path consistency, the new reasoner design, referred to as CHOROS 2, suggests several optimizations for reducing the number of compositions of basic relations and for speeding-up its run-time performance. CHOROS 2 serves also as a framework for a realistic evaluation of several alternative reasoner designs discussed in this paper. Perhaps, one of the most interesting alternatives that deserves further investigation relies on the idea of decomposing directional relations into two smaller sets of basic relations yielding fewer compositions. CHOROS 2 infers all implied relations and detects inconsistencies while retaining soundness, completeness and tractability over the supported relations sets. Experimental results demonstrate that all variants of CHOROS 2 run up to several times faster than both CHOROS 1 (its previous implementation) and SOWL, a spatial reasoner implemented in SWRL which runs under Protégé.