Parallelizing description logics

Abstract

Description Logics (DL), one of the major paradigms in Knowledge Representation, face efficiency problems due to large-scale applications, expressive dialects, or complete inference algorithms. In this paper we investigate the potential of parallelizing DL algorithms to meet this challenge. Instead of relying on a parallelism inherent in logic programming languages, we propose to exploit the application-specific potentials of DL and to use a more data-oriented parallelization strategy that is also applicable to imperative programming languages. We argue that object-level propagation is the most promising inference component for such a parallelization, as opposed to normalization, comparison, or classification.

We present two alternative PROLOG implementations of parallelized propagation on a loosely coupled MIMD (Multiple Instruction, Multiple Data) system, one based on a farm strategy, the other based on distributed objects. Whereas the farm strategy yields only poor results, the implementation based on distributed objects achieves a considerable speedup, in particular for large-size applications.