Keywords: Description Logics, Automated Reasoning, Query Answering, Parallelization, Tableau Algorithm
Abstract: Automated reasoning support is an important aspect of logic-based knowledge representation. The development of specialised procedures and sophisticated optimisation techniques significantly improved the performance even for complex reasoning tasks such as conjunctive query answering. Reasoning and query answering over knowledge bases with a large number of facts and expressive schemata remains, however, challenging. We propose a novel approach where the reasoning over assertional knowledge is split into small, similarly sized work packages to enable a parallelised processing with tableau algorithms, which are dominantly used for reasoning with more expressive Description Logics. To retain completeness in the presence of expressive schemata, we propose a specifically designed cache that allows for controlling and synchronising the interaction between the constructed partial models. We further report on encouraging performance improvements for the implementation of the techniques in the tableau-based reasoning system Konclude.
First Author Is Student: No
Subtrack: Ontologies and Reasoning