Abstract
In the last years, the vision of the Semantic Web fostered the interest in reasoning over large and very large sets of assertional statements in knowledge bases. Traditional tableau-based reasoning systems perform bad answering queries over large data sets, because these reasoning systems are based on efficient use of main memory data structures. Increasing expressivity and worst-case complexity further tighten the memory burden. The purpose of our work is to investigate how to release the main memory burden from tableau-based reasoning systems and perform efficient instance checking over SHI-knowledge bases.
The key idea is to reduce instance checking for an individual in a knowledge base to smaller subsets of relevant axioms. Modularization techniques are introduced and further refined in order to increase the granularity of modules.
For evaluation purposes, experiments on benchmark and real world knowledge bases are carried out. The principal conclusion is that the main memory burden for instance checking can be released from tableau-based reasoning systems for semi-expressive Description Logics, by using modularization techniques.