Evalutate OWL Reasoners for real industry use cases

Technology Review

Estimated Effort:

250 h

Abstract

A reasoning engine is a piece of software able to infer logical consequences from a set of asserted facts. OWL is a family of knowledge representation languages popular in the Semantic Web. For sublanguages such as OWL-DL typical reasoning tasks are decidable and a reasoner will be able to infer all implicit knowledge. However, OWL reasoning might be infeasible in terms of space and time complexity. Reasoners implement different strategies to deal with this problem. For reasoning on real world knowledge graph therefore profiling tests have to be performed to show the actual performance of reasoners.

The facts

Expected results

• Comparision of the performance of popular reasoners on a given data set from the automation domain
• Design of a test setup that allows to alter the data in meaningful ways, and to observe and describe the effects on reasoning performance.

Literature

• Test cases for OWL2 inferences:
  • wiki: http://owl.semanticweb.org/
  • all test cases in RDF format: http://owl.semanticweb.org/exports/all.rdf