Topllet: An Optimized Engine for Answering Metric Temporal Conjunctive Queries

Lukas Westhofen, Christian Neurohr, Jean Christoph Jung, Daniel Neider

Abstract: We present Topllet, a software tool for answering Metric Temporal Conjunctive Queries over temporal knowledge bases with ontologies formulated in expressive Description Logics. Its main use case is the formal specification of requirements and their evaluation against test data when confronted with a highly complex operational domain of the system under test, e.g., urban automated driving. It is implemented as a module in the well-established reasoner Openllet, which offers good performance in the core reasoning tasks (such as consistency checks) but lacks support for temporal properties. Although the underlying problem of answering queries in the examined logics is ExpTime-hard, this work shows how we practically tackle this theoretical complexity. Despite being the first implementation of the task, Topllet already exhibits satisfactory performance due to our optimizations.