ReqTLo: An Ontology for Exchanging and Managing Bounded Temporal Requirements for Simulation-Based Verification and Validation

Abstract

Complex cyber-physical systems (CPS) are an interdisciplinary combination of subsystems that are increasingly developed in collaborative settings, where specialized suppliers contribute components built with different tools and proprietary models. A growing share of this development is digitized, carried out through advanced simulation models that link to physical components. Recent developments enable simulation-driven Verification and Validation (V&V) through standards that allow the exchange of simulation models and the definition of system topologies. These standards include the System Structure and Parameterization (SSP), which standardizes system architecture, and the Functional Mock-up Interface (FMI), which standardizes component-model exchange through Functional Mock-up Units (FMUs). However, these standards do not provide an interoperable representation for the time-bounded behavioral requirements that govern co-simulation behavior. Requirements remain external to the co-simulation package, typically exchanged as unstructured text or tool-specific files, which removes explicit requirement structure and weakens traceability to model signals. To mitigate this, this paper introduces ReqTLo, a requirement exchange format built on World Wide Web Consortium (W3C) semantic web standards. ReqTLo embeds bounded Metric Temporal Logic (MTL) in a Web Ontology Language (OWL) knowledge graph and binds it to SSP components and FMU variables, with SHACL (Shapes Constraint Language) shapes validating structural well-formedness and SPARQL providing retrieval. An induction motor co-simulation case study demonstrates that system-level and component-level requirements remain discoverable through SPARQL competency questions after encoding and transfer. A comparative analysis against the string-based Requirements Interchange Format (ReqIF) further identifies representational capabilities that graph-based encoding provides natively for temporal formula structure and simulation variable bindings. Finally, a scalability study on 100 industry and literature-sourced requirements reports millisecond-level query times across the evaluated dataset.