Multi-Agent Specification-based Metamorphic Testing of FMU-Based Simulations
Researchers have developed a novel LLM-powered multi-agent system for metamorphic testing of simulation models. This approach automates the extraction of metamorphic relations (MRs) from functional and interface specifications, which are typically manual and error-prone processes. The system uses a Given-When-Then pattern to define MRs and generates test cases to verify simulation consistency, as demonstrated on a Lube Oil Cooling system FMU. AI
IMPACT Automates complex testing for simulation models, potentially reducing manual effort and improving verification.