Model verification is the process of determining that a model or experimentation implementation and its associated data accurately represents the researcher’s conceptual description and specifications. In modeling and experimentation methodology, model verification usually done by unit level testing. Experimentation workbench is treated as a black box for unit tests. The strategy should be to test at small scale and verify emulation testbed is producing results as expected – “crawl before walking”. Test cases are built around model specification and requirements.
Model validation is the process of determining the degree to which a model or experimentation and its associated data are an accurate representation of the real world from the perspective of the intended uses of the model. Validating Army’s C4ISR system of systems model against real world data is extremely challenging. Ideally, comparing the model input-output transformations to corresponding input-output transformations for the real system is the perfect methodology for model / experimentation validation. Test data generated from Army’s Brigade Modernization program known as Network Integration Evaluation (NIE) exercise is an ideal real system data for validating system of systems experimentation testbed. Before getting into comparative analysis, first it is important to layout model assumptions to validate against real system data. Followed by accurately replicating test architecture, operational data traffic and scenarios into emulation testbed which contains verified models.