Quantifying validation of discrete event simulation models

Simulation model validation and its rigorous assessment is known to be a difficult task. Quantification of validation is necessary to answer the question “how much validation is adequate?” One can answer this question by developing adequacy criteria to measure the validation performed on a simulation model. Developing test adequacy criteria for verifying computer programs has been useful for improving the quality and increasing confidence of regular software systems. We argue that from the validation and verification (V&V) perspective, simulation models are no different than software that are generally termed as “non-testable” due to the absence of a test oracle. There has been little research to develop V&V related adequacy criteria for this type of software. We present a validation coverage criterion, show in detail how it applies to discrete-event simulation models (DES), and discuss how it can be extended to develop V&V coverage criteria for other “non-testable” software.