Building confidence in Simulation Models using automated analyses

Different descriptions of validation tests are available in literature, but few attempts have been made to operationalize them. In this paper, an attempt is made to standardize validation tests and implement them in an automated testing sequence. This testing sequence, called PySD testing battery, provides modelers and reviewers with insight on model behavior and creates input for reviewing models. Tests currently included are time step and extreme condition tests. Interpretation guidelines are elaborated wherever applicable. The use of automated analyses in the modeling workflow will greatly increase the rigor applied to model testing and should enable modelers to deliver higher-quality models.