Development and application of a validation framework for traffic simulation models

Discusses the concept of validation and proposes a multi-stage validation framework for traffic simulation models. The framework consists of conceptual validation and operational validation. The operational validation involves two levels of statistical tests: a two-sample t-test and a 2D two-sample Kolmogorov-Smirnov test. The validation experience employing the proposed framework demonstrates the fact that while a model can be valid for one level of detail, it can be invalid for another. The validation results also illustrate that the proposed multi-stage validation procedure can account for the complexity of the validation task and its conclusions