An interval statistical procedure for use in validation of simulation models

We describe a new statistical procedure for use in conducting validation of simulation models via hypothesis testing when the amount of model accuracy required is specified as a range (interval). This statistical procedure provides for the model to be accepted as valid if the difference between the system and the model outputs is within the specified range of accuracy and be rejected as valid otherwise. Both Type I and Type II errors are considered through the use of the operating characteristic curve. The model builder’s and the model user’s risk curves are also developed.