A simulation model validation method based on design of experiments

When simulation factors are numerous while real-world observed data are sparse, the issue of validating the simulation models is problematic. An extreme case is focused that limited real-world observations are available cross the factor space, and only a single replicate is available on per simulation factor setting. A method based on design of experiments is proposed by which the validation experiments could be well arranged across the factor space through optimal design. The p-value test technology is employed to evaluate the statistical consistency of the static data, and for a set of validation experiments obtained by DoE, the combined analysis of all the p-values resulted from these experiments can be taken based on the inverse-CDF theorem, to make an overall characterization of degree of the simulation credibility on the entire factor space. An example of validation of a guided missile simulation is taken to demonstrate that the method is useful.