Modeling input processes

Computer models for various applications are closely scrutinized both from the standpoint of questioning the correctness of the underlying mathematical model with respect to the process it is attempting to model, and from the standpoint of verifying that the computer model correctly implements the underlying mathematical model. A process that receives less scrutiny, but is none the less of equal importance, concerns the individual and joint modeling of the inputs. This modeling effort clearly has a great impact on the credibility of results obtained from simulation studies. Model characteristics are reviewed that have a direct bearing on the model input process and reasons are given for using probabilistic based modeling with the inputs. Discussions are presented on how to model distributions for individual inputs and how to model multivariate input structures when dependence and other constraints may be present.