Simulation statistical software: an introspective appraisal

Simulation experiments are sampling experiments by their very nature. Statistical issues dominate all aspects of a well-designed simulation study-model validation, selection of input distributions and associated parameters, experiment design frameworks, output analysis methodologies, model sensitivity, and forecasting are examples of some of the issues which must be dealt with by simulation experimenters. There are many factors which complicate analyses, such as multivariate input distributions, serially correlated model inputs and outputs, multiple performance measures, and non-linear system response, to name a few. The purpose of this panel is to discuss any and all issues related to software tools available for dealing with these and other problems. I asked five experts from within the simulation community to share their opinions and insights on the availability and quality of software to meet the statistical needs of the simulation community. The position statements provided by them are intended to serve as a springboard for a more extensive exchange of ideas during the discussion at the conference.