This paper gives a survey on how to validate simulation models through the application of mathematical statistics. The type of statistical test actually applied, depends on the availability of data on the real system: (i) no data, (ii) only output data, and (iii) both input and output data. In case (i), the system analysts can still experiment with the simulation model to obtain simulated data. Those experiments should be guided by the statistical theory on design of experiments (DOE); an inferior - but popular - approach is to change only one factor at a time. In case (ii), real and simulated output data may be compared through the well-known Student t statistic. In case (iii), trace-driven simulation becomes possible. Then validation, however, should not proceed as follows: make a scatter plot with real and simulated outputs, fit a line, and test whether that line has unit slope and passes through the origin. Instead, better tests are presented. Several case studies are summarized, to illustrate the three types of situations.
|Place of Publication||Tilburg|
|Number of pages||21|
|Publication status||Published - 1998|
|Name||CentER Discussion Paper|