Validation of Simulation Models: Mine-Hunting Case-Study.

摘要

Stringent validation requires that simulation and real-life responses have the same mean. The responses, however, may show not only sampling error but also measurement error. Moreover, simulated and real responses are not comparable if they are obtained under different environmental conditions or scenarios. Modules within the simulation model should be submitted to sensitivity analyses based on experimental design theory and regression analysis. A weaker validation procedure tests whether the estimated simulation and real responses are positively correlated (they do not necessarily have a common mean). These issues are illustrated through a study on mine hunting at sea by means of a sonar.