Techniques for Sensitivity Analysis of Simulation Models: A Case Study of the CO2 Greenhouse Effect

摘要

Any simulation model requires sensitivity analysis. That model can be treated as a black box, if the techniques of regression analysis and experimental design are applied. The regression model is a metamodel of the simulation model, and guides the experimental design. The design leads to efficient and effective experimentation. The case study demonstrates that application of these statistical techniques requires knowledge of the underlying simulation model and real world system. For example, potentially important factors and their ranges must be given by the analysts. Some statistical expertise is needed to select the generators for the design.