Gl. Stevens et al., ANALYSIS OF DETERMINISTIC SIMULATION-MODEL PERFORMANCE USING NON-REPLICATED FACTORIAL 2-LEVEL EXPERIMENTS, Agricultural systems, 52(2-3), 1996, pp. 293-315
The usefulness of deterministic simulation models has increased in agr
icultural applications. This increase must be matched with the use of
analytical tools that can verify and validate their performance and he
lp interpret their output. Values of model input variables can be sele
cted by the analyst and are analogous to the concept of factors in con
ventional experiments. However, deterministic models cannot produce in
dependent replications of selected input variables; many statistical a
nalysis procedures cannot therefore be utilized. Consequently, this pr
esentation describes the use of techniques applicable to the analysis
of unreplicated, two-level factorial experiments. It is demonstrated t
hat these concepts are a solution to the problem of how to evaluate de
terministic simulation model performance and analyse deterministic sim
ulation experiments. Copyright (C) 1996 Published by Elsevier Science
Ltd