Abstract
Deterministic simulation models are important in agricultural applications and their use is becoming increasingly common. Therefore, statistical procedures that interpret the output and evaluate the performance of deterministic models are necessary. The fact that deterministic computer simulation experiments cannot be replicated provides opportunities for using several procedures applicable to unreplicated factorial experiments. We discuss a classification scheme that selects the correct technique for most deterministic simulation experiments. The value of these techniques is their capability to estimate the experimental error variance for unreplicated computer experiments. Using these estimates of error, model developers and practitioners can more thoroughly analyze their deterministic simulation experiments.
Keywords
Deterministic simulation models, simulation experiments, unreplicated factorials, key to classification, statistical analysis, estimation of error
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.
Recommended Citation
Willers, J L. and Vinyard, B. T.
(1995).
"A CLASSIFICATION OF UNREPLICATED FACTORIAL EXPERIMENTS FOR USE WITH THE ANALYSIS OF DETERMINISTIC SIMULATION MODELS,"
Conference on Applied Statistics in Agriculture.
https://doi.org/10.4148/2475-7772.1337
A CLASSIFICATION OF UNREPLICATED FACTORIAL EXPERIMENTS FOR USE WITH THE ANALYSIS OF DETERMINISTIC SIMULATION MODELS
Deterministic simulation models are important in agricultural applications and their use is becoming increasingly common. Therefore, statistical procedures that interpret the output and evaluate the performance of deterministic models are necessary. The fact that deterministic computer simulation experiments cannot be replicated provides opportunities for using several procedures applicable to unreplicated factorial experiments. We discuss a classification scheme that selects the correct technique for most deterministic simulation experiments. The value of these techniques is their capability to estimate the experimental error variance for unreplicated computer experiments. Using these estimates of error, model developers and practitioners can more thoroughly analyze their deterministic simulation experiments.
