Active nonlinear tests (ANTs) of complex simulation models
Article Abstract:
The use of active nonlinear tests (ANTs) to validate the structure and robustness of a complex simulation model was described. ANTs, a general class of simple, automatic, nonlinear search algorithms, were designed to break the implications of a simulation model as well as improve and refine its design by probing for key weaknesses in the model's behavior. The algorithms can also be used to maximize the deviation between the original model's prediction and that produced from the perturbed model by searching across a set of reasonable model perturbations. The application of ANTS in testing the World3 model of global dynamics showed that the technique is effective in uncovering small but powerful nonlinear effects that may determine vulnerabilities in the original model.
Publication Name: Management Science
Subject: Business, general
ISSN: 0025-1909
Year: 1998
User Contributions:
Comment about this article or add new information about this topic:
Factor Screening in Simulation: Evaluation of Two Strategies Based on Random Balance Sampling
Article Abstract:
Computer simulation can be useful in the study of very complex real-world systems. Simulation models can be very complex themselves and expensive. Models could be more manageable if the most important factors could be identified. Factor screening methods are statistical methods that attempt to identify the more important variables. Two screening strategies are evaluated. Tables of numerical results are included. A graph of curves defining boundaries to zones of inadmissibility is also included.
Publication Name: Management Science
Subject: Business, general
ISSN: 0025-1909
Year: 1984
User Contributions:
Comment about this article or add new information about this topic:
Factor Screening in Simulation: Evaluation of Two Strategies Based on Random Balance Sampling
Article Abstract:
Computer simulation is a useful technique in management science. However simulation models tend to be large and complex. Two factor screening methods are developed. Both are based on random balance sampling. Performance is evaluated. Tables and graphs are given. An appendix gives proofs.
Publication Name: Management Science
Subject: Business, general
ISSN: 0025-1909
Year: 1984
User Contributions:
Comment about this article or add new information about this topic:
- Abstracts: A Simulation Analysis of Alternative Pricing Strategies for Dynamic Environments. Markup variability and flexibility: theory and empirical evidence
- Abstracts: Deriving and analysing optimal strategies in Bayesian models of games
- Abstracts: Neural network models for time series forecasts. An Urban Allocation Model Combining Time Series and Analytic Hierachical Methods