In most instances, designers of digital simulation models ignore the statistical design and the statistical implications associated with these models. Most ignore the problem because of the added difficulties related to autocorrelated data. Fishman and Kiviat of the RAND Corporation suggested the use of spectral analysis as a technique for evaluating the significance of statistics emanating from a simulation model, comparing model output with 'real-world' data for model verification and comparing model outputs for two or more policy change evaluations. These authors have developed this concept into a synthesized model and verified the model through a spectral analysis on data derived from the logistics composite model (LCOM). LCOM was developed jointly by the RAND Corporation and the Air Force Logistics Command for the purpose of simulating aircraft flight and base support processes in response to mission requirements. Some forty statistics were evaluated through spectral analysis for confidence limits. Further, certain policy changes regarding inventory re-order were implemented and a spectral analysis was performed to judge whether or not the separate results were statistically different.
Statistical evaluation and verification of digital simulation models
Statistische Auswertung und Verifizierung von digitalen Simulationsmodellen
Computers and Industrial Engineering ; 3 , 1 ; 75-88
1979
14 Seiten, 92 Quellen
Article (Journal)
English
Comparison of statistical digital simulation methods
Tema Archive | 1988
|Statistical Validation of Traffic Simulation Models
Transportation Research Record | 2004
|Statistical versus Simulation Models in Safety
Transportation Research Record | 2009
|Failure-Adverse Closed-Loop Statistical Verification
AIAA | 2019
|Proactive Ergonomic Verification Through Human Biodynamic Modeling and Digital Simulation
SAE Technical Papers | 1999
|