Empirical model-building and response surface
Empirical model-building and response surface
Statistical tools for simulation practitioners
Statistical tools for simulation practitioners
A guide to simulation (2nd ed.)
A guide to simulation (2nd ed.)
An experimental procedure for simulation response surface model identification
Communications of the ACM
The global simulation clock as the frequency domain experiment index
WSC '88 Proceedings of the 20th conference on Winter simulation
Frequency domain metamodelling of a feedback queue
WSC '87 Proceedings of the 19th conference on Winter simulation
A model for frequency domain experiments
WSC '87 Proceedings of the 19th conference on Winter simulation
A spectral method for confidence interval generation and run length control in simulations
Communications of the ACM - Special issue on simulation modeling and statistical computing
Solution to the indexing problem of frequency domain simulation experiments
WSC '91 Proceedings of the 23rd conference on Winter simulation
The effects of batching on the power of the test for frequency domain methodology
WSC '91 Proceedings of the 23rd conference on Winter simulation
| Hi-index | 0.01 |
We present three extensions of the frequency domain approach proposed by Schruben and Cogliano (1987). The first is the assignment of multiple frequencies to input factors. The frequency selection in this case is nearly identical to that of one frequency per factor. The second extension is the use of the time series of batch means to flatten the noise spectrum and make the identification of peaks in the output spectrum easier. Finally, we show that using common random numbers (as suggested in Schruben and Cogliano, 1987) for the signal and noise runs do indeed tend to decrease the peaks in the noise spectrum and also facilitate factor identification.We discuss several limitations to the use of frequency domain methodology as it currently exists. One limitation is the case in which a factor-frequency interaction is present. Although none of the simulation models currently being examined seems to have this property, it is one present in certain chaotic dynamical systems. These systems demonstrate qualitatively different behavior when oscillated. Another limitation involves the time index. For complicated systems it may not be feasible to define a single “clock” with which to oscillate parameters. Current work (Jacobson, et. al., 1988) shows that certain models may use the global time clock, but it is not clear that there is always a solution.