Data Input, Collection, and Analysis


Input Analysis

The accuracy of the analysis of any output generated by a simulation model is highly dependent on the appropriateness of the inputs that are used to drive the simulation model. Often, data are available and you want to use those data to estimate the parameters of a theoretical distribution and then sample from that fitted distribution to generate inputs to your model. In this case, you can use the Fitted option in the Numeric Source block to access the JMP automatic distribution-fitting procedure. After you specify a value for File Path and click the Fit Distribution button in a Numeric Source block, JMP automatically fits a series of distributions to the specified data and ranks the results. Either you can select the best fit that is suggested by JMP or you can investigate other distributions and use the analysis options available in JMP to make a distribution choice. After you select a distribution, the parameters for that distribution are automatically passed back to the Numeric Source block. See Appendix D: Input Analysis, for details.

In addition to selecting a theoretical distribution to generate inputs to a model, the Numeric Source block also enables you to generate samples from discrete and continuous empirical distributions, which can be especially useful when it is difficult to find an appropriate theoretical distribution that accurately represents the data. Finally, you can also use the Numeric Source block to specify a nonhomogeneous Poisson process that is based on either count data or rate data. Simulation Studio uses the count data or rate data to generate a time-dependent arrival process for a specified time interval that can be used as an input to a model. For more information about using empirical distributions and nonhomogeneous Poisson processes, see Appendix B: Random Variation in a Model.