Skip to content
Sign inStart free

User-Defined Distribution

A user-defined distribution lets you build a distribution from your own numbers instead of a formula: either a set of observed values, or a list of values each with a percentage. Once it is defined and named, you sample it with UD('name').

You create one in the Expression Builder: supply the data, give it a name, and it is saved with the model. There are two ways to provide the data:

  • From observations: paste a column of measured values and ProcessModel samples from them, either fitting a smooth shape or picking exact values.
  • From value and percent pairs: list each value and how often it should occur, for example 10 minutes 20 percent of the time and 20 minutes 80 percent of the time.

Then reference it anywhere a distribution is allowed:

UD('ServiceTimes')

Use a user-defined distribution when no standard shape fits your data, or when the measured values themselves are what you want to sample. For just a few fixed outcomes, the discrete distribution is simpler.