Data Value Generation During Simulation

Regular process execution relies on process-relevant data being created or at least referenced during process execution. To generate these data for simulation is a very complex task because the modeler has to provide proper assumptions about the value distribution of possibly even complex data. Hence, the Simulation Feature normally does not rely on data generation for simulation purposes.

Hence, even if you decide to use the data generation facility of the Simulation Feature, the generated data are not correlated with the transition traversal probabilities.

However, a purpose of using the Simulation Feature might be to populate audit trail databases for mass and/or performance tests.

For these use cases it might be desirable to create the set of process-relevant data modeled for process definitions even if the values of these data do not or not sufficiently reflect real world value distributions. The following sections describe how to generate such data value sets during simulation.

Note, that data value creation is only relevant with regard when you are writing simulation results to audit trail databases.

Currently, data value generation is only supported for primitive numeric and string data.

Data Value Generation Algorithm

During simulation data values are created for each process instance, whenever the execution of a process instance is accessing a specific process data for the first time in an incoming or outgoing data mapping. Hence, if a process path containing an activity with the only data mapping for a specific process data is never traversed in a simulation run, no value for this data is created.

Numeric Data Value Generation

To specify the behavior of data value generation for numeric data such as

open the Simulation > Data Value Generation property page in the property dialog for your numeric primitive data.

Null Value Generation

To simply create null values for all process data according to the generation algorithm described above, select the Generate null-values option for the Data Value Generator.

Generation based on Gaussian Distribution

To create values for numeric process data distributed with a Gaussian value distribution according to the generation algorithm described above, select the Random Data According to Normal Distribution option for the Data Value Generator.

Generation based on Poisson Distribution

To create values for numeric process distributed with a Poisson data according to the generation algorithm described above, select the Random Data According to Poisson Distribution option for the Data Value Generator.

Generation based on Value Range

To create null for numeric process data distributed equally across a value range according to the generation algorithm described above, select the Integer Value from Specified Range option for the Data Value Generator.

String Data Value Generation

To specify the behavior of data value generation for string data open the Simulation > Data Value Generation property page in the property dialog for your string primitive data.

Null Value Generation

To simply create null values for string process data according to the generation algorithm described above, select the Generate null-values option for the Data Value Generator.

Generation From File

To simply create null values for string process data according to the generation algorithm described above, select the Random Data From External File option for the Data Value Generator.