For each parameter the user can chose one of the distributions
Distribution | Parameters | Return datatype | Remark |
Uniform | min, max | DOUBLE | |
Normal | mean, sigma | DOUBLE | |
Exponential | lambda | DOUBLE | |
Poisson | lambda | INT | |
Binomial | n, p | INT |
If min and max are specified and the generated value exceeds them, new random numbers are generated until one fits the interval. The number of trials is limited. If no number is found, the parameter set is skipped.
Generated parameter sets can be further restricted by constraints.
Arbitrary number of calibration variables (either INT or DOUBLE, depending on distribution used)
<cal id="VariableName" ...
For each calibration variable the distribution and it's parametrization.
<cal id="VariableName1" <input id="distribution" datatype="CHAR" <input id="mean" datatype="DOUBLE" ... <input id="sigma" datatype="DOUBLE" ... <input id="lambda" datatype="DOUBLE" ... <input id="p" datatype="DOUBLE" ... <input id="n" datatype="INT" ... </cal> <cal id="VariableName2"> <input id="distribution" ... <input id="mean" ... ... </cal>
See the list above for possible Distributions and their required parameters.
Name | DATATYPE | Description |
cCount | INT | the number of parameter sets to be generated, mandatory |
cRandomSeed | CHAR | optional random seed for the generator. Can be a long integer, therefore it has datatype CHAR |
The number of created simulations can be smaller than Count.
The random seed is set only on initialization of the generator. That means, that a subsequent iteration or a subsequent project can not set it's individual random seed.