Tools that use the Random Number Generator environment use algorithms that use the seed and distribution to produce a sequence of random numbers.
- If a tool using the random values (for example, Create Random Raster, Create Random Points, or Calculate Value) runs twice with the same seed, the output is identical. Control over the seed allows you to reproduce your results.
- If a random number generator is defined globally, one stream of random values is created. Each subsequent tool or tools within a ModelBuilder model pulls the next available value from the stream as needed.
- Multiple random number generator streams can be used in a ModelBuilder model by creating them locally for the desired tools that use random numbers. Each tool can have its own stream from which to pull random values as needed. Only the tool with the local stream can pull values from that stream. There is no limit to the number of local streams.
- Seed—The seed is an integer value and is used to initiate the random number generator. The default value is 0.
- Random Generator Type—The random generator algorithm.
- ACM599—ACM collected algorithm 599. This is the default.
- MERSENNE_TWISTER—Mersenne Twister mt19937.
- STANDARD_C—Standard C Rand.
arcpy.env.randomGenerator = random_generator_option
A RandomNumberGenerator can be created using the CreateRandomValueGenerator function or returned from the randomGenerator environment.
import arcpy # Set the random generator to ACM599 with random seed of 99 arcpy.env.randomGenerator = "99 ACM599"