57.9 Technical Details - Random Number Generator
Monte Carlo simulations in TreeAge Pro make use of a robust pseudo-random number generator (RNG) algorithm, the Mersenne Twister, which has the following useful properties:
-
Has a period of 2^19937, or approximately 10^6001 unique sequences.
-
Has negligible serial correlation between successive values in the output sequence.
-
Works fast.
-
Passes numerous tests for statistical randomness.
Distribution sampling and discrete simulation random walks utilize the RNG. By default, each RNG is “seeded” using the computer clock. This “random” seeding can be overridden by the user specifying a random seed value in the Simulation options. The Simulation Options can be found in Tree Preferences > Analysis Settings > Monte Carlo Options > Random Number Seeding Options.
In a multi-processor simulation, each thread has a separate RNG, which is started at a different position based on the clock seed or the user-specified.
More information about seeding options can be found in Tree Preferences section Monte Carlo Options
