Stochastic Simulation

RoadRunner supports stochastic simulation through the use of Gillespie algorithm, which is a variation of Monte Carlo method.

To set the current instance of RoadRunner to use the Gillespie solver, try:

>>>  rr.setIntegrator('gillespie')

RoadRunner also supports a simplified method to run Gillespie solver through gillespie().

One of the important component of stochastic simulation is setting the seed. Random number genetration in computers are known to be ‘pseudo-random’, meaning it can only ‘approximate’ randomness. Seed is an initial key value for generating a sequence of numbers. This means that when a seed is given, it is possible to regenerate identical sequence of numbers from random number generator. This might be desireable for reproducibility perposes but undesireable for multiple iterations of random number creation.

By using getSettings(), you can check settings accepted by Gillespie solver.

>>>  rr.getIntegrator().getSettings()
('seed',
 'variable_step_size',
 'initial_time_step',
 'minimum_time_step',
 'maximum_time_step',
 'nonnegative')

For information on what the settings represent, try getDescription(). Check Roadrunner Solvers for additional information.