A model can be stochastic as well as deterministic. There are many different ways to incorporate nondeterminism into a model. For example, you can assign a randomly generated time value to a transition, event rate, or delay operation. Or a random value or its derivative can be used to determine a message destination address, evaluate a guard expression, or otherwise impact the model behavior.
There is also a case when the model can have stochastic behavior, even if you do not specify it explicitly using randomly generated values: this is random serialization of simultaneous events. If several events are available at the same time, AnyLogic can make non-deterministic choice with equal probability for each event. Otherwise, the model behavior is deterministic and 100% reproducible irrespective of the seed of the random number generator.
To turn random serialization on
- In the Projects view, select the experiment you are currently working with.
- Go to the Randomness section of the Properties view.
- Select the option Random for the parameter Selection mode for simultaneous events.
To enable conducting complicated experiments over stochastic models, AnyLogic supports replication mechanism. Please refer to Optimizing stochastic models article for more information.
How can we improve this article?