Statistics storage is used to hold information on action choice states and cycle types collected on the event history since the beginning of actor operation.
Additionally, statistics storage can hold probability profiles assigned to action choice states.
The proper use of probability profiles may substantially increase operating optimality of a system you develop.
Statistics storage is automatically created for an actor by the function
qsmm_actor_create and destroyed by the function
Storage API functions provide means to get the number of spur types supported by storage, to read and write a condition for an action choice state and statistics for a cycle type, to remove information on an action choice state from storage, to enumerate action choice states and cycle types, information on which is held in storage.
A powerful mechanism, which can be used by a developer, is setting storage redirection functions that return initial conditions for action choice states and initial statistics for cycle types when Storage API functions need this information. The mechanism can be especially helpful when specifying the same probability profile for a large number of nodes of a multinode model. Because probability profiles are stored within data structures of action choice states and cycle types, instances of those structures will not be used (and possibly allocated) until conditions for action choice states and statistics for cycle types are updated the first time. Such delayed use (and possible allocation) of the instances may reduce the memory footprint of storage and speed up program startup.
|• Types of Storage|
|• Structures for Accessing Storage|
|• Reading and Writing Statistics|
|• Enumerating States and Cycle Types|
|• Providing Initial Statistics|
|• Getting the Reason of a Storage Failure|
|• Example of Using the Storage API|