Simulation data and properties

LASER is designed to handle large populations with high numbers of independent agents. In order to track agents and their properties, LASER includes functions that store and update this information.

The classes described below are only a subset of LASER components. For a full list of all LASER classes, see the API reference.

LaserFrame and PropertySet

LaserFrame can be thought of as a database system for LASER models. It is a class that is used to dynamically manage and allocate data for nodes and agents, and supports both scalar and vector properties.

PropertySet stores agent properties in a dictionary-like object. Properties can be dynamically updated and this class is used to help define simulation parameters.

The tutorial Build SIR models demonstrates how these two classes can be implemented.

SortedQueue

Timing of events is an important function for agent-based models. Management and tracking of values, especially as the number of agents reaches the millions, can be problematic. SortedQueue is a custom process created for LASER to track and sort these values, and works directly with LaserFrame object arrays.