Alaska EPSCoR Data Portal

{{title}}
{{description}}
Catalog Record: Geometry of the behavioral spaces: A computational approach to analysis and understanding the agent based models and agent behaviors.
Filename: imported_locations

Organizations

EPSCoR-ACE
Primary

Funding
NSF

Geometry of the behavioral spaces: A computational approach to analysis and understanding the agent based models and agent behaviors.

Systems with non-linear dynamics frequently exhibit emergent system behavior, which is important to find and specify rigorously to understand the nature of the modeled phenomena. Through this analysis, it is possible to characterize phenomena such as how systems assemble or dissipate and what behaviors lead to specific final system configurations. Agent Based Modeling (ABM) is one of the modeling techniques used to study the interaction dynamics between a system's agents and its environment. Although the methodology of ABM construction is well understood and practiced, there are no computational, statistically rigorous, comprehensive tools to evaluate an ABM's execution. Often, a human has to observe an ABM's execution in order to analyze how the ABM functions, identify the emergent processes in the agent's behavior, or study a parameter's effect on the system-wide behavior. This paper introduces a new statistically based framework to automatically analyze agents' behavior, identify common system-wide patterns, and record the probability of agents changing their behavior from one pattern of behavior to another. We use network based techniques to analyze the landscape of common behaviors in an ABM's execution. Finally, we test the proposed framework with a series of experiments featuring increasingly emergent behavior. The proposed framework will allow computational comparison of ABM executions, exploration of a model's parameter configuration space, and identification of the behavioral building blocks in a model's dynamics.

Complete

Data and Resources

  • View website

    Website :: Publisher's website

    http://scitation.aip.org/content/aip/journal/chaos/26/11/10.1063/1.4965982

Status: Complete
Type: Project
Data Types: Report
Primary Contacts
Name: Tania Clucas
Email: thclucas@alaska.edu
Primary Organizations
Type
EPSCoR - Alaska Adapting to Changing Environments
Academic

Funding Organization
Type
EPSCoR - Alaska Adapting to Changing Environments
Academic

Other Organizations
National Science Foundation

Regions
Alaska

Direct Record Link
http://northern.epscor.alaska.edu/catalog/entries/8868-geometry-of-the-behavioral-spaces-a-computati
Direct Full Map Link
http://northern.epscor.alaska.edu/catalog/entries/8868-geometry-of-the-behavioral-spaces-a-computati/map