A Tool for Agent-Based System Produced Emergent Networks (ASPEN)

Period of Performance: 09/28/2012 - 12/31/2014

$1000K

Phase 2 SBIR

Recipient Firm

Charles River Analytics, Inc.
625 Mount Auburn Street Array
Cambridge, MA 02138
Principal Investigator

Abstract

Social media is an increasingly important way for people to communicate, but interpreting the large and complex social networks that emerge from social media is challenging. A promising approach is to build synthetic networks based on realistic behaviors that mimic the form and structure of real networks, and then infer behaviors that underlie the real networks from the models used to generate the synthetic ones. In ASPEN Phase I, we integrated: (1) behavioral theory in an agent-based simulation to inform the generation of synthetic networks; (2) social network analysis techniques to test for similarity to empirical networks; and (3) evolutionary algorithms to iteratively increase the fidelity of the synthetic networks. In Phase II, we will extend our use of evolutionary algorithms to explore the relationship between agent behavior and network structure by analyzing data from a social media testbed. This will allow us to move ASPEN from being strictly a synthetic network generator to a powerful research and development tool for social media analysis and influence operations support.