Multi-scale Modeling of Functional Groups
Abstract
The Multi-scale Modeling of Functional Groups project will develop a modeling framework for explaining the emergence of large-scale social phenomena through the dynamics of functional groups at lower scales. Functional groups exist at multiple scales in social systems and produce behaviors of interest, like the decisions of agrarian households to send daughters to school or the polarization of mask-wearing by media organizations. Understanding these behaviors can help explain and predict the emergence of high-level social phenomena like gender norms and social justice movements. The aim of this project is to develop a fast, semi-automated method for learning and simulating the group structures and decision-making patterns present in functional groups. Multi-scale agent-based models are a good framework for modeling the emergence of high-level dynamics from lower-level behavior, provided that they include enough levels to explain the causal interactions between these behaviors. The goal of this project is to extend current multi-scale modeling practices to meet two challenges. The first is modeling within-group and between-group interactions at a single level, e.g. modeling the interactions within and between households. The second is modeling interactions between entities at different scales, e.g. modeling the interactions between households and gender norms. The approach will be to reduce as much of the burden on the modeler as possible through existing technologies, such as methods for discovering social networks and AI-assisted model-building. The focus of the project is on agricultural value chains, which provide a rich combination of functional groups and social, economic and agronomic factors. Few decisions along the agricultural value chain are made individually, so existing individual-based models will struggle to model high-level phenomena like consumer unrest due to shortages or farmer strikes due to agricultural policy. The framework developed in this project ought to be able to answer questions about the behaviors of farmer cooperatives, women’s groups, and market networks, without incurring a high computational cost for the user. This project will be done by Allegra Beal Cohen as a post-doctoral fellow at the University of Florida, advised by Professor Gerrit Hoogenboom.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 30, 2021
- Source ID
- HR00112110009
Entities
People
- Gerrit Hoogenboom
Organizations
- Defense Advanced Research Projects Agency
- University of Florida