Brazil has initiated two territorial public policies for a rural sustainable development, the National Program for Sustainable Development of the Rural Territories (PRONAT) and Citizenship Territory Program (PTC). These public policies aims, as a condition for its effectiveness, the equilibrium of the power relations between actors which participate in the Collegiate for Territorial Development (CODETER) of each Rural Territory. Our research studies the hypotheses that, in the Rural Territories submitted to the PRONAT and PTC public policies, the power and reciprocity relations between actors engaged in the CODETER effectively have evolved in favor of the civil society representatives to the detriment of the public powers, notably the mayors.

The SocLab approach has been applied in two case studies and four models representing the Southern Rural Territory of Sergipe (TRSS) and the São Francisco Rural Territory (TRBSF) were designed for two referential periods, 2008-2012 and 2013-2017. These models were developed to evaluate the empowerment of the civil society in these rural territories due to thes two public policies, PRONAT and PTC.

MASTOC-LLM extends the classic Multi-Agent System Tragedy of the Commons (MASTOC) model by replacing hard-coded behavioral rules with autonomous decision-making powered by large language models (LLMs). Three heterogeneous agents manage herds of cows on a shared grassland commons. Each tick, an agent receives a structured prompt describing current resource levels, its own herd size, peer behavior, and — optionally — a rolling memory of recent rounds and messages from neighboring agents. The LLM returns a stocking decision (add, remove, or hold cows) together with a natural-language rationale and, when communication is enabled, a short message to broadcast to peers.

The model is designed to test whether LLM agents spontaneously develop Ostrom-style common-pool resource governance (mutual monitoring, graduated sanctions, graduated rule revision) or instead fall into identifiable failure modes. Preliminary experiments with Claude Haiku 4.5, GPT-5.4-mini, and DeepSeek R1:32b have revealed four recurring collapse patterns — Cooperative Paralysis, Defection Cascade, Overshoot-Panic, and Hybrid Architecture Failure — whose onset timing is sensitive to memory length, inter-agent communication, and the post-training alignment approach of the underlying model.

MASTOC-LLM is intended as a laboratory for generative agent-based modelling (GABM) methodology: it provides a clean, well-understood commons baseline against which LLM behavioral hypotheses can be systematically tested and compared across models, parameter sweeps, and alignment regimes.

The model explores how two types of information - social (in the form of pheromone trails) and private (in the form of route memories) affect ant colony level foraging in a variable enviroment.

RAGE models a stylized common property grazing system. Agents follow a certain behavioral type. The model allows analyzing how household behavior with respect to a social norm on pasture resting affects long-term social-ecological system dynamics.

This model contains MATLAB code describing the virtual worlds framework used in the paper entitled “Polarization in Social Media: A Virtual Worlds-Based Approach.” The parent directory contains driver code for replicating results from the paper. Additionally, the source code is structured by three directories:

• Data Structures: Contains classes and objects used in the code, such as the virtualWorlds.m
• Metrics: Contains code which computes metrics, such as congruentLinks.m
• Visualization: Contains code for generating pictures and plots, such as drawSystemState.m

…

This model represents technological and ecological behaviors of mobile hunter-gatherers, in a variable environment, as they produce, use, and discard chipped stone artifacts. The results can be analyzed and compared with archaeological sites.

The basic idea behind developing MIXTRUST was to represent a network of agricultural stakeholders composed of farmers and a cooperative in a mixed landscape to test its performances in response to risks. A mixed landscape here is a landscape where crop and livestock systems interact by the intermediary of material flows of agricultural products. It can be within mixed farms, or between farms, often specialized, (e.g. straw-manure).

This agent-based model explores the dynamics between human behavior and vaccination strategies during COVID-19 pandemics. It examines how individual risk perceptions influence behaviors and subsequently affect epidemic outcomes in a simulated metropolitan area resembling New York City from December 2020 to May 2021.

Agents modify their daily activities—deciding whether to travel to densely populated urban centers or stay in less crowded neighborhoods—based on their risk perception. This perception is influenced by factors such as risk perception threshold, risk tolerance personality, mortality rate, disease prevalence, and the average number of contacts per agent in crowded settings. Agent characteristics are carefully calibrated to reflect New York City demographics, including age distribution and variations in infection probability and mortality rates across these groups. The agents can experience six distinct health statuses: susceptible, exposed, infectious, recovered from infection, dead, and vaccinated (SEIRDV). The simulation focuses on the Iota and Alpha variants, the dominant strains in New York City during the period.

We simulate six scenarios divided into three main categories:
1. A baseline model without vaccinations where agents exhibit no risk perception and are indifferent to virus transmission and disease prevalence.
…

This model implements a Bayesian belief revision model that contrasts an ideal agent in possesion of true likelihoods, an agent using a fixed estimate of trusting its source of information, and an agent updating its trust estimate.

3 simple models to illustrate diffusion of innovations.
The models are discussed in Introduction to Agent-Based Modeling by Marco Janssen. For more information see https://intro2abm.com/