This is an initial exploratory exercise done for the class @ http://thiagomarzagao.com/teaching/ipea/ Text available here: https://arxiv.org/abs/1712.04429v1
The program:
Reads output from an ABM model and its parameters’ configuration
Creates a socioeconomic optimal output based on two ABM results of the modelers choice
Organizes the data as X and Y matrices
Trains some Machine Learning algorithms
…

Violence against women occurs predominantly in the family and domestic context. The COVID-19 pandemic led Brazil to recommend and, at times, impose social distancing, with the partial closure of economic activities, schools, and restrictions on events and public services. Preliminary evidence shows that intense co- existence increases domestic violence, while social distancing measures may have prevented access to public services and networks, information, and help. We propose an agent-based model (ABM), called VIDA, to illustrate and examine multi-causal factors that influence events that generate violence. A central part of the model is the multi-causal stress indicator, created as a probability trigger of domestic violence occurring within the family environment. Two experimental design tests were performed: (a) absence or presence of the deterrence system of domestic violence against women and measures to increase social distancing. VIDA presents comparative results for metropolitan regions and neighbourhoods considered in the experiments. Results suggest that social distancing measures, particularly those encouraging staying at home, may have increased domestic violence against women by about 10%. VIDA suggests further that more populated areas have comparatively fewer cases per hundred thousand women than less populous capitals or rural areas of urban concentrations. This paper contributes to the literature by formalising, to the best of our knowledge, the first model of domestic violence through agent-based modelling, using empirical detailed socioeconomic, demographic, educational, gender, and race data at the intraurban level (census sectors).

Policymakers decide on alternative policies facing restricted budgets and uncertain future. Designing public policies is further difficult due to the need to decide on priorities and handle effects across policies. Housing policies, specifically, involve heterogeneous characteristics of properties themselves and the intricacy of housing markets and the spatial context of cities. We propose PolicySpace2 (PS2) as an adapted and extended version of the open source PolicySpace agent-based model. PS2 is a computer simulation that relies on empirically detailed spatial data to model real estate, along with labor, credit, and goods and services markets. Interaction among workers, firms, a bank, households and municipalities follow the literature benchmarks to integrate economic, spatial and transport scholarship. PS2 is applied to a comparison among three competing public policies aimed at reducing inequality and alleviating poverty: (a) house acquisition by the government and distribution to lower income households, (b) rental vouchers, and (c) monetary aid. Within the model context, the monetary aid, that is, smaller amounts of help for a larger number of households, makes the economy perform better in terms of production, consumption, reduction of inequality, and maintenance of financial duties. PS2 as such is also a framework that may be further adapted to a number of related research questions.

The DiDIY-Factory model is a model of an abstract factory. Its purpose is to investigate the impact Digital Do-It-Yourself (DiDIY) could have on the domain of work and organisation.

DiDIY can be defined as the set of all manufacturing activities (and mindsets) that are made possible by digital technologies. The availability and ease of use of digital technologies together with easily accessible shared knowledge may allow anyone to carry out activities that were previously only performed by experts and professionals. In the context of work and organisations, the DiDIY effect shakes organisational roles by such disintermediation of experts. It allows workers to overcome the traditionally strict organisational hierarchies by having direct access to relevant information, e.g. the status of machines via real-time information systems implemented in the factory.

A simulation model of this general scenario needs to represent a more or less abstract manufacturing firm with supervisors, workers, machines and tasks to be performed. Experiments with such a model can then be run to investigate the organisational structure –- changing from a strict hierarchy to a self-organised, seemingly anarchic organisation.

This is a replication of the SequiaBasalto model, originally built in Cormas by Dieguez Cameroni et al. (2012, 2014, Bommel et al. 2014 and Morales et al. 2015). The model aimed to test various adaptations of livestock producers to the drought phenomenon provoked by climate change. For that purpose, it simulates the behavior of one livestock farm in the Basaltic Region of Uruguay. The model incorporates the price of livestock, fodder and paddocks, as well as the growth of grass as a function of climate and seasons (environmental submodel), the life cycle of animals feeding on the pasture (livestock submodel), and the different strategies used by farmers to manage their livestock (management submodel). The purpose of the model is to analyze to what degree the common management practices used by farmers (i.e., proactive and reactive) to cope with seasonal and interannual climate variations allow to maintain a sustainable livestock production without depleting the natural resources (i.e., pasture). Here, we replicate the environmental and livestock submodel using NetLogo.

One year is 368 days. Seasons change every 92 days. Each day begins with the growth of grass as a function of climate and season. This is followed by updating the live weight of cows according to the grass height of their patch, and grass consumption, which is determined based on the updated live weight. After consumption, cows grow and reproduce, and a new grass height is calculated. Cows then move to the patch with less cows and with the highest grass height. This updated grass height value will be the initial grass height for the next day.

The purpose of this model is to understand the role of trade networks and their interaction with different fish resources, for fish provision. The model is developed based on a multi-methods approach, combining agent-based modeling, network analysis and qualitative data based on a small-scale fisheries study case. The model can be used to investigate both how trade network structures are embedded in a social-ecological context and the trade processes that occur within them, to analyze how they lead to emergent outcomes related to the resilience of fish provision. The model processes are informed by qualitative data analysis, and the social network analysis of an empirical fish trade network. The network analysis can be used to investigate diverse network structures to perform model experiments, and their influence on model outcomes.

The main outcomes we study are 1) the overexploitation of fish resources and 2) the availability and variability of fish provision to satisfy different market demands, and 3) individual traders’ fish supply at the micro-level. The model has two types of trader agents, seller and dealer. The model reveals that the characteristics of the trade networks, linked to different trader types (that have different roles in those networks), can affect the resilience of fish provision.

This MAS simulates the traffic of Barcelona Eixample. Uses a centralized AI system in order to control the traffic lights. Car agents are reactive and have no awareness of the intelligence of the system. They (try to) avoid collisions.

Previous research on organizations often focuses on either the individual, team, or organizational level. There is a lack of multidimensional research on emergent phenomena and interactions between the mechanisms at different levels. This paper takes a multifaceted perspective on individual learning and autonomous group formation and turnover. To analyze interactions between the two levels, we introduce an agent-based model that captures an organization with a population of heterogeneous agents who learn and are limited in their rationality. To solve a task, agents form a group that can be adapted from time to time. We explore organizations that promote learning and group turnover either simultaneously or sequentially and analyze the interactions between the activities and the effects on performance. We observe underproportional interactions when tasks are interdependent and show that pushing learning and group turnover too far might backfire and decrease performance significantly.

We present a socio-epistemic model of science inspired by the existing literature on opinion dynamics. In this model, we embed the agents (or scientists) into social networks - e.g., we link those who work in the same institutions. And we place them into a regular lattice - each representing a unique mental model. Thus, the global environment describes networks of concepts connected based on their similarity. For instance, we may interpret the neighbor lattices as two equivalent models, except one does not include a causal path between two variables.

Agents interact with one another and move across the epistemic lattices. In other words, we allow the agents to explore or travel across the mental models. However, we constrain their movements based on absorptive capacity and cognitive coherence. Namely, in each round, an agent picks a focal point - e.g., one of their colleagues - and will move towards it. But the agents’ ability to move and speed depends on how far apart they are from the focal point - and if their new position is cognitive/logic consistent.

Therefore, we propose an analytical model that examines the connection between agents’ accumulated knowledge, social learning, and the span of attitudes towards mental models in an artificial society. While we rely on the example from the General Theory of Relativity renaissance, our goal is to observe what determines the creation and diffusion of mental models. We offer quantitative and inductive research, which collects data from an artificial environment to elaborate generalized theories about the evolution of science.

We construct an agent-based model to investigate and understand the roles of green attachment, engagement in local ecological investment (i.e., greening), and social feedback.