Agent-Based Modeling for Policy-Making in Inequity Contexts

Abstract

Recent advancements in AI have allowed for more nuanced agent-based models (ABMs). These simulations offer a non-invasive way to evaluate policies in diverse and complex social contexts. Our project focuses on developing ABMs to examine the impact of legal and social norms on inequity, particularly in scenarios where systemic discrimination affects human well-being. Our research is grounded on the Capability Approach, which provides a comprehensive framework to assess inequity in terms of real opportunities, underpinning the United Nations Sustainable Development Goals (SDGs). By defining representative (i) agent profiles, (ii) agent decision-making, and (iii) agent environment, this work aims to enhance the realism of ABMs and provide valuable insights for policy-making. To achieve this, we are in the process of developing (i) a novel population synthesis method to generate agent profiles that include motivators of behaviour, (ii) a decision-making model based on Markov decision processes (MDPs) that integrates values and needs in short-term and long-term rewards, and (iii) an ABM simulation to assess the impact of norms on inequity in terms of the real opportunities of individuals, among other indicators.