Fairness in multi-criteria decision-making sorting problems

• Grant number: 24/18109-3
• Support Opportunities: Regular Research Grants
• Start date: May 01, 2025
• End date: April 30, 2027
• Field of knowledge: Engineering - Production Engineering - Operational Research
• Principal Investigator: Renata Pelissari Infante
• Grantee: Renata Pelissari Infante
• Host Institution: Escola de Engenharia (EE). Universidade Presbiteriana Mackenzie (UPM). São Paulo , SP, Brazil

Abstract

Recent studies have shown that algorithmic decision-making can be inherently prone to injustice. In fact, the increasing use of machine learning (ML) algorithms for decision-making in various fields has highlighted the problem of unintentionally replicating historical discrimination embedded in data related to sensitive attributes such as gender, race, skin color, and nationality. Although the discussion on algorithmic fairness has primarily been developed on the field of ML, other research areas, such as multi-criteria decision analysis (MCDA), have started to explore the topic. Since in MCDA decisions are based on the opinions of decision-makers rather than training data, bias can arise from the discriminatory views of the decision-makers themselves. So far, in the context of MCDA, most studies have focused on ranking problems, leaving challenges related to multi-criteria sorting unexplored. The overall objective of this research project is to develop new MCDA sorting methods capable of addressing fairness issues. (AU)