SMA
Personalify: A framework for implementing MBTI agents
Lundi 13 janvier 2025Jaime Sichman (Universidade de São Paulo, Brazil)
The use of Artificial Intelligence for the evolution of the study of human behavior has been receiving great attention from researchers over the years. The study of techniques that allow simulation of human characteristics using computers has contributed to a bet- ter understanding of several factors, including people’s personality types and behavioral preferences. In this sense, instruments such as the Myers-Briggs Type Indicator (MBTI) have been used, seeking to classify and categorize individuals’ behavioral preferences to describe their particularities and support a greater understanding of human factors in different contexts, such as in work. Studies show that human capital is one of the most critical resources for organizations that, despite this, still lack means that allow them to broadly understand the factors associated with the formation of high-performance teams. The use of tools that allow individuals greater self-knowledge of their preferences and characteristics can help in adaptations that create more harmonious environments and are favorable to teamwork, positively impacting teams’ results. In this work, we sought to explore, through agent-based simulations, the observation of behavioral aspects that can influence decisions and the consequent performance of artificial agent teams. In particular, we propose a framework called Personalify, which extends the Belief-Desire- Intentions (BDI) architecture in order to incorporate multiple decision attributes designed with characteristics described in MBTI model. The Personalify was implemented on the Gama platform, and tested in various buyers and sellers scenarios, which have similarities with work contexts, aiming to provide simulations closer to organizational realities. In this way, it is expected that the study will contribute significantly to a better understanding of elements that can influence modeled agent behavior with personalities inspired by MBTI theory, observing how certain characteristics can have a greater impact on the decisions made by agents and how these could be related to overall results presented by the team.
Keywords: Multiagent systems. Agent-based simulations. MBTI. BDI. Human Behavior. Team Building.