[4R3-OS-8b] OS-8

[4R3-OS-8b-04] Optimization of Emergent Opinions and Decision-Making Processes in Small Groups through Reinforcement Learning.

〇Shota Shiiku^1,2, Yugo Takeuchi¹ (1. Shizuoka University, 2. Max Planck Institute for Empirical Aesthetics)

[[Online]]

Keywords:Decision-Making, Emerge Opinion, Multi Agent System, MCGDM, Small Groups

In the decision-making process within small groups, members often have numerous opportunities to explicitly express their own opinions and thoughts. This contrasts with larger groups, where decisions are more frequently made by majority vote or delegated to specific individuals. As a result, small groups may exhibit more complex and emergent interactions, such as compromises among members and the proposal of new ideas. However, these creative aspects of interaction, commonly observed in small groups, have not received much attention to date. Therefore, predicting how members of a small group will react or behave towards the conclusions reached after decision-making is challenging. This study constructs a reinforcement learning model that incorporates the satisfaction of individual members with the interactions and outcomes during the decision-making process of small groups. It reveals that this approach leads to members positively accepting the conclusions reached by the group while also reducing the time required to reach these decisions during the decision-making process.