Skip to content

Agent based modelling of LLMs power actors for thermal comfort

Notifications You must be signed in to change notification settings

kelu124/sim_chilly

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

"Simulating Consensus: An Agent-Based Model of Group Decision-Making on Thermal Preferences Using Large Language Models" - aka sim_chilly

In this paper, we introduce a novel approach to simulating group dynamics and decision-making processes within heterogeneous groups, leveraging the capabilities of Large Language Models (LLMs) to develop an agent-based model. Our focus is on a specific scenario: a group of individuals with varying thermal preferences, confined within a single room, endeavoring to reach a consensus on the room's temperature setting. This situation, seemingly mundane, encapsulates the broader challenges of group decision-making, conflict resolution, and collaborative problem-solving in diverse groups.

The complexity of human behavior, especially in group settings, presents a substantial challenge for traditional modeling approaches. However, the recent advancements in LLMs, characterized by their ability to understand and generate human-like text, offer a unique opportunity to simulate human-like decision-making and interactions more accurately. Our model employs these LLMs to represent individual agents, each with distinct thermal preferences and decision-making strategies, simulating a microcosm of a larger societal or organizational decision-making process.

Through this simulation, we aim to explore several key aspects: the dynamics of compromise and consensus-building in groups with diverse preferences, the impact of individual behavior on group outcomes, and the potential emergence of leadership or influential roles within such settings. Furthermore, we examine the role of communication patterns, persuasion techniques, and argumentation in reaching a group consensus.

Our research contributes to the fields of agent-based modeling, behavioral sciences, and the application of artificial intelligence in social simulations. By providing a detailed and nuanced understanding of group decision-making processes, this study offers valuable insights for organizational management, design of collaborative workspaces, and the development of AI systems that better understand and predict human behavior in group settings.

Example of results

    Copyright (C) 2023  Luc Jonveaux

    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation, either version 3 of the License, or
    (at your option) any later version.

    This program is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    GNU General Public License for more details.

    You should have received a copy of the GNU General Public License
    along with this program.  If not, see <https://www.gnu.org/licenses/>```