Decision support for building thermal comfort monitoring with a sustainable GenAI system

Arslan, Muhammad; Munawar, Saba; Mahdjoubi, Lamine; Manu, Patrick

Abstract

Staff expenses can account for up to 70% of business costs, with indoor conditions playing a critical role in employee health, behavior, and productivity. Optimal thermal comfort, typically around 21°C with 40-70% humidity, maximizes productivity. However, effective monitoring requires comprehensive data, particularly as energy regulations push for smarter building management. Although Building Information Modeling (BIM, a digital representation of a building's physical and functional characteristics) and sensor integration support facilities management, many existing systems are proprietary, expensive, and inflexible. To address these challenges, this study introduces ThermalComfortBot, a sustainable Generative Artificial Intelligence (GenAI)-powered Chatbot designed as an advanced Information System (IS). Utilizing Large Language Models (LLMs, AI models for natural language understanding) and Retrieval-Augmented Generation (RAG, a method that combines data retrieval with LLMs-generated insights), ThermalComfortBot integrates data from BIM, sensors, and other relevant sources. Built on open-source technology, it is cost-effective and fully customizable, allowing users to tailor datasets to their needs. The Chatbot delivers actionable insights through a Question-Answering (QA) interface, enabling data-driven decisions on thermal comfort to improve workplace conditions and enhance operational efficiency.

• Presentation Conference Type: Conference Paper (published)
• Conference Name: 2024 International Conference on Decision Aid Sciences and Applications
• Start Date: Dec 11, 2024
• End Date: Dec 12, 2024
• Acceptance Date: Nov 23, 2024
• Deposit Date: Nov 28, 2024
• Publisher: Institute of Electrical and Electronics Engineers (IEEE)
• Peer Reviewed: Peer Reviewed
• Public URL: https://uwe-repository.worktribe.com/output/13470785