Leveraging generative AI for knowledge-driven information retrieval in the energy sector

Department of Design Manufacturing and Engineering Management, University of Strathclyde, 75 Montrose Street, Glasgow G1 1XJ, Scotland, UK

^* Corresponding author: a.alsayegh@strath.ac.uk

Abstract

This paper presents an innovative approach to knowledge management in the energy sector through the development of the Advanced Agent Architecture (AAA). AAA integrates Retrieval-Augmented Generation (RAG) techniques with a tailored local knowledge base (LKM) and web search functionalities, aiming to enhance the accuracy, robustness, and flexibility of information retrieval. We conducted a detailed case study involving a solar power system to evaluate the effectiveness of AAA compared to traditional Large Language Models (LLMs) such as Llama 3. Our results demonstrate that AAA significantly outperforms conventional methods in delivering accurate and relevant answers to complex domain-specific queries. However, the system also shows higher energy consumption and slower response times, identifying critical areas for future research. This study sets the stage for further exploration into optimizing AAA’s energy efficiency and processing speed, expanding the range of queries, and providing a more comprehensive benchmarking against traditional systems. Our findings indicate that AAA has the potential to substantially improve knowledge management practices, facilitating more informed decision-making and operational efficiencies in the energy sector.