A Complex Network-based Dialogue Hybrid Intelligent Information System for Solving Tasks in Thermal Power Engineering
Evgeny A. Belousov, Danila R. Gromozdov, Dmitry V. Aladin, Marina O. Ponomareva, Nikita S. Klimov, Nikolay M. Gorkunov · 2025
This paper presents a complex network-based dialogue Hybrid Intelligent Information System (HIIS) that utilizes complex networks to solve tasks in the field of thermal power engineering. As the volume of data in this area increase, the need for advanced tools that facilitate problem solving and enhance operational efficiency becomes evident. The proposed HIIS integrates large language models (LLMs) with a robust architecture that enables natural language interaction while effectively addressing complex technical tasks. The system's architecture includes three components: input, output, and a complex network-based inference, ensuring seamless communication and precise data analysis. Despite the advantages of LLMs, the paper discusses challenges such as model hallucination and the “black box” phenomenon, which are addressed to ensure transparency and reliability. This innovative approach aims to provide practical recommendations and insights, ultimately assisting professionals in making informed decisions in the dynamic field of thermal power engineering.