BuildEvo: Designing Building Energy Consumption Forecasting Heuristics via LLM-driven Evolution

2025-07-16

AI Summary

Existing building energy consumption forecasting methods suffer from low accuracy, poor generalizability, or limited interpretability: physics-based models oversimplify thermal dynamics, while data-driven approaches lack transparency and physical consistency. To address this, we propose the first large language model (LLM)-driven evolutionary framework that automatically synthesizes physically plausible, high-accuracy heuristic rules. Our method embeds building thermodynamic priors into LLM prompts and employs an evolutionary algorithm to iteratively optimize both rule structure and parameters, enabling end-to-end interpretable modeling on real-world operational data. Evaluated on multiple public benchmarks, our approach significantly outperforms state-of-the-art models—reducing mean absolute error by 23.6% on average—while generating human-readable, physically verifiable explicit prediction logic. It thus achieves a balanced trade-off among predictive accuracy, cross-scenario generalizability, and model interpretability.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Planning, Routing, and Scheduling: Planning with Language Models
• Natural Language Processing: (Large) Language Models

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Accurate building energy forecasting is essential, yet traditional heuristics often lack precision, while advanced models can be opaque and struggle with generalization by neglecting physical principles. This paper introduces BuildEvo, a novel framework that uses Large Language Models (LLMs) to automatically design effective and interpretable energy prediction heuristics. Within an evolutionary process, BuildEvo guides LLMs to construct and enhance heuristics by systematically incorporating physical insights from building characteristics and operational data (e.g., from the Building Data Genome Project 2). Evaluations show BuildEvo achieves state-of-the-art performance on benchmarks, offering improved generalization and transparent prediction logic. This work advances the automated design of robust, physically grounded heuristics, promoting trustworthy models for complex energy systems.

Problem

• Improving building energy forecasting accuracy and interpretability
• Automating heuristic design using LLMs and evolutionary processes
• Enhancing generalization by incorporating physical building insights

Innovation

• LLM-driven evolution for heuristic design
• Incorporates physical insights from building data
• Achieves state-of-the-art transparent performance