From User Preferences to Optimization Constraints Using Large Language Models

2025-03-27

AI Summary

This study addresses the critical challenge of automatically translating users’ natural-language preferences—elicited in Italian renewable energy communities (RECs)—into formal constraints for smart-home energy optimization. Method: We propose the first LLM-driven preference-to-constraint translation framework and introduce, to our knowledge, the first open-source Italian benchmark dataset of aligned preference-constraint pairs, alongside reference implementation code. Using zero-shot, one-shot, and few-shot prompting strategies, we systematically evaluate multiple Italian-capable LLMs—including GPT-4 and Claude series—on constraint generation performance. Contribution/Results: Experiments demonstrate that state-of-the-art Italian LLMs possess foundational capability for this task, with GPT-4 and Claude models significantly outperforming others in both accuracy and robustness. The study also identifies key domain-specific adaptation challenges and practical limitations. Our work establishes a reproducible empirical paradigm and foundational infrastructure for LLM-augmented human-AI collaboration in energy system optimization.

Technology Category

• Humans and AI: Learning Human Values and Preferences
• Machine Learning: Large Multimodal Models (LMMs)
• Constraint Satisfaction and Optimization: Constraint Learning and Acquisition

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

This work explores using Large Language Models (LLMs) to translate user preferences into energy optimization constraints for home appliances. We describe a task where natural language user utterances are converted into formal constraints for smart appliances, within the broader context of a renewable energy community (REC) and in the Italian scenario. We evaluate the effectiveness of various LLMs currently available for Italian in translating these preferences resorting to classical zero-shot, one-shot, and few-shot learning settings, using a pilot dataset of Italian user requests paired with corresponding formal constraint representation. Our contributions include establishing a baseline performance for this task, publicly releasing the dataset and code for further research, and providing insights on observed best practices and limitations of LLMs in this particular domain

Problem

• Translate user preferences into energy optimization constraints
• Evaluate LLMs for Italian natural language processing
• Establish baseline performance for constraint conversion

Innovation

• LLMs translate user preferences into constraints
• Zero-shot, one-shot, few-shot learning evaluated
• Dataset and code released for research