Global Constraint LLM Agents for Text-to-Model Translation

2025-09-10

AI Summary

Automatically translating natural language specifications into MiniZinc models faces significant challenges due to the high cognitive load of integrating logical reasoning and constraint programming knowledge, resulting in low modeling accuracy. To address this, we propose a multi-agent collaborative framework that decomposes the modeling task based on global constraint categories. Specialized large language model (LLM) agents independently identify and generate corresponding constraint submodules, which are then synthesized into a complete MiniZinc model by an integration agent. This division of labor reduces per-agent reasoning complexity while enhancing modeling accuracy and scalability. Experiments across multiple state-of-the-art LLMs demonstrate that our approach consistently outperforms baseline methods—including single-shot and chain-of-thought prompting—achieving an average 23.6% improvement in modeling accuracy on standard benchmarks. These results validate the effectiveness and generalizability of our constraint-driven task decomposition mechanism.

Technology Category

• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories
• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Agent Architectures

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
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Natural language descriptions of optimization or satisfaction problems are challenging to translate into correct MiniZinc models, as this process demands both logical reasoning and constraint programming expertise. We introduce a framework that addresses this challenge with an agentic approach: multiple specialized large language model (LLM) agents decompose the modeling task by global constraint type. Each agent is dedicated to detecting and generating code for a specific class of global constraint, while a final assembler agent integrates these constraint snippets into a complete MiniZinc model. By dividing the problem into smaller, well-defined sub-tasks, each LLM handles a simpler reasoning challenge, potentially reducing overall complexity. We conduct initial experiments with several LLMs and show better performance against baselines such as one-shot prompting and chain-of-thought prompting. Finally, we outline a comprehensive roadmap for future work, highlighting potential enhancements and directions for improvement.

Problem

• Translating natural language descriptions into correct MiniZinc models
• Addressing logical reasoning and constraint programming expertise requirements
• Decomposing modeling tasks by global constraint type with specialized agents

Innovation

• Agentic framework with multiple specialized LLM agents
• Decomposes modeling task by global constraint type
• Generates and assembles constraint snippets into complete model