Text2Zinc introduces a cross-domain dataset for translating natural language descriptions into formal optimization and constraint models.

• Creates a unified framework for both satisfaction and optimization problems
• Enables LLMs to act as co-pilots for complex optimization tasks
• Supports diverse problem domains using the MiniZinc constraint modeling language
• Advances automation of formal model creation from informal requirements

This research significantly enhances engineering capabilities by reducing the expertise barrier for implementing optimization solutions, enabling more efficient resource allocation and complex system design.

Text2Zinc: A Cross-Domain Dataset for Modeling Optimization and Satisfaction Problems in MiniZinc