Automated Constraint Specification for Job Scheduling by Regulating Generative Model with Domain-Specific Representation

Advanced Planning and Scheduling (APS) systems have become indispensable for modern manufacturing operations, enabling optimized resource allocation and production efficiency in increasingly complex and dynamic environments. While algorithms for solving abstracted scheduling problems have been extensively investigated, the critical prerequisite of specifying manufacturing requirements into formal constraints remains manual and labor-intensive. Although recent advances of generative models, particularly Large Language Models (LLMs), show promise in automating constraint specification from heterogeneous raw manufacturing data, their direct application faces challenges due to natural language ambiguity, non-deterministic outputs, and limited domain-specific knowledge. This paper presents a constraint-centric architecture that regulates LLMs to perform reliable automated constraint specification for production scheduling. The architecture defines a hierarchical structural space organized across three levels, implemented through domain-specific representation to ensure precision and reliability while maintaining flexibility. Furthermore, an automated production scenario adaptation algorithm is designed and deployed to efficiently customize the architecture for specific manufacturing configurations. Experimental results demonstrate that the proposed approach successfully balances the generative capabilities of LLMs with the reliability requirements of manufacturing systems, significantly outperforming pure LLM-based approaches in constraint specification tasks. Note to Practitioners—This paper presents a practical solution for automating the conversion of raw manufacturing information into job scheduling specifications, addressing a common challenge in implementing APS systems. The proposed architecture can process diverse manufacturing documentation formats, from semi-structured route sheets to natural language instructions, while ensuring reliability through domain-specific representations. Manufacturing practitioners can use this system to reduce the manual effort in specifying digitalized constraints for production, particularly beneficial for facilities with frequent requirement changes or small-batch, multi-variety production. The system’s ability to automatically adapt to different manufacturing scenarios makes it accessible without requiring extensive programming expertise, offering a practical balance between automation and accuracy in production planning.