Self-Supervised Job Grouping for Parallel Machine Scheduling with Family Setups

This study addresses the parallel machine scheduling problem with family setups in large-scale manufacturing, motivated by challenges in the air conditioner assembly process, aiming to minimize total tardiness and setup costs. The proposed framework integrates self-supervised learning for job grouping and incorporates tardiness evaluator for iterative refinement. Experimental evaluations on datasets inspired by real-world scenarios demonstrate that the proposed framework effectively reduces total tardiness and setup costs compared to baseline methods. These results highlight the framework’s scalability, computational efficiency, and potential as a practical solution to real-world scheduling challenges.