CoRS speeds up flexible job-shop scheduling with smart state reuse
CoRS: Fast Neural Flexible Job-Shop Scheduling via Composite Interval State Reuse
Artificial Intelligence
Summary
Scheduling jobs on machines with changing requirements is tricky and slow for AI because it updates complex data after every small decision. The authors present CoRS, a system that saves time by reusing parts of this data across several steps and only updating the necessary parts after each action. They add features that help the system understand the job setup better, keeping schedules feasible and good quality while making decisions faster. Tests show CoRS finds quality schedules much faster than before, especially useful when there’s limited time to decide.
What this means in practice
- •For manufacturing operations teams: Generate production schedules quickly for complex factories with flexible machine assignments, improving responsiveness under tight deadlines.
- •For logistics planners: Produce adaptable job schedules in logistics hubs where machine and job flexibility affects overall throughput and timing.
Authors
Jiang Sutong, Zhao Peng, Wu Xuan, Wang Yan, Wang Jiaqi, Zhang Rui, Jiang Yuan, Zhou You, Wang Di
Abstract
Neural solvers for the Flexible Job-Shop Scheduling Problem (FJSP) typically refresh graph representations and update feasibility after every operation--machine decision. However, only feasibility must remain exact at every step, whereas graph encoding is substantially more expensive and can potentially be refreshed less frequently. We present CoRS, which decouples them: an interval-level graph representation is reused across multiple sequential decisions, while a stepwise feasibility state is updated after every action. This two-timescale design amortizes graph encoding without parallelizing action generation or relaxing scheduling constraints. CoRS further develops a feasibility-preserving composite-interval decoding scheme with dynamic constraint tracking and boundary state synchronization, which reduces representation refresh frequency while maintaining continuous feasible schedule construction. Furthermore, to alleviate representation staleness caused by interval-level state reuse, CoRS introduces enhanced static features as a compensation mechanism. By providing more comprehensive structural information of the scheduling problem, these features improve the model's perception of overall problem characteristics. Extensive experiments on synthetic and multiple public FJSP benchmarks demonstrate that CoRS significantly reduces inference time while maintaining competitive solution quality, achieving favorable speed-quality trade-offs under tight wall-clock budgets and low-latency scenarios.