Time-lag-aware deep RL agent for flexible job-shop scheduling in modular-construction factories

arXiv·low signal

An arXiv paper (2607.11725) by Zhang and Zhang applies time-lag-aware deep reinforcement learning to flexible job-shop scheduling on prefabricated-construction (PPVC) factory floors, where curing and transfer delays make naive schedulers brittle. It's a concrete example of RL 'agents' moving into gritty industrial operations rather than chat, and of modeling real-world temporal constraints that generic agent frameworks ignore. Niche, but a useful data point on where autonomous decision agents are actually being deployed.