A Deep Reinforcement Learning (DRL)-Based Transformer Method for Solving the Open Shop Scheduling Problem

In three linesTransformer-based deep reinforcement learning method for open shop scheduling problem (OSSP). Encoder-decoder model trained on Taillard benchmarks (4x4 to 10x10) generalizes to 40x40-100x100 instances with 12.89-15.12% gaps from lower bound, outperforming classical heuristics SPT/LPT.

Summary generated by Claude — human-verified