A Deep Reinforcement Learning Method Based on a Transformer Model for the Flexible Job Shop Scheduling Problem | Synapse