A flexible job shop scheduling method based on deep reinforcement learning

The traditional scheduling methods suffer from short-sightedness, long computational time, and difficulty in determining algorithm parameters in solving large-scale flexible job shop scheduling problems. Therefore, an effective deep reinforcement learning method for solving the large-scale flexible job shop scheduling problem is proposed. Firstly, the flexible job shop scheduling problem is transformed into a multi-agent Markov decision process. Secondly, an actor-critic model for solving the flexible job shop scheduling problem is constructed. The actor network outputs the scheduling rule according to the state, the agent selects the appropriate operation according to the scheduling rule, and the critic network evaluates the action of the actor network according to the state and the reward. Finally, the flexible job shop scheduling problem instances of different sizes are used to verify the performance of the method. The experimental results show that the solution quality of the method is better than the heuristic scheduling rules and the solution efficiency is better than the meta-heuristic algorithms.