Application of Improved DQN Algorithm in Three-Dimensional Garage Scheduling

With the increase in the number of urban cars, the demand for stereo garages is increasing. Therefore, it is necessary to study the scheduling of stereo garages. Traditional stereo garage dispatching mostly uses heuristic algorithms. In recent years, machine learning has achieved good results in many fields, among which reinforcement learning has shown great potential in scheduling, has been applied to many scheduling problems, and has achieved very good results. Therefore, this paper proposes a reinforcement learning algorithm to improve experience priority—UP-DQN. In the algorithm, it proposes a way to model the state of the garage and designs a reward function, as well as an improvement to the priority of experience, so that Scheduling has a better effect. The experiment compares the three algorithms in the same environment, and the experimental results show that the algorithm has a better scheduling effect.