Multi-AGVs dispatching strategy in automobile assembly line based on Deep Reinforcement Learning

Abstract

AGVs (Automated Guided Vehicles) are widely used in automobile assembly lines, optimizing the AGVs dispatching strategy in automobile assembly line is of important academic significance and application value. In actual production process, assembly lines require multiple types of AGVs to work in coordination, which requires a intelligent dispatching strategies to solve this complex dispatching problem and can be applied to different situations. Hence, a deep reinforcement learning method based on DQN algorithm is proposed to optimize the dispatching strategy. Appropriate reward function is designed according to the demand materials of every station, and experience replay mechanism is used to improve the algorithm performance. In addition, this method is applied to AGVs dispatching simulation, and the results are given.