Research on efficient automated guided vehicle transportation based on improved A* and DQN algorithms in solar cell shop

Efficient logistics is the goal pursued by production enterprises. Planning reasonable transportation routes quickly and formulating crossroads traffic strategies are effective methods to achieve efficient logistics. In this paper, for the problems of fast path planning and crossroads traffic in solar cell shop, firstly, the task, goal and constraint models are established. Secondly, the turning penalty length and the multi-minimum equal path straight rule are introduced, and the A* algorithm is improved to quickly plan the automatic guided vehicle (AGV) task path. Thirdly, each crossroads is regarded as an agent and the model, including state, action, reward and strategy, is built. The crossing strategy using deep Q network (DQN) learning is proposed. Finally, an information interaction mechanism between Python and Plant simulation is proposed to realize dynamic simulation analysis, which verifies the superiority of the proposed method.