Intelligent scheduling of parking lots based on improved DQN

With the development and popularity of automated parking lots, the problem of high energy consumption generated by unreasonable scheduling of parking robots has become increasingly prominent. In order to solve the scheduling problem of automatic parking lot, an intelligent scheduling strategy based on deep Q-network algorithm is proposed to solve the problem of high energy consumption caused by unreasonable scheduling in automatic parking lot. By introducing double deep Q-network architecture and advantage function, the scheduling strategy of the robot is optimized. In this study, the time difference error is used as the weight of the sample cache and the negative energy consumption value is used as the reward function to evaluate the performance of the improved deep Q-network algorithm and the traditional algorithm. The simulation results show that the improved deep Q-network algorithm performs well in vehicle scheduling optimization with a reward value of 46.8. By analyzing the distribution of vehicles in the garage under different algorithms, the effectiveness of the improved deep Q-network algorithm in preferentially utilizing high-quality garage resources is further verified. It demonstrates how the enhanced deep Q-network algorithm can efficiently optimize the automated parking lot’s scheduling strategy while utilizing the premium garage to lower operational energy usage.