Deep reinforcement learning and biomechanical modeling are integrated to optimize the scheduling problem of intelligent logistics and warehousing robots

This study introduces an innovative approach to optimizing the scheduling of intelligent logistics and warehousing robots by integrating deep reinforcement learning (DRL) with biomechanical modeling. Leveraging a comprehensive dataset from a large-scale logistics company, the research formulates the scheduling problem as a Markov Decision Process (MDP) and incorporates biomechanical principles to accurately model robot energy consumption. A Deep Q-network (DQN) is employed to learn the optimal scheduling policy, which is further refined using policy gradient optimization. This integrated framework aims to maximize task completion efficiency while minimizing energy usage, addressing the complexity of balancing these competing objectives. Extensive simulations validate the proposed approach, demonstrating significant improvements in task completion rates, average travel distances, and energy consumption compared to baseline algorithms such as random scheduling and greedy algorithms. The methodology presents a robust and efficient solution for enhancing operational efficiency in intelligent logistics and warehousing systems.