Dynamic Adjustment of AGV Path Based on Particle Swarm OptimizationDynamic Adjustment of AGV Path Based on Particle Swarm Optimization

Ren Le^*, Wang Chaofan

Shanghai Maritime University

Abstract：

As the demand for intelligent port operations grows, dynamic path planning for AGVs (Automated Guided Vehicles) has become key to improving operational efficiency. Traditional AGV scheduling methods lack flexibility when dealing with unexpected situations, such as ships arriving early, making it difficult to achieve overall system optimization. To address these issues, this paper proposes a method for dynamic AGV path adjustment based on improved particle swarm optimization (PSO). It constructs a “macro-micro” hierarchical optimization framework that decouples task assignment from path execution and establishes a mixed-integer programming model aimed at maximizing task completion rates and minimizing the average travel distance of AGVs. By designing a hierarchical PSO architecture, this method achieves collaborative optimization and conflict resolution between task allocation and path planning, effectively enhancing system robustness and operational efficiency. It holds significant value for application in various scenarios, including automated terminals and smart warehouses, as it can improve the efficiency of logistics operations and reduce operational costs in these environments.

Key Words：

AGV; particle swarm optimization; orchestration layer; microscopic path layer; DQN