An adaptive dispatching approach for automated guided vehicles system

Abstract Automatic guided vehicles (AGVs) are widely used in material handling of various industries due to their advanced automation and high performance. Effective job scheduling is critical for optimizing AGV system (AGVS) efficiency. This study proposes an adaptive dispatching approach leveraging neural networks to evaluate four key factors: the AGV-job distance, job waiting time, output buffer space of current workstation, and destination input buffer space. The BP neural network dynamically updates attribute weights using real-time system data, and the Hungarian Algorithm and cost function guide AGV-job assignments. Compared with conventional rules, this method has good efficiency and performance in metrics.