20 March 2026 in IEEE Transactions on Consumer Electronics
This paper addresses real-time task allocation and path conflict resolution for heterogeneous agents in large-scale automated guided vehicle (AGV) systems within smart manufacturing. To overcome limitations of existing methods in dynamic and high-density environments, we propose a distributed market-based task allocation algorithm integrated with a cooperative multi-AGV obstacle avoidance strategy to ensure coordinated motion planning. A dynamic task buffer pool suppress system oscillations caused by frequent task arrivals. In addition, an incremental path caching mechanism is developed to reduce redundant computations during the 2-opt optimization process. The proposed Two-level priority mechanism effectively resolves multi-AGV collaborative obstacle avoidance and path conflict issues. Compared with existing approaches, the proposed method maintains robust conflict resolution among multiple AGVs while reducing task replanning frequency and computational cost. A standardized task allocation comparison framework is constructed to evaluate system performance under varying task scales. Simulation results based on real manufacturing layouts validate the feasibility and effectiveness of the proposed hybrid algorithm. The results provide practical guidance for efficient task scheduling and collaborative obstacle avoidance in industrial AGV systems.