A Hybrid Floyd Algorithm for Multi-Agent System in Warehouse

The Floyd-Warshall algorithm, traditionally used for finding all-pairs shortest paths, faces challenges in large-scale warehouse settings due to its high time complexity. This paper proposed a hybrid improved Floyd-Warshall algorithm integrated with a path time window model for multi-agent systems in warehouse environments. Optimizing the Floyd algorithm by simplifying path network connections through the use of unidirectional agent routes and introducing a time window model to predict and avoid conflicts between agents. The integration of these techniques allows for the efficient calculation of shortest paths, minimizing collisions and improving task completion times. The effectiveness of the algorithm is verified through MATLAB simulations, demonstrating significant improvements in collision avoidance and overall task execution efficiency, with reductions in both path length and total completion time.