A rolling horizon optimization for dynamic path planning of multiple warehouse AGVs

Automated Guided Vehicle (AGV) can improve the efficiency of warehouse picking, but the existing multi-AGV path planning method cannot efficiently solve the collision-free path in the narrow aisle environment. Therefore, a multi-AGV dynamic path planning method with an improved space-time network model is proposed. The proposed method improves the space-time network model, and decomposes the global programming problem with high computational complexity into several local programming models with low complexity, and presents a rolling horizon optimization method for an online solution. The simulation results show that the proposed method is superior to the classical multi-AGV path planning algorithm in terms of total distance and computation time.