Research on Intelligent Warehouse Path Planning Based on Multi-Objective Reinforcement Learning

In the context of intelligent warehousing, path planning is a crucial means for the efficient and safe handling of materials. Traditional path planning methods usually only consider the shortest path and fail to balance multiple objectives, such as maximizing cargo load and avoiding congestion. To address this issue, we propose a novel multi-policy multi-objective reinforcement learning approach for intelligent warehouse path planning, focusing on the path planning of Automated Guided Vehicles (AGVs) within the warehouse environment. Our method simultaneously optimizes multiple objectives, using our algorithm to achieve trade-offs among them and obtain multiple policies. Experimental results in simulated warehouse environments show that our method outperforms other state-of-the-art approaches in terms of multiple objectives. The proposed approach offers a promising solution for intelligent path planning in modern warehouses.