Improved Genetic Algorithm Based Task Allocation for AGV Warehousing and Logistics with Adjustable Load

With the rapid development of the e-commerce economy, the demand for enhanced logistics speed has become increasingly urgent. Among various logistics processes, the transition from factory to delivery is particularly critical. Consequently, the order picking system based on multi-load automatic guided vehicles (AGVs) has drawn significant attention. Compared with single-load AGV systems, scheduling a multi-load AGV system presents a more complex challenge. Addressing the optimization of automatic picking by multi-load AGVs to minimize costs and maximize efficiency is an urgent issue. In response, we propose a task scheduling model for a multi-load AGV system. Each AGV is capable of carrying multiple picking orders, with the sequence of orders for each AGV determined by minimizing the driving path, reducing the number of AGVs, and ensuring the timely submission of ordered goods. Simultaneously, an improved multi-link genetic algorithm is proposed, which employs a novel labeling method to address the issues of genetic algorithm encoding and decoding for multi-load AGVs. We compared the performance of the algorithm across different parameters and validated its efficacy in real-world scenarios. The results demonstrate that multi-load AGVs can complete designated order picking tasks more quickly and efficiently within the specified time constraints, using fewer driving paths and AGVs. Additionally, in scenarios involving a large number of AGVs and tasks, the algorithm continued to effectively optimize task submission.