Abstract: Abstract: Steel logistics platforms often need to split steel products into multiple shipping bills for transportation when handling customer orders. Less-than-truckload (LTL) cargo, which fails to meet the minimum load requirements of a truck, needed to be consolidated with goods from other customer orders to optimize transportation efficiency. Although previous studies had proposed some solutions, none had simultaneously considered the issues of detour distances and prioritizing high-priority cargo in consolidated shipments. Therefore, a multi-objective optimization framework for steel cargo consolidation under multiple constraints was proposed in this study. The globally optimal cargo consolidation decisions were achieved by designing a hierarchical decision network and a representation enhancement module. Specifically, a hierarchical decision network based on Proximal Policy Optimization (PPO) was used to determine the priorities of three optimization objectives first, and then the LTL cargo was consolidated and selected based on these priorities. Meanwhile, a representation enhancement module based on Graph Attention Network (GAT) was employed to dynamically represent cargo and LTL cargo information, which was then input into the decision network to maximize long-term multi-objective gains. Experimental results on a large-scale real-world freight dataset show that this method improves the weight ratio of high-priority cargo and reduces average detour distances while slightly sacrificing the total weight of tail cargo shipments, effectively enhancing the efficiency of consolidated transportation.

Key words: consolidation decision-making, Markov decision process, Proximal Optimization(PPO), Graph Attention Network(GAT), Decision optimization

摘要： 摘 要: 钢铁物流平台在处理客户订单时，常需将钢材产成品拆分成多个运单进行运输。而未达到货车最低载重限制的“尾货”，往往需要与其他客户订单的货物拼载以优化运输效率。尽管先前的研究已经提出了一些解决方案，但均未能同时考虑拼货运输中可能产生的绕行距离及高优先级货物优先发运问题。为此，本研究提出一个多约束条件下多目标优化的钢铁拼载决策框架。该框架通过设计分层决策网络和表征增强模块，实现全局最优的拼货决策。具体而言，采用基于近端策略优化的分层决策网络，先确定各个优化目标的优先级，再基于这些优先级进行尾单的组合与选择。同时，利用基于图注意力的表征增强模块来实时表征货物信息和尾货信息，并输入决策网络实现多目标的长期收益最大化。在大规模真实货运数据集上的实验结果表明，该方法能在略微牺牲尾货发运总重量的前提下，提升高优先级货物重量占比并减少平均绕行距离，有效提升了拼载运输的效率。

关键词: 拼货决策, 马尔可夫决策过程, 近端策略优化, 图注意力网络, 决策优化