机器学习驱动的出口跨境电商供应链网络优化

Machine learning driven optimization of supply chain network for export cross-border E-commerce

针对出口跨境贸易中的需求、汇率、关税等的不确定性,利用丰富的电商数据,研究机器学习驱动下的出口跨境电商企业的供应链网络优化。通过构建包括海外仓、边境仓选址与库存决策在内的两阶段随机规划模型;并利用历史数据,采用机器学习的方法构建模型所需随机贸易场景,实现了从预测到决策的流程研究。将所建模型应用于某服装企业面向东南亚的跨境电商销售,并结合新近实施的《区域全面经济伙伴关系协定》,展开灵敏度分析。结果表明,随机森林模型在算例中的预测效果更优,《区域全面经济伙伴关系协定》的实施能够显著降低跨境物流费用,并有利于海外仓模式的发展。

For the uncertainties of demand,exchange rates and tariffs in export cross-border trade,the machine learning driven supply chain network optimization of export cross-border E-commerce enterprises is studied by utilizing rich-data in E-commerce.A two-stage stochastic programming model including overseas warehouse,border warehouse location and inventory decisions is developed.Based on historical data,the machine learning techniques are adopted to construct stochastic trading scenarios required by the proposed model.Thus,the‘predictive to perspective study is realized.Moreover,the model is applied to an apparel enterprise who sells its products to Southeast Asian.The sensitivity analysis is performed with the consideration of“Regional Comprehensive Economic Partnership(RCEP)”which comes into operation recently.The results show that the random forest algorithm outperforms other machine learning techniques in our case and the implementation of RCEP can bring the significant decrease of the cross-border logistics costs and the development of overseas warehouses.