基于DEA效率嵌入的数字化供应链动态供需预测模型研究

Research on Dynamic Supply and Demand Prediction Model of Digital Supply Chain Based on DEA Efficiency Embedding

摘要: 本文聚焦数字化供应链，针对传统供需预测模型在数据维度、效率量化和动态性方面的不足，构建“DEA效率嵌入预测”模型。通过将DEA效率评估与供需预测模型相结合，形成“效率–预测”协同框架，利用动态窗口DEA (BCC模型)捕捉供应链节点效率时序变化并融入预测模型。研究选取某零售企业3个区域仓库2023年1~6月数据进行实证分析，结果表明该框架提升预测准确性，在需求波动期预测误差(RMSE)较传统模型降低15%~20%，资源配置冗余率减少10%~15%。同时，通过企业内部多源数据构建的数据协同机制，增强了供应链的动态调整和风险应对能力。研究还指出该框架在数据获取、模型复杂性与适用性以及对突发事件考虑等方面存在局限性，并提出引入强化学习、区块链和数字孪生技术等未来研究方向，为供应链管理提供新视角和决策工具，对特定规模零售企业的管理优化提供局部经验，行业政策制定需结合多规模企业验证。

Abstract: This paper focuses on the digital supply chain. Aiming at the deficiencies of traditional supply-demand forecasting models in data dimensions, efficiency quantification, and dynamics, a “DEA efficiency-embedded forecasting” model is constructed. By combining the DEA efficiency evaluation with the supply-demand forecasting model, an “efficiency-forecasting” collaborative framework is formed. The dynamic window DEA (BCC model) is used to capture the time-series changes of supply chain node efficiency and integrate them into the forecasting model. The research selects the data of three regional warehouses of a retail enterprise from January to June 2023 for empirical analysis. The results show that this framework significantly improves the forecasting accuracy. During the demand fluctuation period, the forecasting error (RMSE) is reduced by 15%~20% compared with traditional models, and the resource allocation redundancy rate is reduced by 10%~15%. At the same time, the data collaboration mechanism constructed by integrating multi-source heterogeneous data enhances the dynamic adjustment and risk-response capabilities of the supply chain. The research also points out the limitations of this framework in data acquisition, model complexity and applicability, and consideration of extreme events, and proposes future research directions such as the introduction of reinforcement learning, blockchain, and digital twin technologies. This provides a new perspective and decision-making tools for supply chain management, and has important guiding significance for enterprise management and industry policy-making.