A NOVEL STATISTICAL PREDICTION TECHNIQUE BASED ON THE DYNAMIC RELATIONSHIP IDENTIFICATION ALGORITHM TO FORECAST SUPPLY CHAIN DEMAND

作者：Cao, Jian^*; Chen, Yangyang; Zhou, Gengui

来源：Economic Computation and Economic Cybernetics Studies and Research, 2015, 49(1).

• Supply chain demand forecasting
• Dynamic relationship model
• Identification algorithm
• Forecasting precision
• Adaptive modeling techniques

A key driver for the success of the supply chain (SC) is effective customer demand forecasting. However, increasingly individualized demands and fierce competition between SCs create more uncertainty of predicted results in Mainland China nowadays. Thereby, appropriate, forecasting techniques are critically important for the SC's decision-makers to analyze and uncover the patterns of historical correlative data of the demands and then project those patterns into the future. A novel statistical forecasting method based on the dynamic relationship identification algorithm (SF-DRIA) was presented in this paper Combined with the adaptive modeling techniques, the dynamic relationship model reflecting the relational pattern between the forecasting result and its correlative influencing variables was constructed initially. Then, in accordance with the continuously optimized forecasting precision and the newly added correlative data, the parameter estimates and the structure coefficients of the model were regulated to achieve the optimal forecasting results. A case performed by the core enterprise in a supply chain reveals that the proposed method produces higher accurate forecasting and better tracking capability to the trend of the demands compared with several traditional statistical forecasting approaches. Moreover; the high robustness of the SF-DRIA method is guaranteed by the employment of the forecasting precision.

• 出版日期2015
• 单位浙江工业大学