Revisiting the Effects of Forecasting Method Selection and Information Sharing Under Volatile Demand in SCM Applications
研究发现在产能紧张度低时，供应商自身预测比订单信息共享更有效；产能紧张度高时信息共享才有价值。先进预测方法（如GARCH、神经网络）在多数情况下能显著降低成本，但模型误设会导致系统性能下降。
Using a two-stage capacitated supply chain with four retailers and one supplier, this paper reveals that volatile demand and the supplier's own forecasting intelligence may limit the value of information sharing (IS). The results show that capacity tightness (CT) is the most important factor that affects the performance of the supply chain compared with the effects of forecasting methods and IS policies. When CT is low, order information sharing is not preferred, because the supplier's own forecasts are more beneficial. When CT is high, IS becomes valuable. In addition, this study demonstrates that advanced forecasting methods such as generalized autoregressive conditional heteroscedasticity (GARCH) and properly configured neural network models significantly reduce costs relative to the conventional forecasting methods under most scenarios examined. However, misspecified models often result in poor system performance. The findings also reveal significant interaction effects among forecasting method, IS, and CT. In order to achieve cost reduction, supply chain managers should jointly consider all the critical factors when selecting forecasting method and IS policy.