Forecasting orders accurately plays a crucial role for household electric appliance (HEA) enterprises keeping low inventory level. In order to reduce the influence of bullwhip effect, sales data are applied to forecas...Forecasting orders accurately plays a crucial role for household electric appliance (HEA) enterprises keeping low inventory level. In order to reduce the influence of bullwhip effect, sales data are applied to forecast orders of the next period instead of ordering data. Mobile agent is applied to achieve the sales data from retailers in the data preparation. And the converting approach from sales data of retailers to prediction data is proposed. Autoregressive and moving average (ARMA) model is introduced to forecast orders and a comparison amongst final prediction error (FPE), Akaike information criterion(AIC), Bayes information criterion(BIC) and Akaike's corrected information criterion(AICC) criterion is shown. The sample test verifies the superiority of AICC; therefore it is applied to identify ARMA order. Forecasting architecture is established and then the prototype system is tested, at last a case shows the orders prediction of the next quarter and verifies the effectiveness of the proposed method.展开更多
基金Hubei International Cooperation Projects,China(No.2007CA008)
文摘Forecasting orders accurately plays a crucial role for household electric appliance (HEA) enterprises keeping low inventory level. In order to reduce the influence of bullwhip effect, sales data are applied to forecast orders of the next period instead of ordering data. Mobile agent is applied to achieve the sales data from retailers in the data preparation. And the converting approach from sales data of retailers to prediction data is proposed. Autoregressive and moving average (ARMA) model is introduced to forecast orders and a comparison amongst final prediction error (FPE), Akaike information criterion(AIC), Bayes information criterion(BIC) and Akaike's corrected information criterion(AICC) criterion is shown. The sample test verifies the superiority of AICC; therefore it is applied to identify ARMA order. Forecasting architecture is established and then the prototype system is tested, at last a case shows the orders prediction of the next quarter and verifies the effectiveness of the proposed method.
