Institute of Transportation, MOTC

Shu-Hsien Li, 2008.06
Graduate Institute of Transportation Science, Tamkang University

　　Under the keenly competitive environment and avoid to waste cost by bullwhip effect, the enterprises beginning to join the supply chain collaboration. The recent collaborative initiative, termed Collaborative Planning, Forecasting, and Replenishment (CPFR), has begun to gain wide acclaim for the benefits it delivers.

　　The new evolution of CPFR is to extend the core elements to include the transportation component, termed Collaborative Transportation Management (CTM). CTM is a holistic process that improve the operating performance of all parties involved in the relationship by eliminating inefficiencies in the transportation component of the supply chain through collaboration. CTM shipment forecasting and freight demand forecasting are critical foundation in the CTM business process, that are prerequisite to carriers’ tactical and operational planning, such as network planning, routing, scheduling, and fleet planning and assignment. However, few literatures have been paid to the forecasting modeling for CTM. This study attempts to develop a series of forecasting models for shipment and freight demand forecasting under the CTM framework.

　　This study extends and improves grey forecasting theory and constructs hybrid models to develop a series of shipment forecasting and freight demand forecasting models for CTM. In shipment forecasting, consider different collaborative frameworks, both grey systematic forecasting and grey time-series forecasting are developed. This study first attempts to integrate the grey number in forecasting models, in order to analyze shipment forecasting under partical information sharing in  CTM framework. Furthermore, an aggregated freight demand forecasting model was also developed. This study then use grey calamity forecasting model to predicting the shipment exceptions. A case study with an IC (Integrated Circuit) supply chain and other relevant data was provided to illustrate the results. These models are shown to be more accurate prediction results than multiple regression, ARIMA and neural network models, as well as shipment exception forecasting. Finally, the results indicate that the more information sharing under CTM, the carriers can predict more accurately.

 　　This study demonstrates how the proposed forecasting models might be applied to the CTM system and provides as the model theoretical basis for the forecasting module developed for the CTM.