Applying TLFs to design the exception thresholds in collaborative forecasting

Collaborative forecasting (CF) is an effective method that can help supply chain members to combat the bullwhip effect through the information-sharing process and collaborative tasks. To improve benefits for the members, our study presents a new and simple method for a two-echelon supply chain of a typical firm, consisting of a single manufacturer and multiple retailers. We are concerned with the out-of-bound (exception) conditions of over- and under-forecasting errors in CF, in which the Taguchi loss functions (TLFs) are used to design the exception thresholds of the forecasting errors, and to identify which retailer has first priority to deal with its out-of-bound condition by weighting the acceptable Taguchi loss of each loss element. Finally, an example of an iron and steel manufacturer is used to illustrate the proposed method.