Multi-Echelon Supply Chain Demand Forecast Based on Support Vector Machines

The term "bull-whip effect" refers to a phenomenon that demand variability amplifies while one moves up a supply chain. Alleviating the bullwhip effect can remarkably enhance the overall benefit of the supply chain. For the past decade, alleviating and controlling the bullwhip effect has been the international hotspot, in which the demand forecast level is the direct cause that induces the bullwhip effect. To enhance the demand forecast level in each echelon of the supply chain, there are forecast methods, such as time series and artificial neural network, adopting in the literature. In this paper, the demand forecast of multi-echelon supply chain is based on support vector machines (SVM) and all the stages adopt the method; combined with the strategy of information sharing, they will enhance the overall forecast level, alleviate the bullwhip effect and improve the supply chain performance. The numerical value of supply chain demand forecast based on SVM is got through the computation and compared with artificial neural network's.