Support vector regression and Adaptive neuro fuzzy to measure the Bullwhip effect in supply chain

Abstract

Support Vector Regression (SVR) and Adaptive Neuro-Fuzzy (ANFIS) are the advance statistic knowledge. They are employed to construct the forecasting signal models for a tin milk industry. Then the models are used to measure the Bullwhip Effect (BWE) in the supply chain with the beer game methods. This study tried to minimize the BWE with forecasting methods through the comparative for both methods and attempt to prove that a small Mean Square Error (MSE) of the BWE. The data is collected from the price and demand variables, then it is generated in random normal distribution. It informs that in two conclusions. ANFIS is better than SVR in case of MSE comparison. Moreover, the totals BWE of ANFIS and SVR are 209 and 1,237 respectively. The Second conclusion is the better forecast model leads to reduce the amplification across the supply chain.