A New Hybrid Prediction Reduces the Bullwhip Effect of Demand in a Three-level Supply Chain

In this paper, we present a new predictive hybrid model using discrete wavelet transform (DWT), and the artificial neural network (ANN) to reduce the bullwhip effect of demand in supply chain to obtain a real amount of final customer demand. Also, we compare our result with more comprehensive sample of previous research to extend the scope of our study. In this new research our methodology is combine two discrete wavelet transform (DWT), and the artificial neural network (ANN) was used to analyze the data. Results indicated that in comparison with the previous methods of prediction to reduce the bullwhip effect in supply chains, the use of DWT and ANN is more favorable leading to less error against other methods. Moreover, we discrete our data in liner data and nonlinear data because since the combinational method uses nonlinear data and gives importance to these data rather than linear data, it can be concluded that in comparison with linear data, nonlinear data have more importance in predicting the bullwhip effect. According to this new combinational technique, organizations can obtain suitable amounts of demand at all stages of supply chain, which makes a low distance between true and forecasting demands. Therefore, organizations can avoid some costs that playing an inessential role in their products.

Keywords:

Supply Chain , Bullwhip Effect , Demand , Artificial Neural Network , Discrete Wavelet transform