Demand forecasting methods in a supply chain: Smoothing and denoising

A widespread forecasting method within supply chain models is the exponential smoothing method. The use of a particular forecasting method affects the costs of a supply chain. To improve the efficiency of the supply chain costs, this paper introduces the theory of wavelets. An application of this theory to the field of forecasting is wavelet denoising. Results obtained by the exponential smoothing method are compared to the results obtained by wavelet denoising. This comparison is supported by simulation experiments which include incorporation of forecasting algorithms within supply chain models. Different series of simulated data are used for testing these two methods and it is shown that wavelet denoising has an edge over the exponential smoothing method cost-wise.