Based On Bayesian Updating Short Life Cycle Product Demand Forecasting And Inventory Control

In modern society, products with short life cycles are becoming increasingly common in personal computer (PC) and mobile phone industries that are prone to value depreciation. This kind of products have such characteristics as lacking of historical demand data, quickly value depreciation and fast product substitution. It is because of these features making the traditional forecasting and inventory strategy inaccurate and leading to inaccurate demand forecasts, inventory costs increasing, service levels decreasing when applying to the short life cycle product demand forecast and inventory control.This paper analyzes the defects on traditional forecasting methods to the short life cycle product demand forecasting, and presents a dynamic demand forecasting algorithm based on Bayesian parameter updating. Firstly, the algorithm analyzes the knowledge structure of the product life cycle and selects the Bass model to model the demand. Secondly, the algorithm analyzes the previous similar products and extracts the initial parameters based on 1stopt software and calculates the initial demand of short life cycle products. With the launch of the product, the actual demand data become available, it updates the model parameters on the basis of actual demand data by the Bayesian method and produces the rolling demand forecast to improve model prediction accuracy, and the model is revised with seasonal factor. The numerical illustration proves that our dynamical forecasting algorithm could improve model prediction accuracy for products with short life cycle and better other methods on MAD, RMSE and MAPE.Based on the above rolling demand forecasting, this article combines newsboy model and part of the period balance method to control inventory of the short life cycle products. It uses newsboy model give the optimal order quantity of each order cycle, then determines the optimal order point in time through part of the period balance method, and makes dynamic adjustments to the ending stocks based on the optimal order quantity and the optimal order point of each order cycle. The numerical illustration proves that our method reduces the obsolete stocks and lack stocks.