Research On The Tobacco Demand Combined Forecast Model Based On Supply Chain

The tobacco industry is under the background of the reform of "supply by orders", business collaboration and precision marketing. So the tobacco commercial enterprises need to accurately grasp the demand change trend, and establish precise cigarette demand prediction system. Accurate prediction of cigarette demand based on supply chain is conducive to the promotion of business cooperation, can improve the ability to predict cigarette demand, so as to determine reasonable order quantity and inventory levels, seeking optimal balance between customer satisfaction and inventory level. So the tobacco industrial and commercial enterprises in the supply chain can organize the procurement, production and distribution activities orderly and effectively, and then improve the overall supply chain efficiency.This thesis introduces the basic cigarette demand forecast theory based on supply chain, summarizes the characteristics of demand for cigarettes and four basic elements of cigarette demand forecast, and then puts forward cigarette demand forecasting process and steps on the basis of analysis of the impact of the comprehensive factors on cigarette demand. At last it analyzes the process evaluation index.Then this paper researches on cigarette demand forecasting theory and method, from qualitative and quantitative respects of current cigarette demand forecasting methods commonly used in research. Through summarizing the advantages and disadvantages of different prediction methods and applicability to cigarette demand forecast, it lays the theoretical foundation on choosing a suitable model for certain city.Then this paper takes a tobacco company as an example to analyze the influence factors of cigarette demand forecast, constructs the forecast index system, and then selects15quantifiable indicators from the index system to take correlation analysis and grey relational analysis. The thesis establishes a multiple linear regression model and the improved grey system model for cigarette demand forecasting, and then take the predictive results as the input layer of BP neural network model to the combined predict model, adjust and correct the model. The results show the prediction accuracy has been greatly improved and the prediction error significantly decreased. The thesis enriches cigarette demand forecasting method based on supply chain, which enables the development of cigarette demand plan close to the actual demand volume, and provides effective demand forecasting method for the city tobacco company decision makers.