Research On Demand Forecast And Bullwhip Effect Of Supply Chain In Big Data

With the rise of big data and the Internet technology rapidly merges into various fields,it has made new changes in the structure of the supply chain and reclaimed some new directions for the development of the supply chain.Due to the faster and more convenient acquisition of data in big data,it is possible for enterprises to achieve information sharing in the new supply chain.In the supply chain management,demand forecasting is an essential part of the whole supply chain.When there is a lack of accurate market demand information,the bullwhip effect that generated in the process of information transfer to the upstream of the supply chain will affect the coordination and control of each link in the supply chain,greatly weakening the competitiveness of the whole supply chain.However,traditional forecasting methods have been unable to obtain satisfactory prediction accuracy under the background of big data,which has brought great challenges to the supply chain management.Based on the above problems,this paper propose the research topic of the supply chain demand forecasting and the bullwhip effect in big data,which has profound theoretical and practical significance for adapting to the new market demand changes,improving the core competitiveness of enterprises,and achieving the best overall efficiency of the supply chain.This paper combines supply chain management with big data to build a type of supply chain model in big data.Based on this,we further analyse supply chain demand forecasting and the bullwhip effect.The main innovations and specific work of the paper are as follows:(1)According to the characteristics of the supply chain in the background of big data,we build a new supply chain system,and introduce Agent theory to abstractly simulate the supply chain structure model in big data with Multi-Agent system.(2)According to the structure of the supply chain and the characteristics of the market demand in the background of big data,we propose a support vector machine prediction model based on genetic algorithm,Firstly the general traditional model of support vector machine is established to discuss the influence of parameters in the model on the fitting prediction.Then on the basis of fully considering the characteristic which enterprise has to quick response to market demand and the large prediction error of the traditional predictive model in big data,we use a genetic algorithm to optimize and improve the parameters of the original model,and construct the vector machine prediction model based on genetic algorithm.Finally,we verify the prediction performance of the new model and the effect on the bullwhip effect in the supply chain by means of the fitting prediction of the sales quantity of an electronic product.(3)According to the characteristics that many factors influence the market demand in the supply chain under the background of big data,the information entropy theory is adopted to select the main characteristic factors and to consider the mutual influence of multiple variables.On the basis of this,we propose a demand forecasting method based on multivariable support vector machines to maximize the mining of a variety of effective information.Because the prediction model of the support vector machine lacks generalization capability and has a high prediction error in high dimensions,this paper uses BP neural network to adjust the error of prediction results for the multivariable support vector machines.Finally,we use the computer simulation to accomplish the example analysis.The results showed that the multivariate support vector machine prediction model has better prediction performance for the market demand under the influence of the multiple factors,and effectively improve the prediction accuracy.Then,we find that the optimization demand forecast can effectively mitigate the bullwhip effect in the supply chain.by analyzing the bullwhip effect in the multivariate support vector machine prediction model.