Research On Bullwhip Effect Of Enterprise Supply Chain Based On Neural Network

The development of technology and the diversification of customer needs have prompted companies to face the problems of shortening product production cycles and accelerating product development to quickly respond to customer needs,thus creating a supply chain.The supply chain formed by three or more upstream,core enterprises,and downstream can more quickly transfer products and information from the supply side to the final consumer,while also allowing enterprises at each node of the supply chain to focus on their own Core business,enhance competitiveness.The bullwhip effect is an objective phenomenon in the supply chain.In order to reduce the instability of the production plan caused by the deviation of actual demand and planned quantity,each node enterprise in the supply chain often needs to actively choose to increase the quantity of safety stock to ensure normal production activities.And trigger the bullwhip effect.Solving the problem of the bullwhip effect in the supply chain has become a core issue for improving the efficiency of the entire supply chain.In supply chain management,demand forecasting is the basis for driving business processes.The enterprise's setting of safety stock is based on the deviation value of demand forecasting.The smaller the deviation value,the smaller the safety stock quantity.The existence of safety stock is the direct cause of the bullwhip effect.Node companies evade the risk of out-of-stocks by actively increasing safety stocks,which causes the bullwhip effect to gradually increase in the supply chain.Therefore,the accuracy of demand forecasting can reduce the quantity of safety stocks and weaken the impact of the bullwhip effect,so as to improve the ability of each node company to respond to market demand fluctuations in a timely manner.Based on the above background,the thesis first analyzes the cause and specific performance of the bullwhip effect based on the business process in the supply chain with the manufacturer as the core.Through quantitative analysis of the safety inventory,demand forecast and bullwhip effect,the improvement is determined Demand forecasting accuracy can weaken the influence of the bullwhip effect,and proves that this method is applicable to all nodes in the supply chain.Considering the limitations of some traditional time series forecasting methods commonly used by enterprises in the past,non-linear data features cannot be captured,and their accuracy is limited in practical applications.In recent years,the application of artificial intelligence technology,especially neural network algorithms,has been widely used in In various fields,this article uses a neural network-based model for demand forecasting,which provides new ideas and methods for supply chain management.Due to the many factors influencing demand forecasting in the supply chain in a complex market environment,the thesis uses the grey correlation analysis method to select the main characteristic factors such as price,seasonality,and promotional strength.Secondly,considering the shortcomings of the traditional forecasting model,based on the study of the core ideas and methods of the neural network,the thesis puts forward the LSTM-BP combination model,and uses this model to realize the demand forecast in the future period of time.In the application of the model,according to the current production management disorder of a manufacturing enterprise and other current situations,the company's product orders in the past two years are used as the research object for empirical analysis.The results prove that the product forecast deviation under this model is greatly reduced,and the degree of the influence of the bullwhip effect on the node enterprises is effectively alleviated.Therefore,the LSTM-BP combination model proposed in this thesis has good application value for weakening the bullwhip effect in the supply chain.