Research On FMCG Supplier Inventory Control Based On Combined Demand Forecasting

Demand forecasting and inventory control,as an important part of supply chain optimization management,directly affect the profitability and service capacity of all enterprises in the whole chain.At present,with the increasingly rich market demand for fast-moving consumer goods(FMCG),the production and sales work of enterprises gradually inclines to the market.The traditional demand forecasting method and single inventory control method can no longer help FMCG enterprises accurately control the market and improve the inventory management level.Therefore,based on the reality,this paper proposes an improved demand prediction and inventory control method for FMCG suppliers.Taking Tobacco Company in C city as the empirical object,correlation analysis is carried out by using part of cigarette data to prove that the method can achieve the expected effect.The specific research contents are as follows:(1)SARIMA-LSTM combined prediction method based on the sum of squares of optimal errors.Taking the monthly sales data of the product as an example,the seasonal characteristics of the product were investigated,and the Seasonal Autoregressive Integrated Moving Average(SARIMA)model and Bi-directional Long Short-Term Memory(LSTM)Neural Network were used to forecast the market demand.At the same time,in order to capture more useful information features,the two single prediction methods are combined linearly according to the sum of the squares of error and the minimum target,and solved by genetic algorithm,the weight is given to the single method,and the new composite model is obtained.Finally,taking the monthly cigarette sales data of Tobacco Company in C city as an example,the prediction results of the combined model were compared with the two single prediction methods,and the Mean Absolute Percentage Error(MAPE)of the combined prediction method was verified to be 5.52%,higher than the prediction accuracy of SARIMA model and LSTM neural network.(2)Research on multi-product inventory control strategy.In order to realize the inventory classification management,this paper classifies the product categories from the two levels of business importance and price.(Q,R),(t,R,S),(t,S)and Vendor Managed Inventory(VMI)replenishment strategy based on stock and sales coefficient are adopted for different categories,and the construction of VMI replenishment model driven by demand based on stock and sales coefficient is emphasized.To verify the effectiveness of the improved inventory control strategy,four kinds of marketing regulations of Tobacco Company in C city were selected for example analysis.After verification,the prediction accuracy of SARIMA-LSTM combined prediction model after application in C company is 94.84%,16.2% higher than before.At the same time,after the improvement of accuracy,enterprises in determining the maximum inventory,safety inventory and other parameters more in line with the market reality,after the improvement of inventory management costs of various products,inventory turnover rate increased.The method proposed in this paper can provide some reference for demand forecasting and inventory control management in supply chain management of various industries.