Based On The Category Of Goods Of Retail Supply Chain Demand Forecasting And Inventory Management Model

With the development of economic globalization, a large number of foreign retail enterprises are rapidly entering China, making the competition environment of our retail industry increasingly serious. Facing the new opportunities and challenges, retail companies should strive to enhance the competitiveness of the key part. For the retail, inventory management is not only the core part of business management, but also throughout the entire business process. The monetary value involved in it is particularly large. While inventory management impact on customer service level directly, forcing retailers must focus on it. With the consumer habit changing, the product life cycle is gradually shortened and the demand uncertainty from the market is increasing day by day. The traditional inventory management has been unable to meet the requirement of low-cost and high service for the modern retail. For this reason, the company should establish a new inventory management model which based on the system idea and close collaboration between members of the supply chain. However, the retail holds various commodities which have very different properties. With the constraints of cost and effort, the company should use category management to search for the inventory management model for each category, making sure that every good classification matching the right inventory management model.The research is focus on the demand forecasting and inventory management model based on category using"how to classify– how to forecast demand– how to match model"as the logic thread. First of all, based on two dimensions of importance and risk, we construct the retail classification system. Then we use AHP to determine the index weight and make an empirical analysis with the data of a retail company. Secondly, according to the demand forecasting method which is the critical problem in inventory management, we analyze the forecasting effect difference of exponential smoothing, SARIMA, neural network and combined forecasting model, and discuss the reason for the differences through category characteristics and data feature. Finally, according to the characteristics of key commodity, bottleneck commodity, important commodity and common commodity, we use the idea of category management to match them with CPFR, JMI, VMI and tradition inventory management. The proposed research ideas and methods in this paper also have some reference meanings for the inventory management in other retailers.