The Design And Implementation Of ARMA Sales Forecasting Model For Product Life Cycle

Nowadays, many traditional manufacturing enterprises are faced with more and more challenges, such as high cost, high competition and so on. With efficient production management, we can reduce the cost of production and improve the production quality and efficiency in order to make more profits for enterprises. Therefore, manufacturing enterprises have begun to take more attention to production management. And informationization may enhance the forecasting capability of future product sales in order to enhance the efficiency of production of enterprises. Arranging production plans plays an important role in production management, so its informationization is very important.If we could predict the future sales of new products precisely according to the sales history of its relevant product, we may arrange our production plan and make products based on the result of sales forecasting. If we do it, we can not only reduce the cost of inventory, but also reduce the risk of losing orders because of the shortage of stock in order to enhance the profits. Based on the actual conditions of the enterprise, it will provide great profits and value for the enterprise that we can create a suitable sales forecasting model to predict the future sales of the product precisely.Based on analyzing main forecasting theories, we select a suitable forecasting model according to the actual conditions of Company K. Combined with product life cycle theory in marketing, this thesis constructs an ARMA(Auto-Regressive and Moving Average) sales forecasting model for product life cycle creatively and chooses Flex, a kind of RIA(Rich Internet Application) technique, which has rich capability of presentation to design and implement the sales forecasting model.This thesis has mainly completed the following researching tasks:1. Introduction of the concept, theory and main model of forecasting, including Regression Analysis Prediction Model, Time Series Analysis Model, Grey System Theory Model and Artificial Neural Network Model. Based on the research, we analyze the strength and weakness of different models further. According to product life cycle theory, this thesis introduces the concept and characteristic of product life cycle stages including introduction, development, mature and decline.2. Based on the theory research, this thesis analyzes the problems and needs of Company K, selects a suitable forcasting model, develops the strategies of modeling and constructs a time series analysis model——ARMA sales forecasting model based on the actual conditions of Company K. Furthermore, this thesis designs a method to check and recover the prediction error of the model and improves the model by adding the information of orders into the model so that the model could be more practical.3. Introduction of the concept of RIA(Rich Internet Application) technique and Flex, a kind of main RIA techniques. Based on the technique, we design the software architecture and hardware architecture of sales forecasting system. The software architecture includes three parts: software integral architecture, Web App architecture and batch program architecture. The hardware architecture divides the network of the system into DMZ(Demilitarized Zone), application server zone and utilization management server zone. At last, we design and implement Web App, batch program and the core part of the system: statistic forecasting value calculation. And we validate the forecasting model by the history data of Company K.