Study Of Market Demand Prediction System On ANN In Auto Parts

As the prime mainstay of modern manufacturing industry, Auto industry is the basis and protection of China's modernization, urbanization, which plays an important role in national economy. Auto parts are the backbone of the whole auto industry.In recent years, China's automobile production increased dramatically, which promotes the development of auto parts industry. The competition in auto parts has become more and more intense. In order to win in competition, the units and individuals which are engaged in commodity production and marketing must provide customers products at the lowest cost and in the shortest time. Estimating the changes in the market correctly and the prospects for business development and predicting market demand timely and accurately are key element of business success. The correct forecast of market demand for our products not only contribute to the development of operational management of production planning, procurement and staffing plans, but also is conducive to managers to determine inventory levels, the establishment of the expenditure budget, forecast cash flow and financing needs.Based on status quo of the auto parts market demand, thesis studied auto parts market demand forecasts for the study system through comparing several methods of forecast, main factors of the impact market demand for auto parts are analyzed ,and improved BP neural network prediction model was applied to predict market demand for auto parts system. Combined with an auto parts company's ERP project, an open market demand forecast system for auto parts which can be integrated with enterprise ERP is developed.The works done by the author in this thesis are as follows:1. According to the characteristics and development process of auto parts industry, the existence problem of market demand for auto parts and main factors of the impact market demand for auto parts are analyzed, the existence of complex non-linear relationship between the impact of the main factors and market demand for auto parts is studied.2. Analyzing the application of the multiple linear regression method, exponential smoothing and neural network method in the model of predict market demand for auto parts, And the use of auto parts market historical data to verify the model compared with a forecast of market demand for auto parts, artificial neural network to predict the results of more accurate and reliable.3. The BP network problems was analyzed in this paper, the improved BP network was put out , the relevant parameters were selected and optimized, auto parts for a production to market demand characteristics of a comprehensive, in-depth were analyzed, the auto parts market demand forecast model was developed and with which forecasts the demand for auto parts market.4. This research studied the structure prediction system of market demand for auto parts and its composition and function modules, the data mining prototype developed using SQL language and C++ Builder 5 in a SQL Server 2000 environment.The prototype system was applied in to the some enterprise actually and obtained preferable effect. These confirmed the feasibility of theory study and implement method. So this method and system can be applied to enterprise widely and effectively.