Demand Forecasting Research For Management Of Tyre Inventory

As tyre sales are usually restricted by market, production plans are fluctuated, and different tyre types will induce different dosage of materials, it is difficult to get exact stock plans. In order to deliver the tyres to customers to meet customers' needs as fast as possible, the amount of inventory should be increased. But excessive inventory not only occupied plentiful funds, but also lead to appear unclear storage location and quantity. In order to resolve above problems, it is need to study the project of requirement forecasting problem of inventory, then we can enable the tyre raw materials and products to basically achieve the balance between purvey and requirement, production and marketing.At present, with rapid development of the tire industry, tyre inventory management affects the company's operating efficiency, it is quite required to study the demand for tyres. Experts and scholars in the world had done a lot of research on the demand forecasting, but very few study the tyre demand forecasting. This thesis introduces Bayesian neural network prediction for tyre demand forecasting firstly.The week or month demands of tyres are predicted by simulation using BP neural network model and Bayesian neural network model in this thesis, and the prediction results of BP neural network model and Bayesian neural network model are compared. The simulation results indicated that bayesian neural network model is better than BP neural network model, and satisfactory effect can be got using Bayesian neural network theory .If we have known the demand prediction of tyre, we can easily get the the number of tyres in storage and the safety stock amount of tires by establishing quantitative relationship of tyre sales and materials and collating recent inventory. Thus we can receive the scientific inventory, and then tyre production can meet the market demand, by which effective tyre inventory can be reduced, and the company's cost-effective will also improved . Demand forecast system of tyre inventory is realized in this thesis using visual C++6.0 and MATLAB. We make the human-machine interface using Visual C++6.0, and realize the forecasting system using MATLAB engine if we set Bayesian neural network as backstage in MATLAB.