A Study Of Oil Sales Forecasting Model And Application For ZJ Company

It is generally known that oil is an important resource for National development.In China,there is an imbalance in supply and demand of oil sales and the problem of energy shortage is becoming more and more serious.The key to solve this problem is to find high accurate oil sales forecast solution and then accurately distribute the oil resources to each province.This article takes ZJ Oil Company as the research object,ZJ Company is a major oil sales company in Zhejiang province,mainly sales production is gasoline,and diesel and Kerosene.ZJ Company has three main problems in oil sales forecasting: No system support,long forecast period,low accuracy.Aiming at this problem,this article provides a new method on oil sales forecast.According to the different sales situation of gasoline,diesel and kerosene,analyze the factors which affecting the sales of oil products.In addition,consider China's economy impact on oil sales,confirm the real factors on sale oil production.More comprehensive than only considering oil factor,and can greatly improve the prediction accuracy of oil sales.The traditional grey prediction model GM(1,1)not only has the problem on the input factor is single,but also the prediction accuracy is low.Therefore,this article presents an optimized grey prediction model GM(1,N).This method improves the model function by introducing the background value and multi factors.The research results show that the prediction accuracy of GM(1,N)is better than GM(1,1).Using grey relational analysis to calculate the correlation degree of various factors which affecting the oil sales.Establish grey prediction model,time series model and multiple regression model.GM(1,3)has better than other models,and close to the real data.To build the sales forecasting function based on HANA data warehouse.Finally,the prediction results are displayed by the SAP Business objects.HANA has strong computing power,flexible and fast implementation.Using this tool can solve the problem on sales prediction which sales prediction without system support,the long cycle successfully.This solution also has some reference value for the prediction of China oil market.