Research And Application Of Higway Toll Prediction

With the speedy development of the highway in China, the prediction of highway tolls is one of the core issues in the research of highway management. Great management of toll forecast is the guarantee of the implementation of highway tolls policy as well as the significant section of the construction and development of highway network. The prediction not only reflects the operational evaluation and management status in highway toll station, but also influences the toll rate, toll management, right transfer and etc. in the highway toll. Therefore, the prediction provides important basis for the decision makers in policy formulation and implementation.This thesis analyzes the characteristics of highway tolls data and elaborates the research status of normal prediction theory and thought in the prediction of highway tolls. Considering the disadvantage of existed prediction thinking and choosing time series theory and gray theory as research tools, the research purposed a prediction analysis of highway tolls for short-term and long-term toll forecast. The main contents are as follow. Firstly, the ARIMA(Autoregressive Integrated Moving Average Model) is studied in forecasting highway tolls. The result illustrates that after modification of ARIMA by considering data feature. Secondly, aiming at the limitation of mid-term and long-term prediction of ARIMA, optimized GM(1,1) and a combination of ARIMA and GM(1,1) by applying optimal weighting factor method is built respectively. The results illustrate that the combined model acquires greater accuracy and better fitting with respect to single prediction model. Moreover, it also demonstrates the development of the law of tolls. Finally, the highway toll forecasting system is designed and achieved by utilizing Java, Javascript and Fine Report based on the forecasting models.The research shows that the ARIMA and combined model improve the forecast error obviously which is lower than 5% and meet the demand of the prediction. The prediction system based on forecasting models has evident advantages which improves the normative and efficiency of the toll management of the administration and verifies the feasibility and practicability of those forecasting models in highway tolls prediction.