Research On The Prediction Methods For Small Car Driver Training Deamnd

Rapid development of domestic economy and huge demand of drivers’ licenses have lead to a surge in driver training volume recently. Meanwhile, driver training market in China has gone through dramatic changes. Cities all over the country are setting about reforming local training systems. As significant basis of the reform, prediction of driver training demand becomes the current problem to be solved. This paper analyzes characteristics and future trend of motor vehicle ownership in China and its potential influence on driver training demand; describes driver training system and process in the current stage; discusses characteristics of driver training demand and its inner relation with traffic safety.Owing to the condition that traditional forecasting methods are unable to adapt to the changing context and therefore cannot correctly predict the future trend, this paper divides prediction methods into two categories - stage forecasting and systemic prediction.In stage forecasting, pros and cons of traditional forecasting methods are compared to choose a suitable method as the reference one. Meanwhile, a new forecasting method based on an S curve and saturation theory is proposed. In systemic prediction, this paper researches the periodic feature of small cars and finds out demand characteristics and factors of each stage in its lifecycle; establishes the causal relationship tree and flow charts in four separate subsystems; provides basic methods of parameter estimation and initial value determination; improves validity of the model by feasibility and reliability testing; and carries out simulations with contextual assumptions.Lastly, a case study of small car driver training volume in the city of Huai’an is used as an example to validate these methods and for comparison. Further analysis of the causes of results differences are given eventually. Meanwhile, based on the simulation platform of system dynamics and the specific data of Huai’an, policy simulations under different scenarios are carried out to explore the effects of different policies on driver training demand.