The Prediction Of The Construction Scale Of The National And Provincial Arterial Highway Network In Xinjiang Based On The Neural Network

The reasonable scale of the highway network is one of the core contents of the planning research of the regional highway network.It is an important subject in the highway-network planning that how to determine quantitatively the reasonable construction scale of the highway network.In recent years,scholars in China have done some research on the construction scale of the highway network,and have achieved relatively plentiful results.However,due to the special geographical environment of the Xinjiang Uygur Autonomous Region,the factors influencing the construction scale of the highway network are quite complicated.All of these influencing factors are intricate and complex,which make it’s very difficult to predict the construction scale of the highway network.At present,the methods of predicting the construction scale of the highway network are mainly the experiential summary method and the analogy method,which are of strong subjectivity,so the accuracy of predicting the highway-network construction scale would be doubted.Therefore,this article aims to improve the current methods of predicting the highway-network construction scale,trying to use the neural network method to predict the construction scale of the arterial highway network in Xinjiang Uygur Autonomous Region,so that it can provide guidance and suggestions for the construction and development of the arterial highway network in Xinjiang Uygur Autonomous Region.(1)In this paper,the theories related to the highway-network construction scale were discussed.At the same time,the total volumes of transport demand in the study of the highway-network scale prediction and the related predicting methods of the total volumes of the highway-network scale were discussed and enumerated.Several kinds of total volumes of transport existing currently and the advantages and disadvantages as well as applicable conditions of the methods of predicting the highway-network scale were deeply analyzed,providing reference for choosing the model of predicting the highway-network scale.(2)In this paper,through fully analyzing the factors influencing the total demand for highway transport,the GRNN general neural network model was established to predict and analyze the total demand for highway transport in Xinjiang Uygur Autonomous Region.The model was compared with the traditional model and method,which provides a theoretical basis for properly controlling the future development law of highway transport in Xinjiang Autonomous Region.At the same time,the feasibility of the neural network model in predicting the highway-network scale was verified.(3)This paper constructs the prediction model of total highway network scale based on BP neural network,and forecasts the total highway network scale in 2025 and 2035 in Xinjiang.According to the development trend of trunk highways in Xinjiang and combined with the prediction results of total highway network scale,the construction scale of trunk highway network in 2025 and 2035 in Xinjiang is determined by investigation and statistics method.Finally,the traditional method is used to compare with the model The results can provide decision-making basis for highway development direction in Xinjiang.(4)Finally,combined with the mean impact value(MIV)algorithm,the factors influencing the total demand for highway transport and the total volume of the construction scale were quantified.Combined with the quantified results of the MIV influencing factors,the goal of the grade allocation of arterial highway network with the maximization of transportation demand and the most coordinated technical grade and the least investment was established.Then a multi-objective optimization model of technical grade structure of the arterial highway network was established.Combined with the current construction and development of the highways in Xinjiang Uygur Autonomous Region,the measures and suggestions for the future highwaynetwork development were proposed.(5)The results show that the application of neural network model to calculate the scale of Xinjiang trunk road network is more adaptable than the traditional model,which can provide reference for the future planning of Xinjiang road network,and provide a new idea for the planning of Xinjiang road network.