Research On The Calculation Model Of The Suburban Railway Network Scale Metropolitan Area

Suburban railway plays an important role in promoting regional synergy and optimizing the spatial pattern of metropolitan areas,and the planning and construction of suburban railway in metropolitan areas began to take off.since the 14 th Five-Year Plan in China.As one of the basic contents of the network planning,the reasonableness of the network scale is related to the construction investment and operation service level.At present,most of the conventional network scale determination methods are based on qualitative analysis and single indicator calculation,and the calculation results are usually discrete and cannot serve the network planning well.The paper aims to develop a quantitative network scale calculation model that integrates multiple indicators to provide a theory and method for the planning of suburban railway networks in metropolitan areas.Based on the analysis of the factors influencing the scale of the metropolitan area suburban railway network,this paper constructs a three-level index system affecting the scale of the metropolitan area suburban railway network,applies the theory and method of support vector machine in the field of machine learning to the calculation of the scale of the network,determines the parameters of the model through the training and validation of samples,develops a model for calculating the scale of the metropolitan area suburban railway network,and uses the model calculate the scale of the metropolitan area suburban railway network in the Shanghai metropolitan area.The main work and results of this paper are as follows:(1)Based on the analysis of the factors influencing the scale of the metropolitan area suburban railway network,an index system for the scale of the metropolitan area suburban railway network is proposed.Based on the basic characteristics of metropolitan area and suburban railway,the Interpretative Structural Modeling Method(ISM)is applied to reveal the multi-level control relationships among the influencing factors,and a three-level indicator system for the scale of metropolitan area suburban railway network,such as total economic volume,total population size and scale of existing transportation facilities in the metropolitan area,is constructed,and the rationality of the indicators is verified by applying the Spearman correlation coefficient.(2)A standard support vector machine(SVR)based model for calculating the scale of suburban railway network in the metropolitan area is developed using machine learning theory.To develop a network scale calculation model integrating multiple indicators,a SVR learning method in the field of machine learning is introduced,and 15 metropolitan areas where foreign suburban railway have developed maturely are used as sample data,and the constructed line network size indicator system is used as input to train the calculation model,and the training results are analysed.(3)An improved SVR based model for calculating the scale of suburban railway network in metropolitan areas is developed.In order to solve the problem of poor fitting of the standard SVR model,two optimization-seeking algorithms combining grid search and cross-validation(GSM-CV)and improved particle swarm algorithm(APSO)are designed to obtain the optimal parameters,so that the scale calculation model of the suburban railway network in the metropolitan area based on the improved SVR is established.(4)The model is validated with the calculation results of the scale of the suburban railway network in the Shanghai metropolitan area.The calculation model was used to calculate the scale of Shanghai metropolitan area suburban railway network as 1544.65 km,and the feasibility of the calculation model was verified by comparing the calculation results of Traffic Demand Method and Service Level Method,as well as the scale of network in the approved recent construction plan.