Study On The Prediction Model Of Urban Parking Supply Based On Car Ownership And Road Network Capacity

With the acceleration of urbanization in China and the rapid growth of the number of urban motor vehicles.It brought challenges to the urban road traffic system,and at the same time,it led to severe urban parking problems.Parking problems not only hindered the development of urban traffic,but also affected the convenience of residents.Therefore,setting reasonable parking supply quantity is the key to solve the parking problem in cities.Based on the urban car ownership and road network capacity,the prediction model of urban parking supply has been established,It not only meets the basic parking demand generated by urban car ownership,but also meets the elastic parking demand generated by residents’ daily travel,it is of great practical significance to study the prediction model of urban parking supply.Firstly,the basic theory of urban parking supply and demand has been analyzed.The urban parking supply was divided into two parts:basic parking supply and fexible parking supply.the influencing factors of urban parking supply from five aspects has been analyzed.The principle,advantages and disadvantages of the existing parking demand prediction model has been summarized.Secondly,the prediction model of urban basic parking supply based on car ownership has been established.The influencing factors of urban car ownership has been analysed,Using the grey correlation analysis method to degree of correlation between the influencing factors and the urban car ownership.According to the selection principle of the influencing factors,the influencing factors that meet the conditions has been selected.Summarizing the existing prediction models,and the support vector machine(SVM)model has been selected to forecast the car ownership.Considering the limitation of traditional grid times calendar method in determining the parameters c and g of SVM algorithm,improved grey wolf optimization algorithm has been introduced to optimize the c,g parameters.The IGWO-SVM algorithm was used to forecast the car ownership in the future,i.e.the basic parking supply.Then,the prediction model of urban fexible parking supply based on road network capacity has been established.Using the space-time consumption method to establish road network capacity of calculation model,determining the model correction coefficient from the three aspects of the effective use area of road facilities,the effective operation time of road network and the time-space consumption of individual traffic,the combination prediction model of Urban Road area based on the quadratic exponential smoothing method and the grey prediction method has been established,which brings the forecasting area into the road network capacity of calculation model,the prediction model of the road network capacity has been obtained;The urban Parking supply and demand relationship model is established based on the restriction of road network prediction capacity,combined with the different travel purposes of residents,the study of regional traffic composition,and the difference of residents’ travel mode to determine the quantity of future urban fexible parking supply under the restriction of road network capacity.Finally,taking Jingyang District of Deyang as an example,the model has been applied and verified in practice;SVM,PSO-SVM,GA-SVM,GWO-SVM,IGWO-SVM,nonlinear regression and BP neural network algorithms are selected to predict the test sample data respectively,and the results show that IGWO-SVM algorithm has the best forecast accuracy,and the algorithm is used to forecast the car ownership in Jingyang District of Deyang in 2020,i.e.the basic parking supply quantity;Using the space-time consumption method and combination prediction model predict the the road network capacity in Jingyang District of Deyang,and using the Parking supply and demand relationship model under the capacity constraint to forecast the number of flexible parking supply in Jingyang District of Deyang in 2020,and finally get the total predicted parking supply.Prediction model of urban parking supply has been established based on the urban car ownership and road network capacity,the problem of urban basic parking supply and flexible parking supply has been solved,meets the parking demand generated by car ownerships and social travel activities,which provides a reliable basis for the construction planning of parking facilities and the formulation of parking policies.