| In July 2018,The State Council issued document No.52,which significantly raised the threshold for cities to apply for metro construction,namely GDP of 300 billion yuan(previously 100 billion yuan only),General public finance budget revenue of 30 billion yuan(previously 10 billion yuan),Permanent urban population of 3 million(previously 1 million people),Such a high requirement broke the "metro dream" of many cities.Trams,on the other hand,can be approved by Provincial Development and Reform Departments,not by National Development and Reform Commission,which provides a new alternative to metro for small and medium-sized cities to develop rail transit.In recent years,the discussion on the construction of modern trams in China has been on the rise,among which there are lots of voices of doubt.Some people think that tram investment is not low,but many cities are often not with ideal passenger flow after operation,compared with its cost,it is still suspected of "waste money".Therefore,the number of expected passenger flow is closely related to the tram investment and construction decision.In addition,it is of positive practical significance to predict the passenger flow of the lines that have been opened and operated to improve the service level of tram operating companies,enhance traffic safety and passenger satisfaction.In view of the problem of passenger flow prediction and management,the main work described in this thesis is as following:Firstly,according to the features of passenger flow of tram,the research status of rail transit passenger prediction,combined with the existing literature,the advantages and disadvantages of different methods and models are compared and studied.Based on the characteristics of tram passenger flow,a relatively mainstream and cutting-edge,operable and scientific prediction method is selected to predict the passenger flow during holidays and peak hours.Then,based on selected Huai’an modern tram line 1 partly holiday history real data,using the traditional grey prediction model with time series of traffic situation on crude judgments,and then use the theory of Markov chain method of gray model prediction error correspondingly adjust,test the result of the prediction again to verify that the availability of the model.Then,a current more popular neural network model is adopted.Through comparison and selection,this paper adopts Gray-RBF(Radial Basis Function)neural network to establish the Grey Markov model and Grey-RBF model respectively by taking the passenger flow values of some large holidays of Huai’an modern tram line 1 in recent years as the original data.After comparison of the prediction results,it is found that,in comparison to the single model,the Grey-RBF combination model has better prediction effect and is more suitable for mining and effectively utilizing the relationship between historical passenger flow data in the case of insufficient data,so as to make a scientific and reasonable prediction of the trend of passenger flow changes,which is easily operated and implemented with low cost.It is suitable for predicting and judging short-term changes of passenger flow with less historical data.Finally,through the introduction of the advanced deep learning algorithm LSTM(Long Short-Term Memory)neural network model,for the unique perspective of holiday peak passenger flow prediction research,and try different parameter combinations to calculate and verify the model,achieved good prediction effect.It is proved that the method has good applicability and application prospect in the prediction of tram passenger flow during holidays.In addition,based on the prediction results,in reference to domestic scholars in the rail transit passenger flow management based on the researches,and with the features of trams,from Fleet Organization,Passenger Flow Guide,Traffic Ancillary Facilities,and Emergency Management of big passenger flow management measures are raised and the suggested measures can enhance operational services by tram corporation and improve large passenger flow management for reference,which could also be decision-making information reference to other cities in China for tram construction planning and operation management. |
