Research On Public Transport Route Selection Model Based On Mericulous Population Classification

With the sustainable and rapid development of socio-economic and the speeding up of urbanization process,there are more and more problems such as traffic congestion and environment pollution appear in many first-tier cities and parts of developed second-tier cities in a growing trend.By the way,the development of green transport,promote intensive public transportation is a good recipe to relieve urban malaise,which is internationally recognized.Planning and Development of Transportation Development in Beijing during "13th Five-Year " is explicitly mentioned the need to promote the construction of urban transit.In view of the current situation of public transport,there is a common problem that the proportion of transit travel is low,and the level of service is poor.If we want to relieve these two items and optimize the structure of passenger flow,there is a great significance to analyze and study the mechanism of the transit flow and distribution of passenger flow in the public transport system.Therefore,with the in-depth excavation and analysis of public transport data,this paper does a research on the path selection problem of multi mode public traffic scenarios,which has a great significance for the analysis of passenger route choice behavior theory and the application of big data to improve the bus intelligent decision.In this paper,the data cleaning and fusion of research objects were carried out to obtain kinds of high accuracy and usability passenger flow at first.Then three key parameters such as departure time,travel purpose,travel destination were selected as the inputs of C&RT decision population classification.Based on the framework and parameter threshold of the classifier,an appropriate population classifier was finished.After then,with a comprehensive analysis of the characteristics of path selection results of various passenger types based on the eight path parameters,such as travel time etc,the conclusion is put out that different groups of transit travelers will have widely different route choices.On this basis,the BP neural network parameters prediction model was built up with three inputs which had used in the group classification and three outputs such as travel time,travel cost,travel punctuality.In addition,the path selection model was established with the concept of "distance" further.Subsequently,the accuracy and feasibility of the methods were illustrated from both aspects of theory and example calculation.Finally,the main conclusions of this paper were summarized and the future research directions were pointed out.