Sensitivity Analysis Of Rail Transit Passenger Forecasting

With the development of the Socio-economic and the accelerated process of Urbanization, the problem of the urban traffic becomes more serious. The traditional approaches to control traffic congestion is put into practice by increasing infrastructure investment, and the approaches improve the capability of traffic by the expansion of road width, increasing road mileage, clearing the road network nodes, increasing density of road network, however, the approaches will be greatly restricted because of the scarcity of roads and land resources. The policies that control the congestion focused on the expansion of transport capacity by change the transport equipment on abroad, in which, it is one of the most effective measures to develop the rail transportation. In recent years, urban rail transport has been rapid development in China. Until now, there are about 27 cities are planning to build urban rail transit, urban rail transit construction plans of 22 cities have been approved by the State Council, 22 cities which include Beijing, Shanghai, Guangzhou will build 79 rail transit lines, the total length of 2259.84 km, and the total investment will come to 882.003 billion Yuan.Forecasting passenger play a very important role in the early stage of the construction work of rail transportation, it is the starting point of designing and construction of the rail transportation, the basic of the scale of the project construction and operation of economic evaluation, the key elements of the risk assessment of the project, and the passenger flow forecasting has become a special study in the field of rail transport. Many factors of passenger flow are often unstable, such as the nature of land use, population, and the lines net of the regular bus. Which will easily lead to low precision of the prediction data, in addition, due to rail transport is still a new traffic form in China, the forecast of passenger traffic is basic on some foreign models and experiences because of litter experiences in China, but there are a lot of differences between foreign and China on transportation, cultural, economic and other environmental factors, which have led to imprecise of the passenger flow forecast, and there are big gaps between predictive value for rail transit and the actual value of all passenger traffic in many cities, even some gaps exceeding 200%. In this case, it is a great theoretical and practical significance to study the impact factors of passenger flow forecasting for raising anti-risk ability.First of all, the paper introduced the four-stage method which commonly used in the forecast of passenger flow after the presentation of some theory of flow forecasting. On the basis, combining with the passenger flow Subway Changchun forecast, the paper discusses the various modes of the four-stage method and the model which had been chosen by Subway Changchun.Secondly, the paper summarized the influential factors by analysis the modes of four-stage model which used in the passenger flow forecast of Subway Changchun and the experience of rail passenger flow forecast. On the basic of presentation of the methods and procedures of fuzzy comprehensive evaluation method and the test ofχ2 distribution, the paper extracted the qualitative factors and quantitative factors of the sensitivity factors and summarized the sensitivity factors that need to be analyzed by the degree of the impact.Then, the paper analysis some factors by single factor analysis in the sensitivity analysis, and the factors contain all of long-term factors and some recent factors, the long-term factors contains the process of building urban rail transit network, land use in the nature and size of the population, and these recent factors contain competition and cooperation between the rail transit and other conventional mode of transport, vote, the characteristics of travelers. Besides, the rail transit fares and travel time are analyzed by combing multi-factor sensitivity analysis and logit models. And the paper gave analysis table of the sensitivity by the impact of the sensitivity factors in analysis process.Finally, the differences between the actual passenger traffic and passenger flow forecast is divided into the differences in the total passenger flow and the differences in the distribution of passenger flow, on the basic of the above classification, the paper gives the reasons for these differences, at the same time, in order to improve passenger rail transportation capacity to resist risk prediction, the paper gives some measures for passenger compensation and guidance to differences of the rail in the future.