Research On Metro Network Passager Transfer Routing

In the field of Intelligent Transportation Systems (ITS), traffic network dynamic passenger flow assignment theory is the key technique. Although dynamic passenger flow assignment theory has relatively wide application in traditional public transportation system, it meets many problems while combing with real-time metro passenger flow density index to provide dynamic metro interchange information. Traffic guidance system is the future direction and trend in the development of rail transit operation production management. How to effectively guide the passengers to choose the optimized travel route, in order to reasonably reduce the passenger flow management pressure in the core metro lines at morning and evening peak time, is one of the problems of focus in traffic system engineer and science research. To solve the previous problem, this article, based on the basic data in Beijing rail transit system, focuses on the following critical problems:short term prediction of passenger flow, passenger flow density index system, passenger flow route guidance model, and so on. The article achieves the goal of passenger travel route optimization in rail transit network operation production management. The main study content of this article is as follows:(1) Propose the short term passenger flow prediction algorithm based on the support vector machine. First, study out a passenger flow prediction algorithm by combing genetic algorithm and support vector machine algorithm. Genetic algorithm could optimize the parameter in support vector machine algorithm, resulting that the combined algorithm producing more precise prediction effectiveness. Second, study out a passenger flow prediction algorithm by combing wavelet transform and support vector machine. Wavelet transform could decompose passenger information into high frequency data and low frequency data without loss, and generate multi-scale detailed low frequency sequence. Then the support vector machine makes prediction for a low frequency sequence and many high frequency, sequences. At last conduct wavelet reconstruction to the many sequences gotten by prediction, and get the final passenger flow prediction result. This article is based on the Beijing rail transit network passenger flow data. Under many evaluation criterions, empirical investigation results indicate that the passenger flow prediction algorithms proposed in this article could lead to better results than many of the current popular passenger flow prediction algorithms.(2) Propose a subway density based path choice model. First, divide rail transit network into several layers, and propose district, line, and the whole road network passenger flow density index. Empirical investigation indicates that, the passenger flow density index from different layers all could well reveal the passenger flow density in rail transit network. Second, on the basis of passenger flow density index, an improved rail transit real-time passenger flow Logit assignment model is proposed. The model uses network traffic density and road network-based data to calculate the real-time path cost, dynamic adjustment path allocation ratio and simulate network traffic distribution. Empirical investigation results shows that the passenger flow density index system proposed in this article could give a better description of the passenger flow information in the rail transit network.(3) Propose an time-to-time proportional-switch adjustment model for path guided. First, on the basis of passenger flow density index, study out a real-time choosing generalized cost least information-driven route guidance model according to the information in traffic guidance system, to achieve the aim of dynamically adjust the rail transit network passenger flow pressure. Second, demonstrate the feasibility by mathematic deduction, and deduce and verify many basic characters of the model. Finally, simulate at the three levels of "district-line-road network". Empirical investigation results shows that, traffic guidance information could well guide the passengers to choose the optimized travel route, reduce the passenger flow density in some of the lines and areas of rail transit network, reasonably reduce the passenger flow management pressure in the core metro lines at morning and evening peak time, optimize metro interchange, and achieve the aim of improving the status of rail transit network.