Study On Dynamic Optimal Accessible Paths In Urban Mass Transit

With the rapid development of urban mass transit(UMT), transit network is formed, however, the operation of each line is not coordinate. Due to time constraints the last train, Passengers travelling at evening cannot transfer successfully at the transfer stations. Thus, how to find optimal accessible paths is important for passengers and managers. The objective of this thesis is searching the dynamic optimal accessible paths under the network operation, and some related content is studied.Firstly, the accessibility is defined, and then the accessibility factors are described. It is pointed out, that transfer time is the main factor of accessibility in the established network. Different types of passengers due to differences in walking speed, its cost of the transfer time is different. by using the idea of the distribution function in probability and mathematical statistics theory, a sample survey is conducted for passengers’ walking times to get through every kind of transfer facilities. Fit the distribution of different walking times of transfer facilities with these data. According to the relations between transfer time and walking times through various transfer facilities, the idea that passengers’ transfer time can be described with the lognormal distribution is proposed. Transfer times of different passengers can be decided with endpoint values and middle value of the confidence interval. The study of the transfer time provides data to support model.Secondly, the establishment of UMT network models is based on the weighted undirected simple graph. Based on the network structures and the forms of train operation,3kinds of key nodes in the network models are defined. On the basis of the network models, the accessibility judgment rules are formed. For reasons of the shortest travel time and dynamics of optical path accessibility, establish the mathematical model of dynamic optimal accessible path.Thirdly, the improved GA is detailed designed for solving the model. The generation process of the initial population in GA is optimized. Accessibility factor is introduced in the fitness function, to ensure accessible paths are having priority to be selected. Judgment processes are inserted in each of the operator’s operations to ensure connectivity of accessible paths.Finally, the study of the dynamic optimal accessible paths is applied in the simulation system. Three travel time references of each transfer station in Beijing UMT network are calculated. By computer programming, combining the models and algorithm for the study on dynamic optimal accessible paths, experiments on actual cases are conducted. By setting different starting travel time and different transfer time values, search the optimal accessible paths. The dynamics of optimal accessible path is verified. Meanwhile, the operation results also illustrate the practical significance of this research.